BIOLOGICAL OVERVIEW

In Drosophila, genetic ablation of eyes or mutations in the well-characterized visual phototransduction pathway do not block entrainment of circadian clocks to light. In other words, blind flies still exhibit a photoperiod response. For example, Drosophila mutants norpA (norpA codes for phospholipase C involved in visual phototransduction) and ninaE (ninaE codes for the Rh1 rhodopsin visual pigment in rhabdomeres R1-R6) still have robust and light-sensitive rhythms (Dushay, 1989; Zerr, 1990; Wheeler, 1993; Suri, 1998 and Yang, 1998), indicating that there is at least one additional relevant photoreceptor and signal transduction pathway independent of the now classical photoperiod transcription factors, Period, Timeless, Clock, and Cycle. The existence of such a pathway finds support with the cloning of the cryptochrome (cry) gene, coding for a cryptochrome/photolyase, a member of a flavin based light-intercepting protein family whose members have been previously implicated as sensors of DNA damage. Drosophila Cry may be implicated in DNA repair activity. The relevant residues are well conserved within the family, but some members with comparable conservation (mammalian cryptochromes, for example) fail to manifest repair activity. Members of this family bind DNA and have characteristics of transcription factors. All characterized family members are directly photosensitive and include plant blue light photoreceptors. cry transcription is under circadian regulation, influenced by the Drosophila clock genes period, timeless, Clock, and cycle. Cry protein levels are dramatically affected by light exposure. Importantly, circadian photosensitivity is increased in a cry-overexpressing strain. These physiological and genetic data therefore link a specific photoreceptor molecule to circadian rhythmicity. It has been proposed that Cry is a major Drosophila photoreceptor dedicated to the resetting of circadian rhythms (Emery, 1998 and Stanewsky, 1998).

The characteristics of this additional pathway can be addressed using action spectra. Action spectra have previously been used to measure phase response curves (PRCs) for Drosophila pseudoobscura eclosion; in these earlier studies, the PRCs peaked at 400-500 nm (Frank, 1969 and Klemm, 1976). Only recently have similar Drosophila melanogaster action spectra been derived for adult locomotor activity rhythms and for TIM degradation (Suri, 1998). Both curves are significantly different from that of Drosophila vision, suggesting a novel blue light circadian photoreceptor that affects behavioral rhythms as well as the response of peripheral clocks, even in the eye. In mammals, the retina is very important for circadian light perception and makes synaptic connections with the suprachiasmatic nucleus via the retino-hypothalamic tract. But the relevant photoreceptor(s) is uncertain, since the well-studied mammalian visual photopigments may be insufficient to account for ocular circadian photoreception (Foster, 1998). Therefore, as in Drosophila, the mammalian eye may contain another unidentified photoreceptor specialized for circadian light perception (Miyamoto, 1998 and Soni, 1998).

Given the action spectrum, it has been suggested that the unknown Drosophila photoreceptor might be a member of the photolyase/cryptochrome family, a well-characterized group of blue light-responsive flavoproteins. Photolyases/cryptochromes have five well-characterized subgroups (Kanai, 1997). Three groups of cyclobutane pyrimidine dimers (CPD) photolyases are responsible for the repair of thymidine dimers caused by UV irradiation. Another group contains the 6-4 photolyases, as found in Drosophila (Todo, 1996), responsible for the repair of UV-generated 6-4 thymidine dimers and 6-4 photolyase homologs. A fifth group contains plant blue light photoreceptors (Ahmad, 1993 and Lin, 1998). Evidence from several systems suggests that a flavin-based system, in addition to or instead of retinal-containing proteins, contributes to circadian photoreception. Caroteinoid-depleted Drosophila show normal photosensitivity (Zimmerman, 1971). In Neurospora, a genetic defect in the riboflavin synthesis pathway inhibits the photoentrainment pathway. In Arabidopsis, two cryptochromes have been shown to be involved in three different light-dependent processes: hypocotyl elongation inhibition, phototropism, and photoperiodism of flowering time (Ahmad, 1998a; Guo, 1998; Lin, 1998). Flowering time photoperiodism may be related to circadian rhythms (Guo, 1998). In mammals, a recent report describes the expression pattern of two novel cryptochromes, mCRY1 and mCRY2 (Miyamoto, 1998). The transcripts are expressed in retina and the suprachiasmatic nucleus, tissues highly relevant to circadian rhythms. Moreover, mcry1 transcript undergoes circadian cycling in the SCN. Although there is no evidence that the SCN is directly light sensitive, the authors propose that these cryptochromes are involved in mammalian circadian photoreception. Importantly, there are no physiological data that link these or any other specific cryptochromes to circadian light perception in any animal system (Emery, 1998).

A property shared by many clock gene transcripts is that their abundance is subject to circadian oscillation. To assay CRY mRNA, Northern and RNase protection analyses were performed on RNA isolated from heads after fly entrainment under standard light/dark conditions (LD). Both assays gave identical results: CRY mRNA manifests ~5-fold amplitude cycling, with a peak at ZT1-5 (1 to 5 hours after the lights go on) and a trough at ZT17 (five hours into the dark part of the light/dark cycle). A higher resolution RNase protection experiment reveals that the peak persists from ZT1 to ZT7, and the trough from ZT17 to ZT19. CRY mRNA cycling was also measured under conditions of constant darkness. Cycling persists under these conditions but with a lower, approximately 2-fold amplitude. Lower amplitude cycling in constant darkness has been previously described for other clock genes or for clock gene-derived reporter genes. Because there is good evidence for peripheral clocks in multiple tissues throughout Drosophila, RNA extracted from fly bodies was also examined. CRY body mRNA cycling has a similar phase to head mRNA cycling but only a 2.5-fold amplitude. There is precedence for lower amplitude clock mRNA cycling in bodies as compared to heads (Emery, 1998 and references).

CRY mRNA cycling suggests that the protein abundance also cycles during the day, as shown previously for Per and Tim, and very recently for Clk. To address this question, a rat antibody directed against the N terminus of Cry was used in Western blotting experiments. The antibody shows specific immunoreactivity to in vitro translated Cry. The very strong signal obtained with extracts from heat-shocked flies containing the cry gene under hsp70 promoter control shows that the antibody also specifically recognizes Cry in head extract. The protein migrates with an apparent molecular weight of approximately 60 kDa, as predicted from the cDNA sequence. Two bands were visible on some gels, whereas a single fuzzy band was detected on others. The protein abundance was measured in light/dark entrained wild-type (Canton-S) flies, as well as in y w;timgal4 flies (flies with white, pigmentless eyes). The latter show a very robust 8-fold amplitude cycle, which is strikingly different from the CRY mRNA cycle. Protein is low during the day when the mRNA is high, clearly increases before the mRNA rise, and peaks at ZT23 (an hour before the beginning of the lights on part of the cycle). Two hours later, at ZT1, when the CRY mRNA peaks, Cry protein is already strongly reduced (50% of peak value). This indicates that a translational or posttranslational mechanism makes a strong contribution to Cry regulation. In wild-type Canton-S flies, the cycle was less robust (3- to 5-fold), and the trough is reached between ZT11 and ZT13. Between ZT1 and ZT9, protein levels are approximately 40%-50% of the peak. Lower amplitude Cry cycling in CS versus y w flies is probably due to light shielding by eye pigment, as described for the Tim light response (Suri, 1998 and Emery, 1998).

To determine whether the Cry protein cycling is light driven (for example, if the Cry protein itself is denatured by light), protein levels were measured under constant dark (DD) conditions. Surprisingly, and in sharp contrast with Per, Tim, and Clk, the Cry DD pattern is completely different from that in LD and increases continuously in constant darkness throughout the subjective day and night, beginning at the levels reached at ZT23 of the previous day. In DD conditions as well as in LD conditions, the protein and mRNA profiles are completely different. These results strongly suggest that Cry cycling is in great part under light control. To further address the role of light, Cry protein levels were examined in per⁰, tim⁰, Clk, and cyc⁰ arrhythmic backgrounds. As expected, relative protein levels correlate with the relative RNA levels, higher in Clk and cyc⁰ than in per⁰ and tim⁰. But there was also robust Cry cycling in all these clock mutants, in contrast to what was observed for CRY mRNA cycling under the same LD conditions. Taken together, the results show that Cry expression is light regulated at the translational or posttranslational level as well as clock regulated at the transcriptional level (Emery, 1998).

It is presumed that photon capture by Cry itself increases proteolysis, and the data suggest that Cry is very stable at night and unstable during the day. Perhaps the CRY mRNA oscillation is important for the proper increase and decrease in Cry levels that take place after lights off and lights on, respectively. These might be more pronounced or more important during the gradual changes in illumination that occur during natural light-dark cycles. Interestingly, Arabidopsis CRY2 is also light sensitive and accumulates at low light intensities (Ahmad, 1998b and Lin, 1998).

Light-mediated changes in Cry levels may be important for signal transduction downstream of Cry, or they may occur subsequent to signal transduction. Signaling may take place by electron transfer, protein-protein contact, or by changes in phosphorylation state. These possibilities are just beginning to be explored in the case of plant and animal cryptochromes (Huala, 1997; Zhao, 1997; Ahmad, 1998a). But the ultimate targets in the case of Drosophila Cry are likely to be clock molecules. Indeed, Tim levels are also light sensitive, and there is evidence that this molecular light response is relevant to the behavioral light response (Suri, 1998 and Yang, 1998). Since Tim is unresponsive in the cry^b mutant [a recessive third chromosomal mutation that abolishes bioluminescence rhythms, identified by Stanewsky (1998)] and Cry light regulation is independent of other clock molecules, including Tim, this establishes a clear epistatic relationship between Cry and Tim. Yet CRY mRNA cycling is clock regulated, indicating feedback between the central pacemaker and this clock input component. Conservation of the putative DNA-binding residues in Drosophila suggests another, somewhat audacious possibility: Cry might be a DNA-binding protein and function as a transcriptional regulator of light-induced gene expression. Subcellular localization and mutagenesis studies should aid in determining whether Cry-DNA binding is relevant to the light response in Drosophila. In mammals, mper1 transcription is rapidly induced by light in a rhythm-relevant manner and there are other data that link immediate early gene expression to phase-shifting light pulses. mCRY1 and 2 have not been shown to be functionally relevant to rhythms or to rhythm-relevant immediate early gene expression, and no light-regulated genes are known in Drosophila. But these differences are probably temporary, indicating that Cry might well connect to transcription in both systems. As Cry autoregulates expression (Stanewsky, 1998), it is even possible that it participates directly, or once participated directly, in its own transcriptional regulation. Photoreception and transcriptional feedback loops are ubiquitous features of circadian pacemakers. Therefore, circadian rhythms may have begun long ago when a DNA repair protein acquired the ability to autoregulate expression in a light- and ultimately time-dependent manner (Emery, 1998 and references).

cry^b is an apparent null mutation in a gene encoding Drosophila's version of the blue light receptor Cryptochrome. The mutation maps to a position approximately in the middle of the right arm of chromosome 3. Flies heterozygous for the mutation exhibit poor bioluminescence rhythms. Although a higher proportion of mutation-over-deletion males is rhythmic, as compared to homozygous mutant flies, formal analysis of these rhythms indicates that they are very weak and exhibit anomalous phases; mutation-over-deletion females are thoroughly arrhythmic (Stanewsky, 1998).

Given the drastic effects of this mutation on cyclic expression of Per and Tim proteins (see cryptochrome Targets of Activity), it was expected that the locomotor behavior of cry^b would also be abnormal. Surprisingly, cry^b flies are rhythmic in both LD and DD conditions, given ~24 hr periods. In further LD experiments, flies were first exposed to 5 days of 12 hr:12 hr LD (white light, intensity 640 lux), followed by a second 5 day LD regime in which the lights came on 4 hr later and were changed to dim blue. cry^b flies entrain to the initial LD cycles, since all the mutant individuals exhibit characteristic bimodal activity peaks. These mutant flies also have no trouble reentraining to the second light regimes, even though a much lower light intensity was used. The same behavior is observed for cry⁺ and cry^b flies at all light intensities (Stanewsky, 1998).

Externally blind flies behaviorally entrain to LD cycles but are less sensitive to the synchronizing effects of light in terms of their behavioral rhythmicity. Hence, the light responsiveness of norpA^P41;cry^b double mutants was tested. norpA codes for a phospholipase C (PLC) activity involved in the phototransduction cascade that takes place within the eye. The norpA loss-of-function mutation causes the compound eyes and ocelli to be completely unresponsive to light (Pearn, 1996). The doubly mutant flies exhibit entrainment problems in the initial LD regime. Subsequently, only about 50% of the flies that entrained to the initial LD cycle are able to synchronize to a new light regime that applies 16-lux blue light; and at still lower light intensities, nearly all of the double mutants fail to entrain to the new LD cycles. Since norpA and cry^b mutations by themselves cause minimal or no entrainment problems at these low light intensities, the synergistic effect observed in the double mutant suggests that there are two light entrainment pathways operating in Drosophila's rhythm system. A fast clock period mutation also uncovers entrainment defects in cry^b mutants (Stanewsky, 1998).

The behavioral effects of cry^b indicate that the normal function of the gene is involved in light-mediated entrainment and the ability to reset the clock pacemaker that underlies Drosophila's daily rhythms of locomotion. Yet the protein encoded by cry is not the only photoreceptor involved in this process: rhodopsin molecules function upstream of the NorpA PLC protein in the canonical phototransduction cascade that is a part of the input pathway to the fly's clock. PLC could be functioning within the anatomical elements of such pathways located in the CNS as well as within the external eyes, because norpA's PLC product is found within the brain as well as in the eyes. This PLC could even act downstream of cry in extraocular locations, but if this were the case, a norpA mutation would be sufficient to block entrainment in medium-dim light, and this is not so. Therefore, the combined effects of norpA and cry^b mutations on entrainment are likely to be the result of these two functions acting separately in the eye and in the CNS, respectively. This means that there are anatomically independent light input pathways to the pacemaker that control behavior. One of these might involve brain neurons that feed light into the clock. Alternatively, the extraocular pathway could function intracellularly within clock neurons themselves, by analogy to pinealocytes in lower vertebrates being both circadian pacemaker cells and photoreceptors. But the fact that anatomical eyelessness causes the same decrement in circadian light sensitivity as do norpA mutations suggests that the latter's effects are acting only through external photoreceptors. In this scenario, norpA would participate merely as part of an eye-to-brain throughput pathway, via the compound eyes and optic ganglia, eventually reaching the CNS pacemaker. The fact that Cry is not the only important input factor in the fly's circadian system fits with results from other systems, ranging from microbes to mammals. Multiple input pathways operate with regard to both anatomical structures and physiological processes. The adaptive value of such complexity is that it permits entraining organisms to sample light of several different wavelengths. Indeed, the quality of light reaching the earth's surface changes dramatically from dawn and dusk, in particular being enriched in the blue wavelengths during the dimmest portions of these daily transition times (Stanewsky, 1998 and references)

GENE STRUCTURE

PROTEIN STRUCTURE

Amino Acids - 542

Structural Domains

A BLAST search of the Drosophila EST sequence database reveals the presence of an expressed sequence tag showing high homology with Drosophila 6-4 photolyase (Todo, 1996); there is also a much more distantly related Drosophila cyclobutane pyrimidine dimers photolyase (Yasui, 1994). This second gene was named cry. Cry is more closely related to the two human cryptochromes HsCRY1 and HsCRY2 than to Dm64, suggesting that Cry and the human proteins might share functional properties different from those of Dm64. But an unrooted phylogenetic tree reveals that HsCRY1, HsCRY2, Dm64, and even Arabidopsis 6-4 photolyases are more closely related to one another than to Cry. The blue light photoreceptors of Arabidopsis, including the protein responsible for flowering time photoperiodism, are much more distantly related to Cry. Taken together, the sequence relationships do not convincingly indicate a specific biochemical function for Drosophila Cry (Emery, 1998).

The C-terminal half of the protein shows the highest conservation. This region of E. coli photolyase contains most of the amino acids involved in binding the two cofactors, MTHF (5,10-methenyltetrahydrofolate) and FAD (flavin-adenine dinucleotide; Park, 1995 and Kanai, 1997). The MTHF-binding sites are not very well conserved between Cry and its close relative. Likewise, these residues are not very strongly conserved, even among family members known to bind MTHF. In contrast, the FAD-binding site is highly conserved (12/14 homologous amino acids, 8/14 identical). This is despite the presence of a lysine at conserved position 264, which is either a serine or threonine in all other members of the photolyase/cryptochrome family. Although the crystal structure of E. coli photolyase did not contain nucleic acid (Park, 1995), other experiments indicated that the same C-terminal half of the protein interacts with DNA (Baer, 1993). The putative DNA-binding residues are also well conserved in Drosophila Cry (Park, 1995 and Kanai, 1997).

Cryptochromes are members of a large protein family that includes blue light photoreceptors, 6-4 photolyases (DNA photoreactivating enzymes), and microbial Class I CPD (cyclobutane pyrimidine dimers) photolyases. DCry is most similar to the human cryptochromes, with similarities of 46% and 45% to hCRY1 and hCRY2, respectively. DCry is more similar to mammalian cryptochromes than plant cryptochromes or Class I CPD photolyases, the latter exhibiting similarities ranging from 23% to 30%. The generation of a phylogram for 12 different cryptochromes and photolyases underscores the relatedness of DCry to the hCRY1 and hCRY2 proteins, although DCry is also similarly related to both the Drosophila and Arabidopsis 6-4 photolyases. There is little or no homology among DCry and higher eukaryotic Class II CPD photolyases. Therefore, DCry can be considered a new member of the protein subfamily that includes mammalian cryptochromes and 6-4 photolyases. The aggregate of these comparisons is consistent with the idea that DCry functions as a blue light photoreceptor rather than a photolyase. This conclusion is supported by the functional analysis, which indicates that the DCry protein mediates blue light resetting of the Drosophila circadian clock. Similar to other cryptochromes, DCry has conserved domains that include residues known to be important for the noncovalent binding of the cofactors pterin (folate) and flavin adenine dinucleotide. In plant and mammalian cryptochromes, these cofactors determine light absorption spectra. Also like other cryptochromes, the DCry protein has a nonconserved C terminus (of 41 residues) that is completely unrelated to photolyases, cryptochromes, or any other sequence in the protein databases. It has been suggested that the novel C termini of cryptochromes might be important for interactions with effector molecules, and it has been demonstrated that the C terminus of the human CRY2 protein can physically interact with and inhibit the phosphatase activity of the tetratricopeptide repeat (TPR)-containing protein PP5. The DCry protein does not contain an N-terminal extension similar to those of microbial and plant photolyases that have been implicated in mitochondrial and nuclear targeting (Egan, 1999 and references).

date revised: 27 December 98

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.