Report

Light-Independent Role of CRY1 and CRY2 in the Mammalian Circadian Clock

See allHide authors and affiliations

Science  22 Oct 1999:
Vol. 286, Issue 5440, pp. 768-771
DOI: 10.1126/science.286.5440.768

Abstract

Cryptochrome (CRY), a photoreceptor for the circadian clock in Drosophila, binds to the clock component TIM in a light-dependent fashion and blocks its function. In mammals, genetic evidence suggests a role for CRYs within the clock, distinct from hypothetical photoreceptor functions. Mammalian CRY1 and CRY2 are here shown to act as light-independent inhibitors of CLOCK-BMAL1, the activator driving Per1 transcription. CRY1 or CRY2 (or both) showed light-independent interactions with CLOCK and BMAL1, as well as with PER1, PER2, and TIM. Thus, mammalian CRYs act as light-independent components of the circadian clock and probably regulate Per1transcriptional cycling by contacting both the activator and its feedback inhibitors.

Daily rhythms in physiology and behavior are driven by endogenous oscillators called circadian clocks (1). In all known cases, circadian timekeeping is cell-autonomous, generated at least in part by a feedback loop involving clock proteins that inhibit the transcription of their own genes (2). Regulation of the transcriptional feedback loop by light is thought to mediate setting of circadian clocks to light-dark cycles (2). In mammals, as inDrosophila, a negative feedback loop of Pergene transcription involving PER and TIM proteins is probably central to the clock (2–4). A heterodimeric activator consisting of the basic helix-loop-helix (bHLH)–PAS proteins CLOCK (5) and BMAL1 (6,7) drives mouse Per1 (mPer1) transcription from E-box regulatory sequences (6), and the mPER1 protein in turn acts to inhibit CLOCK-BMAL1 activity (4). In Drosophila, a homologous activator, CLOCK-CYC(dBMAL), plays a similar role (8).

Genetic studies have implicated cryptochromes (CRYs) (9) in light-dependent signaling to circadian clocks in plants (10) and Drosophila (11). Drosophila CRY acts directly as a circadian photoreceptor: it exhibits light-dependent binding to TIM, and it blocks the negative feedback action of the PER-TIM complex on CLOCK-CYC(dBMAL) in a light-dependent manner (12). Targeted disruption of either of the two identified mouse Cry genes results in alterations of circadian period (13, 14), and disruption of both together results in arrhythmicity (14). These results suggest that mammalian CRYs could play a central role in the circadian clock independent of light, leaving debatable the question of whether or not they might have additional photoreceptor functions (15).

To test the hypothesis that mammalian CRYs have light-independent actions in the circadian clock, we examined the activity of human CRY1 and CRY2 (hCRYs) (16) in our previously described assay in which CLOCK-BMAL1 drives a luciferase reporter gene from mPer1 gene E boxes in cultured cells (4, 6). Figure 1A shows results from experiments in which the cells were kept in darkness from the time of transfection until harvesting (17). As we previously reported (4), mPER1 produced a partial but significant inhibition of CLOCK-BMAL1 activity, which is indicative of its role in negative feedback. Both hCRY1 and hCRY2 strongly inhibited CLOCK-BMAL1 activity, with hCRY1 consistently showing a somewhat more potent inhibition than hCRY2. No detectable effect on this strong inhibition by hCRY1 or hCRY2 was produced by coexpression of mPER1 (18). Like mPER1 (4), hCRY1 and hCRY2 produced no significant inhibition of MYOD-E12 activity, a heterodimeric bHLH factor also acting from E boxes, whereas the known inhibitor of MYOD-E12, the ID protein, inhibited strongly (Fig. 1B). These results indicate that hCRY1 and hCRY2 potently and specifically inhibit mPer1 gene transcription by CLOCK-BMAL1 in the absence of light, and they suggest that a role for CRYs in the mammalian circadian clock is to inhibitPer1 gene expression.

Figure 1

Potent and specific inhibition by hCRY1 and hCRY2 of mPer1 gene activation by the CLOCK-BMAL1 heterodimer in the absence of light. Transcriptional activation in cultured mammalian cells of a luciferase reporter gene from E-box sequences derived from the mPer1 gene (A) or the muscle creatine kinase (mck) gene (B) is shown (4). (A) Effect of mPER1, hCRY1, or hCRY2 on transactivation by the CLOCK-BMAL1 heterodimer. (B) Effect of ID protein, hCRY1, or hCRY2 on transactivation by the MYOD-E12 heterodimer. A plus or minus sign denotes an expression plasmid with or without, respectively, the indicated full-length cDNA insert. Cells were kept in constant darkness from the time of transfection until harvesting, except for a brief exposure to a dim red safelight at the time of medium change (17). Shown are the mean and SEM of three to six independent transfections.

To determine whether light could modulate hCRY activity, we next performed transcriptional reporter assays as above in paired sets of cells that had been kept in constant darkness or constant light (19). No systematic or substantive differences were observed in the ability of hCRY1 or hCRY2 to inhibit CLOCK-BMAL1 in light and dark conditions (Fig. 2A). Because inhibition of CLOCK-BMAL1 activity by hCRYs is so potent in our standard assay, it was possible that we missed a modulatory effect of light. We therefore carried out experiments in which the amount ofhCry cDNA expression plasmid in each transfection reaction was systematically decreased over several orders of magnitude, all else being held constant (17). The curves for inhibition of CLOCK-BMAL1 transactivation as a function of the amount of hCry transfected were superimposable for dark and light conditions (Fig. 2B). The slightly greater potency of hCRY1 than hCRY2 observed under our standard conditions (Figs. 1A and2A) was easily discernable as a large displacement on the xaxis between the hCRY1 and hCRY2 curves (Fig. 2B). Thus, not only is light unnecessary for hCRY1 and hCRY2 inhibition of CLOCK-BMAL1 transcriptional activity, it has no detectable modulatory effect over the full dynamic range of hCRY1 and hCRY2 action.

Figure 2

Inhibition of CLOCK-BMAL1 activity by hCRY1 and hCRY2 under constant dark and constant light conditions. (A) Luciferase reporter assays were done as in Fig. 1A, except that duplicate sets of transfected cells were kept in constant darkness or constant light until harvesting (19). Shown are the mean and SEM of three to six independent transfections. (B) Luciferase reporter assays were done as in (A), except that the amounts of hCry expression plasmid in the transfections were systematically reduced starting with the standard amount as in (A), all else being held constant (17). Each point reflects the mean of three to six independent transfections; SEM values, comparable to those in Figs. 1A and 2A, were omitted for the sake of visual clarity.

Could hCRY inhibition of CLOCK-BMAL1 activity in cultured cells reflect a direct interaction with CLOCK or BMAL1 (or both) or an interaction with endogenous TIM or PER proteins (or both), which are known feedback inhibitors of CLOCK-BMAL1 activity? To test whether such interactions occur, we used a yeast two-hybrid assay (20) in which yeast transformants were grown in duplicate sets under constant light or darkness (12). As we recently reported (12),Drosophila CRY interacts with TIM in a light-dependent manner (Fig. 3A), whereas no detectable interaction was observed between CRY and dCLOCK, CYC(dBMAL), or PER under dark or light conditions (12, 18). In contrast, hCRY1 and hCRY2 showed specific interactions with mammalian clock components that were independent of light (Fig. 3B). hCRY1 produced strong interaction signals with both mPER2 and BMAL1, a modest but reproducible interaction signal with mPER1, and no interaction signal above background with CLOCK or hTIM. hCRY2 was similar to hCRY1 in that it produced a strong interaction signal with mPER2, but different in that it produced weak but reproducible interaction signals with CLOCK and hTIM and no detectable interaction signal with mPER1 or BMAL1. These results strongly suggest that hCRY1 and hCRY2 inhibit the CLOCK-BMAL1 heterodimer in mammalian cells by forming direct contacts with it, possibly within a multiprotein complex including PER and TIM proteins.

Figure 3

Light-independent interactions of hCRY1 and hCRY2 with mammalian clock components. Yeast two-hybrid filter-lift assays were performed as described (20). Blue precipitate reports cumulative β-galactosidase activity, resulting from transactivation of the lacZ reporter gene by protein-protein interaction. Triplicate yeast patches expressing the indicated LEXA hybrid (rows) and the indicated VP16 hybrid (columns) were derived from three independent transformants. All results were verified with at least three additional independent transformants generated with new preparations of each expression plasmid. All hybrid proteins were full length, except for VP16-CLOCK, which was CLOCK(1-389) in this experiment. hTIM, human TIM. “Dark” or “light” indicates yeast grown in constant darkness or light from the time of transformation with plasmid DNA (12). (A) Control assays showing light-dependent interaction of Drosophila CRY (dCRY) with Drosophila TIM (dTIM), as reported (12). (B) Light-independent interactions of hCRY1 and hCRY2 with other clock components. Yeast patches showing hCRY1 interactions were grown on the same plate and assayed on the same filter as theDrosophila controls shown in (A). Pink discoloration of some patches in bottom panels is due to the Ade 2 mutation in the L40 strain.

In our functional assays, mammalian CRYs, PERs, and TIM all inhibitmPer1 promoter activation by CLOCK-BMAL1. The ability of CRYs to bind both to the activator (CLOCK and BMAL1) and to other inhibitors (PER2, PER1, and TIM) suggests that CRYs could regulate the actions of the other inhibitors. To test this possibility, we carried out transcriptional reporter assays designed so that CLOCK-BMAL1 activity would be inhibited only partially by hCRY1 or hCRY2. We then looked for synergistic or antagonistic effects of hCRYs on the inhibitory actions of hTIM or mPER2. The typical partial inhibition we observe with hTIM was significantly antagonized by either hCRY1 or hCRY2, whereas the comparable partial inhibition produced by mPER2 was not significantly affected by hCRY1 or hCRY2 (Fig. 4). Furthermore, coexpression of hTIM together with mPER2 produced the same partial inhibition as either alone (18), as we previously reported for hTIM together with mPER1 (4). These results indicate a specific functional antagonism between CRYs and TIM and suggest cross-regulation among the proteins inhibiting CLOCK-BMAL1 activity within the circadian clock feedback loop.

Figure 4

Functional antagonism between hCRY1 or hCRY2 and hTIM. Luciferase reporter assays in cells kept under dark conditions, as in Fig. 1A, are shown. On the basis of the titrations shown in Fig. 2B, the amounts of hCry1 and hCry2expression plasmids transfected were chosen to give approximately 20% inhibition of CLOCK-BMAL1 activity.

We conclude that CRY1 and CRY2 are circadian clock components that negatively regulate Per1 transcription independently of light. Our work indicates that they inhibit CLOCK-BMAL1 transactivation activity, very likely by means of direct contacts between CRY1 and BMAL1 and between CRY2 and CLOCK. We have established that CRY1 and CRY2 can interact with PER2 and, respectively, with PER1 and TIM. In addition, our experiments have revealed a functional antagonism between the seemingly similar actions of CRYs and TIM. Together these results indicate that CRY1 and CRY2 occupy a central position within the mammalian circadian clock, virtually at the interface between CLOCK-BMAL1, the transcriptional activator of the mPer1 gene, and PER1, PER2, and TIM, its feedback inhibitors. The recent finding that mPer1transcripts are constitutively expressed in the suprachiasmatic nuclei of Cry1, Cry2 double mutant mice at a high level, comparable to the peak in wild-type mice, provides strong in vivo support for our conclusions (21).

A role for CRY1 and CRY2 in mammalian circadian clock negative feedback has been reported independently (22), although in that study there was no examination of the effect of light or dark on CRY1 or CRY2 function. The distinct patterns of interaction of CRY1 and CRY2 with PER1, PER2, and TIM revealed in our two-hybrid assays are consistent with the reported ability of CRY1 or CRY2 to coimmunoprecipitate with PER1, PER2, PER3, or TIM from transfected mammalian cells (22). Because of its high sensitivity, coimmunoprecipitation from cellular extracts might detect very low-affinity interactions not detected in two-hybrid assays. Alternatively, coimmunoprecipitation could reflect the presence of two noninteracting proteins within a multiprotein complex containing additional endogenous PER or TIM proteins, which are known to be expressed in various mammalian cell lines (18,22, 23). As shown here, the distinct protein interaction profiles of CRY1 and CRY2 could underlie the different circadian phenotypes observed in Cry1 and Cry2mutant mice.

The light-independent role of mammalian CRYs in circadian clock negative feedback contrasts sharply with that of DrosophilaCRY, which has been demonstrated by us and our collaborators to function directly as a photoreceptor that regulates the action of the PER-TIM complex (12). We suggest that DrosophilaCRY exemplifies the ancestral role of a photoreceptor acting as a light-dependent regulator of the circadian feedback loop, whereas mammalian CRYs have preserved a role within the circadian feedback loop but shed their direct photoreceptor function. We cannot, however, exclude the possibility that mammalian CRYs act as photoreceptors for other possible functions, circadian or otherwise, not detected in our assays.

  • * To whom correspondence should be addressed. E-mail: cweitz{at}hms.harvard.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article