Oscillating Molecules and Circadian Clock Output Mechanisms

NEWS & VIEWS Oscillating molecules and circadian clock output mechanisms

The circadian clock system is crucial for the regulation of daily rhythms in processes ranging from gene expression to the mammalian sleep/wake cycle. This review will summarize evidence from studies in model systems for the involvement of transcriptional and post-transcriptional mechanisms in circadian clock output.

Most neurobiologists, when told about circadian clock output pathways mediating physiological and genes and the seminal advances in understanding the behavioral rhythmicity. molecular mechanisms constituting the circadian clock, will ask: ‘Yes, but how does the clock do what Transcriptional mechanisms and clock output it does?’. This is to say, what are the mechanisms that dictate the accurate transmission of temporal infor- Many investigators in the field of circadian rhythms mation from the clock, located in a discrete region of would speculate that output mechanisms must, in the central nervous system, to the neuroeffector sys- some way, involve the action of transcription factors tems that generate defined circadian rhythms? In the and the regulation of ‘downstream’ genes that encode lingo of the field: what are the signaling or output path- clock output components. This aphorism stems in part ways that mediate the clock control of diverse circad- from several observations suggesting an involvement of ian rhythms? Questions about clock output, posed by transcriptional mechanisms in clock biology. As long such neurobiologists, are indicative of how little we ago as the late 1960s, an explicit model for the circad- really know about such output pathways. This mini- ian clock itself, was formulated that postulated an review will attempt to summarize recent molecular important role for transcriptional activity;7 this model genetic studies that have provided insights about cir- hypothesized that the timing of RNA synthesis from cadian clock output mechanisms. An understanding of a polycistronic complex of DNA (a so-called chronon) circadian clock output in any system may provide provided the temporal information required for circad- insights about the output mechanisms that control ian functions. This hypothesis was based on experi- complex mammalian behaviors. mental results which were interpreted as suggesting an involvement of nucleic acid metabolism in light- induced phase shifting of the clock.8 While the Biochemistry of the clock mechanism Chronon theory has been discounted in its purest form, The isolation and analysis of clock genes in a variety contemporary molecular analysis has established that of organisms, from fruitflies and fungi to mammals1–6 transcriptional mechanisms really are crucial for clock have generated tremendous knowledge about clock function. The best indication of this axiom is that cer- molecules and mechanisms, and have driven the for- tain transcription factors, of the bHLH/PAS class mulation of explicit models that purport to explain the (which contain a basic helix-loop-helix DNA-binding biochemistry of the circadian clock. The prevailing domain and a PAS dimerization domain) are molecular model of circadian clock function includes oscillating constituents of the circadian clock in both fungi and RNAs and proteins as state variables of the clock, and it mammals4–6,9 and a PAS-containing protein called Per- is couched in terms of an autoregulatory loop in which iod (the P in PAS) is known to be a component of the changes in one set of state variables (the proteins) nega- Drosophila clock mechanism.2,3 Furthermore, studies tively regulate the other. This model is supported prim- from several laboratories have demonstrated that acute arily by studies of clock genes1 in Neurospora (frq, wc- perturbations of transcription, using reversible inhibi- 1, wc-2) and Drosophila (per, tim). Currently, however, tors, can alter circadian period and reset the clock the oscillator is described as a closed molecular loop, mechanism.10,11 and it is not obvious how this loop is coupled to the In attempts to identify pacemaker or clock output components, circadian regulated RNAs, encoding transcription factors12–14 and other products15–17 have been Correspondence: FR Jackson, X Zhang and GP McNeil, Dept of described in several different organisms, and at least Neuroscience, Tufts University School of Medicine, Boston, MA some of these RNA oscillations are known to depend 02111, USA. E-mail: [email protected] on transcriptional mechanisms.18 The best charac- News & Views 382 terized of these RNAs, however, encode metabolic overexpression) do not affect all clock-driven RNA enzymes15,17 or structural proteins19 that are not likely rhythms, it has been hypothesized that this negative to function as regulatory elements of a clock output feedback loop acts as a ‘slave’ oscillator, coupling the pathway. Nonetheless, based on models of the circad- clock to specific output pathways.25 At present, how- ian system,20 which postulate the existence of ‘master’ ever, the physiological functions of the presumed out- and ‘slave’ (or driven) oscillators, it is possible that cer- put pathways are unknown. tain oscillating RNAs might be important for clock out- Whereas oscillations in clock state variables (such as put functions. per, tim and frq products) are thought to be driven by RNAs encoding certain basic leucine zipper (bZip) changes in the abundances and/or activities of tran- transcription factors such as D-binding Protein (DBP), scription factors (based largely on the identity of cer- Thyroid Embryonic Factor (TEF), and Hepatocyte Leu- tain clock components as putative bHLH transcription kemia Factor (HLF) exhibit marked circadian fluctu- factors; see earlier discussion), the Atgrp7 rhythm ations in abundance (Ref. 21 and references therein),21 clearly involves the activity of a putative RNA-binding and it has been postulated that circadian changes in protein; ie, the product of the Atgrp7 gene itself. the activities of such factors account for the observed Although it is not yet known whether the Atgrp7 RNA- diurnal regulation of known target genes.22 Interest- binding protein influences transcriptional or post-tran- ingly, DBP mRNA abundance shows circadian oscil- scriptional events to regulate Atgrp7 RNA abundance, lations within the suprachiasmatic nuclei, the anatom- the identity of the protein is consistent with obser- ical location of the mammalian circadian pacemaker, vations in at least two other systems that RNA-binding and DBP-null mutant mice have slightly shorter circad- proteins might have roles as clock output components. ian periods as assayed by locomotor activity rhythms.21 However, the mutants also have lower levels of spon- Gonyaulax bioluminescence rhythms and RNA- taneous activity, and it is known that acute changes in binding proteins activity levels can phase shift the clock and/or alter The unicellular alga Gonyaulax polyedra exhibits well- circadian period.23 Thus, at present there is little evi- characterized circadian rhythms in bioluminescence,28 dence that DBP functions as a clock molecule, and per- which have been used as an assay of clock output by haps it is more likely that DBP plays a role as a clock Hastings and colleagues. Elegant studies by these output component.21 investigators (reviewed in Ref. 29) have demonstrated Perhaps not surprisingly, there is also evidence from that clock-driven translational regulatory mechanisms studies in other model systems that specific transcrip- contribute to the observed circadian fluctuations in the tion factors can mediate clock output functions. In the abundance of proteins which are known to participate cyanobacterium Synechococcus, for example, a sigma- in luminescence (such as luciferase-binding protein or like transcription factor is postulated to function as a LBP). That is to say, in contrast to clock RNAs such as clock output element to control a subset of rhythmi- per, tim,orfrq, lbp mRNA does not oscillate in abun- cally transcribed genes.24 Obviously, transcriptional dance; rather LBP protein synthesis changes in a cir- regulatory mechanisms or changes in RNA stability cadian manner.29 Furthermore, studies in Gonyaulax must be involved in the control of all rhythms in and the unrelated unicellular Chlamydomonas rhein- RNA abundance. hardii, suggest that RNA-binding proteins, with an affinity for a defined GU-rich element within lbp 3′ untranslated sequence, are responsible for the circad- Post-transcriptional mechanisms and clock output ian regulation of lbp mRNA translation.30 It is also of The Arabidopsis Atgrp7 RNA-binding protein and interest that these RNA-binding proteins, which so far clock output have been identified only as lbp-binding factors from The activities of certain oscillating molecules, in parti- extracts of Gonyaulax or Chlamydomonas, have cular the PAS domain-containing clock proteins, are characteristics of repressor proteins; ie, the apparent critical for the function of the circadian system; how- binding activities of these factors are highest at the cir- ever, many of the rhythmically transcribed RNAs more cadian time when LBP synthesis is lowest. likely represent very distal targets of the clock rather than key regulatory elements of a clock output path- Lark, a Drosophila RNA-binding protein mediating way. A clear exception to this statement is the Arabi- clock output dopsis thaliana Atgrp7 gene and its encoded product Screens for new rhythm mutants in Drosophila mel- (a putative RNA-binding protein), which appear to con- anogaster identified a gene known as lark,31 which stitute a circadian autoregulatory feedback loop in encodes a putative RNA-binding protein.32 Mutations which Atgrp7 protein negatively regulates Atgrp7 RNA in the lark gene cause an ‘early-phase’ phenotype for abundance25 (see Figure 1). Interestingly, this mech- the circadian rhythm of eclosion (the emergence of anism is remarkably similar to the one postulated for adults from pupal cases), one of the two circadian the clock itself.1 The Atgrp7 oscillations are clearly rhythms that have been well-characterized in this under clock control, as Arabidopsis mutations such as organism. Specifically, null mutations of the gene have toc1, which shorten the circadian period,26 have com- dominant effects on the eclosion rhythm, resulting in parable effects on the Atgrp7 RNA rhythm.27 Since per- an early-eclosion phenotype (notwithstanding what turbations of the Atgrp7 rhythm (by protein has been published in certain recent reviews), and this News & Views 383

Figure 1 Schematic model for the Atgrp7 oscillator. The ٜ symbol indicates repression. This figure was adapted from one kindly provided by Dr Dorothee Staiger. effect is thought to be a result of decreased gene dosage. idea that the protein acts as a clock output element. Thus, it is of interest that the opposite sort of pertur- These studies have demonstrated that lark protein bation, increased lark-gene dosage, is associated with exhibits rhythmic and high-amplitude (ෂ 10 fold) a ‘late-eclosion’ phenotype.31 Such genetic results sug- changes in abundance in both pupal and adult head gest that lark protein behaves as a repressor molecule protein extracts. The lark rhythm is observed during a (like the Arabidopsis Atgrp7 product and lbp-binding diurnal cycle (ie, while pupae are entrained to a proteins of Gonyaulax) in regulating the daily timing of light/dark cycle), and importantly it persists in con- adult eclosion. The observation that clock period and stant conditions (conditions devoid of temporal cues, environmental input to the clock are apparently nor- in which adult eclosion also continues to exhibit cir- mal in lark mutants31 indicates that the gene product cadian rhythmicity). Remarkably, the phasing of the acts as a clock output element, rather than as a compo- rhythm, with a trough during the night and several nent of the circadian clock mechanism. Consistent hours prior to the daily peak of adult eclosion, is with this supposition, lark mutations (when carried in entirely consistent with the hypothesis that lark mutant heterozygotes) do not have discernible effects behaves as a repressor molecule in mediating the clock on the locomotor activity rhythm, which is thought to control of eclosion. It is equally important to note that be controlled by the same circadian clock. In addition lark mRNA abundance is constant during the diurnal to a role in clock output, lark protein is also essential cycle,32 making it unlikely that transcriptional regulat- for embryonic development,31 and recent studies indi- ory mechanisms contribute to the observed circadian cate that the protein has both maternal and zygotic changes in lark protein abundance. Therefore, clock- roles in development (G McNeil, X Zhang, M Roberts, regulated translational mechanisms (Figure 2, model 1) and FR Jackson, in preparation). As such, lark was the and/or alterations of protein stability (Figure 2, model first essential gene shown to play a role in circadian 2) must be reponsible for the lark protein rhythm, and rhythmicity. one wonders whether the control of lark expression Molecular characterization of the lark gene indicates will resemble the circadian regulation of the Gonyau- that the encoded protein is a novel member of the RNA lax lbp mRNA. It is also possible that lark protein Recognition Motif, or RRM, class33,34 of RNA-binding ‘feeds back’, as part of a slave oscillatory loop (Figure 2, proteins,32 suggesting that the protein has a function in model 3), to autoregulate its own expression, similar RNA processing, export, or translational control. Inter- to the feedback regulation described for the Atgrp7 estingly, the protein is localized within the nucleus of rhythm. most or all neurons, but has a cytoplasmic distribution within a subset of neurons (ෂ 15–20) that contain the Coda neuropeptide crustacean cardioactive peptide (CCAP).35 In other insects, this neuropeptide is thought So, in the words of the philosopher Montaigne, ‘Que to play a prominent role in the physiological regulation sais-je’? What do we really know about clock output of eclosion.36 A differential localization of lark protein mechanisms? While certain studies have demonstrated within the nucleus or cytoplasm of distinct cell types a role for transcription factors and RNA-binding pro- is consistent with the notion that the protein might teins in clock output, we are only just beginning to subserve multiple RNA-binding functions. Alterna- understand the biochemistry of output mechanisms. tively, the protein might shuttle between intracellular The molecular identities of the factors which regulate compartments, in a manner similar to other RNA-bind- circadian rhythms in the Atgrp7 or lark RNA-binding ing proteins,37 and have an essential function in one proteins, for example, are unknown. Moreover, to compartment or the other. quote our omnipresent neurobiologist, we still do not Perhaps the most intriguing finding regarding lark is know how such RNA-binding proteins do what they the recent observation that lark protein abundance is do. However, it is likely that characterization of the regulated by the circadian clock,35 consistent with the RNA targets of such proteins will lead to the identifi- News & Views 384 References

1 Dunlap JC. Genetic and molecular analysis of circadian rhythms. Ann Rev Genet 1996; 30: 579–601. 2 Rosbash M, Allada R, Dembinska M, Guo WQ, Le M, Marrus S et al.ADrosophila circadian clock. CSHSQB 1996; 61: 265–278. 3 Young MW, Wager-Smith K, Vosshall L, Saez L, Myers MP. Mol- ecular anatomy of a light-sensitive circadian pacemaker in Droso- phila. CSHSQB 1996; 61: 279–284. 4 King DP, Zhao YL, Sangoram AM, Wilsbacher LD, Tanaka M, Antoch MP et al. Positional cloning of the mouse circadian Clock gene. Cell 1997; 89: 641–653. 5 Sun ZS, Albrecht U, Zhuchenko O, Bailey J, Eichete G, Lee CC. RIGUI, a putative mammalian ortholog of the Drosophila period gene. Cell 1997; 90: 1003–1011. 6 Tei H, Okamura H, Shigeyoshi Y, Fukuhara C, Ozawa R, Hirose M et al. Circadian oscillation of a mammalian homologue of the Drosophila period gene. Nature 1997; 389: 512–516. 7 Ehret CF, Trucco E. Molecular models for the circadian clock I. The chronon theroy. J Theoret Biol 1967; 15: 240–262. 8 Ehret CF. Action spectra and nucleic acid metabolism in circadian rhythms at the cellular level. CSHSQB 1960; 25: 149–158. 9 Crosthwaite SK, Dunlap JC, Loros JJ. Neurospora wc-1 and wc-2: transcription, photoresponses, and the origins of circadian rhythmicity. Science 1997; 276: 763–769. 10 Raju U, Koumenis C, Nunez-Regueiro M, Eskin A. Alteration of the phase and period of a circadian oscillator by a reversible transcription inhibitor. Science 1991; 253: 673–675. 11 Khalsa SBS, Whitmore D, Bogart B, Block GD. Evidence for a central role of transcription in the timing mechanism of a circadian clock. Am J Physiol Cell Physiol 1996; 271C: 1646–1651. 12 Wuarin J, Schibler U. Expression of the liver-enriched transcriptional activator protein DBP follows a stringent circadian rhythm. Cell 1990; 63: 1257–1266. 13 Green CB, Besharse JC. Identification of a novel vertebrate circadian clock-regulated gene encoding the protein nocturin. Proc Natl Acad Sci USA 1996; 93: 14884–14888. Figure 2 Alternative mechanisms for the LARK biochemical 14 Rouyer F, Rachidi M, Pikielny C, Rosbash M. A new gene encoding rhythm. The + or − symbols indicate the potential for positive a putative transcription factor regulated by the Drosophila circad- .or negative regulation. The ٜ symbol indicates negative regu- ian clock. EMBO J 1997; 16: 3944–3954 lation. 15 Bell-Pedersen D, Shinohara ML, Loros JJ, Dunlap JC. Circadian clock-controlled genes isolated from Neurospora crassa are late night- to early morning-specific. Proc Natl Acad Sci USA 1996; 93: 13096–13101. 16 Anderson SL, Teakle GR, Martino-Catt SJ, Kay SA. Circadian clock- cation of the factors through which they mediate the and phytochrome-regulated transcription is conferred by a 78 bp cis-acting domain of the Arabidopsis CAB2 promoter. Plant J 1994; clock control of physiological processes. In addition, 6: 457–470. 35 immunocytochemical studies of output molecules 17 Van Gelder RN, Bae H, Palazzolo MJ, Krasnow MA. Extent and coupled with a genetic (ie, functional) analysis may character of circadian gene expression in Drosophila melanogaster: help to elucidate the cellular components of clock out- identification of twenty oscillating mRNAs in the fly head. Curr put pathways in animal systems. As certain output Biol 1995; 5: 1424–1436. elements appear to be evolutionarily conserved,38 these 18 Loros JJ, Dunlap JC. Neurospora clock-controlled genes are regulated at the level of transcription. Mol Cell Biol 1991; 11: 558–563. studies could also pave the way for the functional 19 Bell-Pedersen D, Dunlap JC, Loros JJ. The Neurospora circadian characterization of mammalian clock output compo- clock-controlled gene, ccg-2 is allelic to eas and encodes a fungal nents. The hope is that studies of clock output will hydrophobin required for formation of the conidial rodlet layer. allow us to ‘open up’ the oscillatory loop comprising Genes Dev 1992; 6: 2382–2394. the circadian clock so that we can finally see how it is 20 Pittendrigh CS. Circadian oscillations in cells and the circadian organization of multicellular systems. In: Schmitt FO, Worden FG connected to the rest of the organism. (eds). The Neurosciences, Third Study Program. MIT Press: Cam- bridge, 1974, pp 437–458. 21 Lopez-Molina L, Conquet F, Dubois-Dauphin M, Schibler U. The Acknowledgements DBP gene is expressed according to a circadian rhythm in the suprachiasmatic nucleus and influences circadian behavior. EMBO J Supported by NIH grant R01 HL59873 to FRJ. 1997; 16: 6762–6771. 22 Lavery DJ, Schibler U. Circadian transcription of the cholesterol 7 FR Jackson, X Zhang and GP McNeil alpha hydroxylase gene may involve the liver-enriched bZIP protein DBP. Genes Dev 1993; 7: 1871–1884. Dept of Neuroscience 23 Weisberger D, Redlin U, Mrosovsky N. Lengthening of circadian Tufts University School of Medicine period in hamsters by novelty-induced wheel running. Physiol Boston, MA 02111, USA Behav 1997; 62: 759–765. News & Views 385 24 Tsinoremas NF, Ishiura M, Konda T, Andersson CR, Tanaka K, Tak- phila melanogaster that identifies a gene with an essential embry- ahashi H et al. A sigma factor that modifies the circadian onic function. Genetics 1993; 135: 1077–1090. expression of a subset of genes in cyanobacteria. EMBO J 1996; 15: 32 Newby LM, Jackson FR. Regulation of a specific circadian clock 2488–2495. output pathway by lark, a putative RNA-binding protein with 25 Heintzen C, Nater M, Apel K, Staiger D. AtGRP7, a nuclear RNA- repressor activity. J Neurobiol 1996; 31: 117–128. binding protein as a component of a circadian-regulated negative 33 Query CC, Bentley RC, Keene JD. A common RNA-recognition feedback loop in Arabidopsis thaliana. Proc Natl Acad Sci USA motif identified within a defined U1 RNA binding domain of the 1997; 94: 8515–8520. 70K U1 snRNP protein. Cell 1989; 57: 89–101. 26 Millar AJ, Carre´ IA, Strayer CA, Chua N-H, Kay SA. Circadian clock 34 Burd CG, Dreyfuss G. Conserved structures and diversity of func- mutants in Arabidopsis identified by luciferase imaging. Science tions of RNA-binding proteins. Science 1994; 265: 615–621. 1995; 267: 1161–1163. 35 McNeil G, Zhang X, Genova G, Jackson FR. A molecular rhythm 27 Kreps JA, Simon AE. Environmental and genetic effects on circad- mediating circadian clock output in Drosophila. Neuron 1998; 20: ian clock-regulated gene expression in Arabidopsis. Plant Cell 297–303. 1997; 9: 297–304. 36 Gammie SC, Truman JW. Neuropeptide hierarchies and the acti- 28 Hastings JW, Sweeney BM. A persistent diurnal rhythm of lumi- vation of sequential motor behaviors in the hawkmoth, Manduca nescence in Gonyaulax polyedra. Biol Bull 1958; 115: 440–458. sexta. J Neurosci 1997; 17: 4389–4397. 29 Morse DS, Fritz L, Hastings JW. What is the clock? Translational 37 Pinol-Roma S, Dreyfuss G. Shuttling of pre-mRNa binding proteins regulation of circadian bioluminescence. TIBS 1990; 15: 262–265. between nucleus and cytoplasm. Nature 1992; 355: 730–732. 30 Mittag M. Conserved circadian elements in phylogenetically 38 Jackson FR, Banfi S, Guffanti A, Rossi E. A novel zinc-finger-con- diverse algae. Proc Natl Acad Sci USA 1996; 93: 14401–14404. taining RNA-binding protein conserved from fruitflies to humans. 31 Newby LM, Jackson FR. A new biological rhythm mutant of Droso- Genomics 1997; 41: 444–452.