© 1999 Macmillan Magazines Ltd letters to nature NATURE | VOL 398 | 15 APRIL 1999 | www.nature.com 627 25. Jimenez, F. et al. vnd, a gene required for early neurogenesis of Drosophila, encodes a homeodomain protein. EMBO J. 14, 3487±3495 (1995). 26. McDonald, J. A. et al. Dorso-ventral patterning in the Drosophila CNS: The vnd homeobox gene speci®es ventral column identity. Genes Dev. 12, 3603±3612 (1998). 27. MacDonald, R. et al. The Pax protein Noi is required for commissural axon pathway formation in the rostral forebrain. Development 121, 2397±2408 (1995). 28. Jensen, J., Serup, P., Karlsen, C., Nielsen,T. F. & Madsen, O. D. mDNApro®ling of rat islet tumors reveals nkx 6.1 as a beta-cell-speci®c homeodomain transcription factor. J. Biol. Chem. 271, 18749± 18758 (1996). 29. Geisert, E. E. & Frankfurter, A. The neuronal response to injury as visualized by immunostaining of class III beta-tubulin in the rat. Neurosci. Lett. 102, 137±141 (1989). 30. Pattyn, A., Morin, X., Cremer, H., Coridis, C. & Brunet, J. F. Expression and interactions of the two closely related homeobox genes Phox2a and Phox2b during neurogenesis. Development 124, 4065± 4075 (1997). Acknowledgements. We thank P. Rashbass and V. van Heyningen for Sey mice; D. Anderson, J. F. Brunet, H. Edlund, C. Goridis, A. Frankfurter, C. William and S. Wilson for reagents; C. Doe for results on vnd; R. Axel, K. Lee and G. Struhl for comments on the manuscript; and K. MacArthur for help in preparating the manuscript. This work was supported by grants to T.M.J. from the NIH and to J.L.R.R. from Nina Ireland, NARSAD and NIMH; by an HFSP fellowship (J.B.); by the Swedish Institute, the Swedish Foundation for Strategic Research, and the Swedish National Research Council (J.E.); by the Bank of America, Scottish Rite and NIH, NRSA (L.S.); and by the Danish National Research Foundation (P.S.). J.B. and J.E. are associates and T.M.J. is an investigator of the Howardhughes Medical Institute. Correspondence and requests for materials should be addressed to J.L.R.R. (e-mail: jlrr@cgl.ucsf.edu) or T.M.J. (e-mail: tmj1@columbia.edu). Mammalian Cry1 and Cry2 are essential for maintenance of circadian rhythms Gijsbertus T. J. van der Horst*, Manja Muijtjens*, Kumiko Kobayashi², Riya Takano², Shin-ichiro Kanno², Masashi Takao², Jan de Wit*, Anton Verkerk*, Andre P. M. Eker*, Dik van Leenen³, Ruud Buijs§, Dirk Bootsma*, Jan H. J. Hoeijmakers* & Akira Yasui² * MGC, Department of Cell Biology and Genetics, Erasmus University, PO Box 1738, 3000 DR Rotterdam, The Netherlands ² Department of Molecular Genetics, Institute of Development, Aging and Cancer, Tohoku University, 980-8575 Sendai, Japan ³ MGC, Department of Clinical Genetics, Erasmus University, PO Box 1738, 3000 DR Rotterdam, The Netherlands § Netherlands Institute for Brain Research, Meibergdreef 33, 1105 AZ Amsterdam, The Netherlands ......................................................................................................................... Many biochemical, physiological and behavioural processes show circadian rhythms which are generated by an internal time- keeping mechanism referred to as the biological clock. According to rapidly developing models, the core oscillator driving this clock is composed of an autoregulatory transcription±(post) translation-based feedback loop involving a set of `clock' genes 1±6 . Molecular clocks do not oscillate with an exact 24-hour rhythmicity but are entrained to solar day/night rhythms by light. The mammalian proteins Cry1 and Cry2, which are members of the family of plant blue-light receptors (cryptochromes) and photolyases, have been proposed as candidate light receptors for photoentrainment of the biological clock 7±10 . Here we show that mice lacking the Cry1 or Cry2 protein display accelerated and delayed free-running periodicity of locomotor activity, respec- tively. Strikingly, in the absence of both proteins, an instanta- neous and complete loss of free-running rhythmicity is observed. This suggests that, in addition to a possible photoreceptor and antagonistic clock-adjusting function, both proteins are essential for the maintenance of circadian rhythmicity. We were interested in identifying mammalian homologues of the DNA-repair enzyme photolyase, a protein that undoes ultraviolet- induced DNA damage in a single-step process (photoreactivation) requiring light energy captured by blue-light-collecting chromophores 11,12 . In this search, we and others have cloned two genes with strong homology to class I photolyases of lower spe- cies 7±10 . In addition to the photolyase core domain, the gene products appeared to contain a carboxy-terminal extension also found in plant blue-light receptors (cryptochromes), for which the mammalian photolyase-like genes were designated cry1 and cry2. Plant Cry proteins mediate light-dependent processes such as phototropism, growth and ¯owering 13±15 . Since placental mammals as well as endogenous or recombinant mammalian Cry proteins lack clearly detectable photoreactivating activity, the mammalian Cry proteins may act as photoreceptors rather than photolyases 7,9,10 . The biological `master' clock in the suprachiasmatic nucleus (SCN) of the brain controls many physiological processes, from body temperature to the sleep±wake cycle. A major question in mamma- lian chronobiology is how the clock is entrained to solar time, thereby keeping an organism in an exact 24-h rhythm. The absence of photoentrainment in eye-less rodents indicates that the light receptors feeding into the SCN circadian system must reside in the eye 16,17 , but the process does not seem to depend on retinal photoreceptor cells and their visual pigments, as Retinal-degenerate (Rd) mice show a normal circadian response to light 17,18 . Since mammalian cry genes are speci®cally expressed in the ganglion and inner nuclear layer of the retina, the Cry1 and Cry2 proteins are possible candidates for circadian photoreceptors 19 . To explore the biological function of mammalian Cry1 and Cry2, we have generated cry1 and cry2 mutant mice through gene targeting in embryonic stem cells (Fig. 1). Analysis of the transcrip- tional status of the targeted cry1 and cry2 genes, using the reverse transcription-long-range polymerase chain reaction, revealed no detectable transcripts in the corresponding knockout animals, thus demonstrating that we have created null-mutant mice. Targeted cry1 and cry2 alleles both segregate at expected mendelian ratios, indicating that the absence of either Cry1 or Cry2 does not interfere with embryonic development. Moreover, cry1 and cry2 mutant mice are completely healthy and show no overt phenotype (the oldest animals are now 14 and 7 months, respectively). We analysed the possible role for Cry proteins in the biological clock by measuring the circadian wheel-running behaviour of cry-knockout mice under normal light/dark (LD) cycles and in constant darkness (dark/dark; DD). We made two unexpected observations. First, compared with wild-type mice which, when subjected to DD conditioning, have a free-running rhythm close to 24 hours (t  23:77 6 0:07 h (n  14)), the internal clock of cry1 mutants runs signi®cantly faster (t  22:51 6 0:06h (n  9); P , 0:00001) (Fig. 2a, b). In contrast, cry2 mutants exhibit a clear increase in period length (t  24:63 6 0:06 h (n  5); P , 0:00001) (Fig. 2c). Heterozygous animals showed wheel-running patterns comparable to wild-type mice, and there were no clear sex- or age-related differences (data not shown). These ®ndings suggest that Cry1 and Cry2 antagonis- tically modulate the period length of the clock. Second, under LD conditions, both mutants show a circadian periodicity of 24 h (Fig. 2a±c), suggesting that a de®ciency in either cry1 or cry2 does not produce a detectable loss of light entrainment of locomotor activity. However, as cry1 mice still contain a functional Cry2 protein which may (partly) take over the function of Cry1 (and vice versa), functional redundancy may blur the phenotypic out- come. Thus, it was of interest to examine double-mutant mice. Like cry-single-knockout mice, double-mutant animals are viable and show no gross phenotypic abnormalities (by 5 months old). Unexpectedly, these mice (n  8) still display an essentially 24-h circadian rhythm under LD conditions (Fig. 2f, upper part). This suggests that, despite the absence of both Cry proteins, the bio- logical clockÐas re¯ected by locomotor activity±may still receive a light input. However, when double-mutant mice are shifted to a DD regime, they show a striking instantaneous and complete circadian arrhythmicity (Fig. 2f). This indicates that there is no internal circadian clock running with any signi®cant momentum, although we do not exclude the possibility that there is still an ultradian component. Note also that the instantaneous arrhythmicity in double-mutant mice differs from any clock mutant analysed thus far. This includes the only mouse `clock' mutant (clock) described to