Photic Induction of mPer1 and mPer2 in Cry-Deficient Mice Lacking a Biological Clock

See allHide authors and affiliations

Science  24 Dec 1999:
Vol. 286, Issue 5449, pp. 2531-2534
DOI: 10.1126/science.286.5449.2531


Mice lacking mCry1 and mCry2 are behaviorally arrhythmic. As shown here, cyclic expression of the clock genesmPer1 and mPer2 (mammalian Periodgenes 1 and 2) in the suprachiasmatic nucleus and peripheral tissues is abolished and mPer1 and mPer2 mRNA levels are constitutively high. These findings indicate that the biological clock is eliminated in the absence of both mCRY1 and mCRY2 (mammalian cryptochromes 1 and 2) and support the idea that mammalian CRY proteins act in the negative limb of the circadian feedback loop. ThemCry double-mutant mice retain the ability to havemPer1 and mPer2 expression induced by a brief light stimulus known to phase-shift the biological clock in wild-type animals. Thus, mCRY1 and mCRY2 are dispensable for light-induced phase shifting of the biological clock.

Many physiological and behavioral systems are controlled by an internal self-sustaining molecular oscillating mechanism with a periodicity of approximately 24 hours, known as the biological clock. The core oscillator consists of an autoregulatory transcription-(post) translation–based feedback loop involving a set of clock genes (1). In mammals, as inDrosophila, three recently identified mPer genes (mPer1, mPer2, and mPer3) are thought to be oscillator genes (2–4). Transcription of mPer1 is driven by the CLOCK/BMAL transcriptional activator complex and in turn is repressed by its own gene product (5). To maintain synchrony with the solar day/night cycle, the master clock in the suprachiasmatic nucleus (SCN) of the brain needs to be reset by daily light through receptors in the eye (6). In mammals, mCRY1 and mCRY2, members of the light-harvesting cryptochrome/photolyase protein family (7), have been proposed as candidate photoreceptors required for light-entrainment of the biological clock (8). Mouse mutants lacking mCry1 show an acceleration of the free-running clock, as measured by wheel-running activity in constant darkness (9), whereas loss of mCry2 slowed the clock (9, 10). Unexpectedly, mice lacking bothmCry genes completely lack free-running rhythmicity (9). Although these results point to an antagonistic clock-adjusting function as well as to an absolute requirement of mCRY proteins for maintenance of circadian rhythmicity, they failed to resolve a possible function of the proteins as circadian photoreceptors and did not elucidate the consequences of mCRY dysfunction at the molecular level. To investigate how mCRY proteins act in the circadian core oscillation mechanism, we examined temporal and light-induced expression profiles of mCry1, mCry2,mPer1, and mPer2 in the SCN and peripheral tissues of wild-type and mCry mutant mice by quantitative in situ hybridization or quantitative polymerase chain reaction (PCR) using real-time TaqMan technology.

Consistent with a possible function as clock gene, mRNA levels ofmCry1 oscillate in a circadian manner in the SCN and skeletal muscle of wild-type mice (8, 11). Similarly,mCry2 expression exhibits a daily rhythm in the skeletal muscle (11). In contrast to previous reports (8,11), using a highly sensitive method (12, 13), we observed that mCry2 is rhythmically expressed in the SCN in a manner similar to mCry1 (Fig. 1, A and B) (14). Levels of mCry2 mRNA peak at the (subjective) day/night transition whether mice were kept under 12-hour light and dark cycles (LD 12:12) or under constant darkness (DD). Expression of the mCry genes is not significantly up-regulated by a brief light pulse given during the subjective day (CT8) or night (CT20) (Fig. 1, C and D). This indicates that, unlike mPer1 andmPer2 (3, 15), mCry expression is not directly regulated by light.

Figure 1

Expression of mCry1 andmCry2 in the suprachiasmatic nucleus (SCN) of wild-type mice. (A and B) Circadian expression of (A) mCry1and (B) mCry2 mRNA in the SCN in LD 12:12 (open circles) and in DD (closed circles). For eachmCry gene, the relative RNA abundance was determined by quantitative in situ hybridization, with the value at CT12 adjusted to 100%. Values are expressed as means ± SEM (n = 5). Representative autoradiograms for each time point are shown below the graphs. (C and D) Light-pulse experiments for (C)mCry1 and (D) mCry2. A 30-min light pulse (600 lux; fluorescent light) delivered at CT8 (left panels) and at CT20 (right panels) does not induce mCry1 andmCry2 mRNA. The relative RNA level was measured by quantitative in situ hybridization with the value just before onset of light adjusted to 100%. Values are expressed as means ± SEM (n = 3).

Next, we examined the expression profiles of mPer1 andmPer2 in entrained wild-type and Cry mutant mice (LD 12:12; 2 weeks) starting 36 hours after animals were placed in constant darkness (DD). Both mPer1 andmPer2 show a high-amplitude oscillation in the SCN of wild-type animals (Fig. 2, A and B). As expected on the basis of animals' circadian wheel-running behavior (9), robust oscillation of mPer1 andmPer2 mRNA is maintained in animals mutant forCry1 or Cry2. On the absolute time scale, consistent with the differences in free-running period length (9), mPer1 and mPer2 expression peaks earlier in mCry1 mutant mice (tau = 22.5 hours) and later in mCry2 mutant mice (tau = 24.6 hours), compared to wild-type animals (tau = 23.8 hours). In marked contrast,mCry double-mutant mice fail to show significantmPer1 and mPer2 mRNA cycling in the SCN (Fig. 2, A and B); instead, mPer1 and mPer2 transcript levels are intermediate to high at all times examined. In addition, we observed a homogeneous pattern of mPer1 labeling in SCN tissue sections of mice deficient in mCry1 andmCry2 (Fig. 2C), indicating that transcript levels ofmPer1 are similar in all cells. This suggests that the absence of mPer1 cycling in double mutants (Fig. 2A) is not due to a synchronization defect.

Figure 2

Expression of mPer1 andmPer2 in the SCN of wild-type and mCry mutant mice in constant darkness. (A and B) The relative RNA abundance ofmPer1 (A) and mPer2 (B) in the SCN of wild-type (open circles), mCry1 –/–(open triangles), mCry2 /– (open squares), andmCry1 –/– mCry2 /–(closed circles) mice in constant darkness was determined by quantitative in situ hybridization. The peak values observed in wild-type animals were adjusted to 100%. Values are expressed as means ± SEM (n = 4). Representative autoradiograms are shown below the graphs. BecausemCry1/mCry2 double-mutant mice are arrhythmic, we used an environmental time scale (hours in DD) rather than a circadian time (CT) scale. (C) Representative examples ofmPer1 mRNA signals (dotted grains) in emulsion-coated sections of the SCN in wild-type andmCry1 /– mCry2 /–mice. (Left panel) wild-type SCN, 40 hours in darkness (≈CT4; 68 to 80% cells with 10 to 25 grains per cell); (middle panel) wild-type SCN, 52 hours in darkness (≈CT16; <10% cells with 10 to 25 grains per cell); (right panel)mCry1 −/− mCry2 −/− SCN, 52 hours in darkness (62 to 74% cells with 10 to 25 grains per cell). Bar, 10 μm.

Because non-SCN tissues also rhythmically express clock genes (4,16), we examined whether oscillation of these putative peripheral clocks was also affected in the absence of CRY. In contrast to wild-type mice, in the retina (Fig. 3A) (17) and liver (Fig. 3B) (18) of mCry1/mCry2 double-mutant mice,mPer1 mRNA no longer cycles and expression is constantly high. Similarly, mPer2 mRNA did not show a prominent rhythm in retina and liver of mCry1/mCry2 double-mutant mice (19). From the total absence of mPer1 andmPer2 oscillation in the SCN and in peripheral tissues, we conclude that the behavioral arrhythmicity observed inmCry1/mCry2-deficient mice is the direct consequence of a complete impairment of the molecular clock and that mCRY proteins are indispensable components of the core oscillator. Furthermore, the presence of accelerated and retarded clock oscillation inmCry1 and mCry2 single mutants indicates at the molecular level that mCRY1 and mCRY2 have overlapping functions in running the clock and an antagonistic role in determining its pace. The high mRNA levels of mPer1 and mPer2 inmCry1/mCry2-deficient mice suggest that mCRY proteins negatively affect mPer expression. Reppert and co-workers (11), as well as other groups (20, 21), have shown that whereas mPER and mTIM proteins only have moderate inhibitory effects on CLOCK/BMAL transactivating activity, mCRY1 and mCRY2 can almost completely block transcription from the mPer1promoter, which puts the mCRY proteins directly in the negative limb of the circadian feedback loop (11). Our findings that, in the absence of mCRY, mPer1 and mPer2 appear constitutively expressed at high levels are in agreement with this conclusion.

Figure 3

Expression of mPer1 in the retina and liver of wild-type andmCry1 −/− mCry2 −/−mutant mice in constant darkness. (A) The relative abundance of mPer1 mRNA in the retina was determined by quantitative in situ hybridization (five sections per animal; three independent experiments). Values (CT48 was determined as 100%) are expressed as the mean ± SEM. (B) The relative abundance ofmPer1 mRNA in the liver was determined by the reverse transcription PCR method with TaqMan technology as described. The calculated expression level of mPer1 of wild-type at 44 hours divided by that of GAPDH was normalized to 100%. Values are expressed as means ± SEM (n = 3).

Circadian photoreceptors in mammals reside in the eye: loss of eyes abolishes photo-entraining of the biological clock (22). However, rod- and cone-less mice show intact entrainment, indicating that rhodopsins are not involved (23). mCRY proteins are likely candidates for the circadian photoreceptors involved in light-entrainment because (i) they belong to the family of blue-light receptors (7), (ii) their expression is high in the ganglion cells and inner nuclear layers of the retina (8), and (iii) they can exert a clock-adjusting function (9) which would be required in the context of effecting a light input into the clock. However, direct proof for the presumed light input function is lacking because, ironically, the absence of a functional clock inmCry double-mutant mice precludes analysis of a resetting function by classical phase-shifting experiments. Nevertheless, to investigate a potential mCRY-mediated photoreceptor role, we examined the effect of light pulses on mPer1 and mPer2expression: when light is given early during the subjective night concomitant with phase shifting of the clock, wild-type mice respond with an acute induction of mPer1 and mPer2 mRNA (Fig. 4) (3, 14, 24). Surprisingly, mCry double-mutant mice retained the ability to respond to the brief light pulse with acute induction ofmPer1 and mPer2 mRNA (Fig. 4). This indicates that the rapid induction of mPer1 and mPer2 mRNA by light is still intact in complete absence of mCRY proteins and can occur without a functional clock or core oscillator.

Figure 4

Light-induced mPer1 andmPer2 expression in the SCN of wild-type andmCry1 /− mCry2 −/−mutant mice. A 30-min light pulse (600 lux; fluorescent light) delivered 52 hours after the dark transfer induced (A)mPer1 and (B) mPer2 mRNA in the SCN of wild-type and mCry1 −/− mCry2 −/− mutant mice. The relative abundance of mPer1 mRNA was determined by quantitative in situ hybridization, with the peak induced values of wild-type (1 hour formPer1, 1.5 hour for mPer2) adjusted to 100%. Values are expressed as means ± SEM (n = 4). Representative autoradiograms are shown in the lower panels.

In conclusion, our results show that a complete inactivation of the molecular oscillator driving the biological clock underlies the arrhythmicity observed in mCry1/mCry2 double-mutant mice. Thus, mCRY1 and mCRY2 are indispensable components of the core oscillator, and mCry1/mCry2-deficient mice are, to our knowledge, the first mammals completely lacking a circadian pacemaker. In mCry1/mCry2 double-mutant mice, the mechanism of light-induced phase shifting by up-regulation of mPer1 andmPer2 expression is still intact, which points to the involvement of different transcription regulatory processes for core oscillation and light-mediated phase-shifting. Our data suggest that photoreceptors other than mCRY proteins and rod/cone opsins [for example, mammalian homologs of the recently discovered fish VA-opsin (25) or Xenopus laevis melanopsin (26)] may be responsible for photic entrainment. Although mPer transcription repression by mCRY proteins (and thus their function in core oscillation) is light-independent (20), we do not completely rule out that mCRY proteins may act as photoreceptor proteins. First, it is possible that phase-shifting is mediated via more than one photoreceptor system, implying functional redundancy. Second, mCRY proteins may be involved in transmitting light inputs to the clock other than those required for phase shifting, such as changes in the length of the day and night, and information on dusk and dawn. Future experiments should shed light on the mysterious blue-light receptor properties of mCRY proteins.

Note added in proof: Recently, Vitaterna and colleagues have determined mRNA levels of mPer1 and mPer2 inmCry1/mCry2 double-mutant mice at two time points. In addition, photic induction (measured 30 min after a 1-hour light pulse) was observed for mPer2 but not for mPer1(27).

  • * To whom correspondence should be addressed. E-mail: okamurah{at}


Stay Connected to Science

Navigate This Article