Liste des participants
Isabelle Carré (University of Warwick, UK), Jay Dunlap, organisateur (Dartmouth Medical School, Hanover, USA), Nicholas Foulkes (Université Louis Pasteur, Illkirch, France), Albert Goldbeter, organisateur (Université Libre de Bruxelles, Belgique), Susan S. Golden (Texas A & M University, USA), Didier Gonze (Université Libre de Bruxelles, Belgique), Charlotte Helfrich-Förster (Institute of Zoology, Tübingen, Germany), Carl H. Johnson (Vanderbilt University, Nashville, USA), Takao Kondo (Nagoya University, Japan), Jean-Christophe Leloup (Université Libre de Bruxelles, Belgique), Francis Lévi (ICIG, Hôpital Paul Brousse, Villejuif, France), Jennifer Loros (Dartmouth Medical School, Hanover, USA), Andrew Millar (University of Warwick, UK), Hitoshi Okamura (Kobe University School of Medicine, Japan), Till Roenneberg (University of Munich, Germany), Michael Rosbash (Howard Hughes Medical Institute, Waltham, USA), François Rouyer (Institut Alfred Fessard, Gif-sur-Yvette, France), Ueli Schibler (Université de Genève, Suisse), Elaine Tobin (UCLA, USA), Michael Young, organisateur (Rockefeller University, New York, USA)
par Albert Golbeter
4 – 10 septembre 1999
Les rythmes circadiens se produisent dans la plupart des organismes vivants, y compris certaines espèces de bactéries. Ces rythmes, d’une période proche de 24 h, sont au nombre des rythmes biologiques les plus importants d’un point de vue physiologique, car ils permettent aux organismes de s’adapter à la périodicité naturelle de l’environnement. Chez l’homme un grand nombre de fonctions physiologiques varient de manière circadienne, comme le montrent le cycle veille-sommeil et celui de la nutrition. Ces observations possèdent des implications importantes pour la chronopharmacologie dont le but est d’optimiser l’efficacité des médicaments en fonction du temps de leur administration.
Les études génétiques et biochimiques ont permis des progrès importants dans l’étude des bases moléculaires des rythmes circadiens. Ces avancées ont connu une accélération ces dernières années, au point qu’une vision unifiée des mécanismes de l’horloge circadienne commence à émerger dans différents organismes. Ainsi, un mécanisme récurrent dans la genèse des rythmes circadiens chez des organismes comme la drosophile, Neurospora, les mammifères et les plantes implique des boucles d’auto-régulation négative de l’expression de gènes de l’horloge codant pour des facteurs de transcription.
Le but de ce Colloque à la Fondation des Treilles tenu du 4 au 10 septembre 1999 fut de discuter les progrès récents réalisés quant aux mécanismes de régulation sous-tendant l’horloge circadienne dans les modèles expérimentaux les plus étudiés. Historiquement, les progrès sur les bases moléculaires des rythmes circadiens ont été effectués tout d’abord chez la mouche drosophile et le champignon Neurospora. Par la suite, des progrès rapides ont également été effectués chez les mammifères, pour lesquels plusieurs gènes contrôlant l’organisation circadienne chez la drosophile ont été découverts. D’autres avancées ont été réalisées, de manière similaire, chez les plantes et les cyanobactéries.
Les neuf sessions de ce Colloque ont été consacrées, successivement, aux mécanismes moléculaires des rythmes circadiens chez les cyanobactéries, Neurospora, la drosophile, les plantes et les mammifères. D’autres organismes, comme le poisson zèbre, ont également été considérés. Les implications cliniques des rythmes circadiens en médecine ont été envisagées dans l’une des sessions consacrées aux mammifères. La majorité des présentations étaient de nature expérimentale, mais les modèles théoriques pour les rythmes circadiens ont également été discutés.
Compte rendu (en anglais)
The abstracts of the talks are reproduced below in the order of their presentation at the workshop.
Circadian clocks in cyanobacteria: mechanism and fitness
Carl H. Johnson, Vanderbilt University, Nashville
Circadian (daily) rhythms are ubiquitous in eukaryotes, and are also found in eubacteria among cyanobacteria. Previous studies have suggested that longevity, growth, and developmental rate are improved when organisms are maintained on light/dark cycles whose period is similar to the period of the endogenous circadian clock. However, some of those studies have not been reproducible and no studies have demonstrated that reproductive fitness per se is improved by consonance between the endogenous clock and the environmental cycle. Moreover, mutations of circadian clock genes in flies and fungi do not obviously impair reproductive fitness under laboratory conditions. We addressed the adaptive significance of circadian rhythmicity by testing the relative fitness under competition between various strains of the cyanobacterium Synechococcus sp. strain PCC 7942 expressing different circadian periods. Strains that had a circadian period similar to that of the light/dark cycle were favored under competition in a manner that indicates the action of soft selection. Arrhythmic strains could not compete well against wild-type strains in light/dark cycles, but they could compete effectively in constant light. These data indicate that having a circadian clock with a period that matches that of the environmental cycle enhances fitness.
Mutational analyses of clock function in Synechococcus have unveiled a circadian clock gene cluster called kaiABC. Twenty five clock mutations have been mapped to the three kai genes. Inactivation of any single kai gene abolished these rhythms. Continuous kaiC-overexpression represses the kaiBC promoter, whereas kaiA-overexpression enhanced it. Pulses of kaiC-overexpression resets the phase of the rhythms. All three Kai proteins appear to interact with each other and these interactions seem to be crucial for normal clock function. Our current working hypothesis is that negative feedback control of kaiC-expression by KaiC generates a circadian oscillation in cyanobacteria and KaiA sustains the oscillation by enhancing kaiC-expression. Studies of Kai protein expression will be reported at the meeting. The kaiC gene contains two Walker A (P-loop) motifs. Mutation of this motif disrupts circadian rhythms of gene expression. These Walker A motifs appear to mediate binding of nucleotides.
(Much of the work described above was performed in collaboration with Drs. Takao Kondo & Masahiro Ishiura [Nagoya University] and Dr. Susan Golden [Texas A&M University.])
Circadian feedback loop and timing mechanism of cyanobacteria
Takao Kondo, Nagoya University
Cyanobacteria are the simplest model organisms that exhibit circadian rhythms. Thus, elucidation of the cyanobacterial circadian oscillator would be efficient here because various molecular genetic approaches can be applied. The kaiABC gene cluster is identified to be essential for generation of circadian rhythms in cyanobacterium Synechococcus sp. strain PCC 7942.
Circadian oscillation of Synechococcus is likely generated by a negative feedback regulation of kaiC transcription by KaiC, and positive feedback through KaiA that transactivate kaiC. To understand how this putative loops generate circadian oscillation, we have studied the biochemical functions of Kai proteins and their associates.
The following results imply that biochemistry of Kai protein and associates are key to understand circadian timing mechanism of cyanobacteria.
1) KaiA, KaiB and KaiC directly associated each other in yeast cells, in vitro and in cyanobacterial cells. Some mutation altered binding in vitro.
2) Two-hybrid screen for KaiC-binding proteins has identified a bacterial two-component histidine kinase gene. Disruption of this gene severely lowered amplitude of the rhythms but weak rhythmicity still persisted. Therefore, this kinase is a close associate of Kai-based circadian oscillator.
3) KaiC has two potential ATP-/GTP-binding motifs. We demonstrated that KaiC bound both ATP and GTP in vitro and a point-mutation of P-loop completely nullified circadian rhythm and greatly reduced ATP-binding activity.
4) Expression of kaiA is regulated by a newly cloned clock associate gene, pex.
Genes other than kaiABC that are important for circadian rhythms in Synechococcus
Susan S. Golden, Mitsunori Katayama, Oliver Schmitz, Stan Williams, Nikos Tsinoremas, and Alex Sivan, Texas A&M University, College Station
In the cyanobacterium Synechococcus the clock controls daily rhythms of gene expression. We can readily monitor the period, amplitude, and phasing of the circadian rhythm from any Synechococcus promoter by using luciferase gene fusions (Vibrio harveyi luxAB, or firefly luc), such that light production reports transcription in real time. Previous research using this strategy identified a three-gene locus (kaiA, kaiB, kaiC) that is fundamentally important for cyanobacterial circadian rhythms. Mutations in any of the kai genes can cause a change in circadian period, and inactivation of any of them results in arrhythmia.
We are interested in identifying other loci that are important for generating circadian rhythms, as well as the components of input pathways for synchronizing the clock with the environment and output pathways that control cyclic gene expression. We have isolated transposon-tagged (Tn5) mutants that are likely to be affected in clock input and output pathways, and one that appears to unmask a second kai-independent oscillation. The two input mutants include one defective in a histidine protein kinase (CikA, circadian input kinase) that is a member of the phytochrome family of proteins and another lacking an apparent FeS center-binding protein (TnpK). The phenotypes and the protein sequences suggest that the former is defective in interpreting light and dark signals in resetting the clock, whereas the latter may fine tune the period of the rhythm in relationship to photosynthetic activity.
Several output mutants have been isolated that affect the circadian rhythmicity of a subset of genes, including kaiA, the three psbA genes, and a ‘generic’ Escherichia coli promoter, but not kaiBC or the dawn-peaking purF gene. These include mutants defective for a group 2 sigma factor (RpoD2, low amplitude phenotype) and two novel proteins lacking in recognizable functional motifs. One of these (CpmA) affects phase angle (and for some genes, amplitude) and one (Tnp5) greatly disturbs the rhythm, nearly eliminating troughs and changing phase angle of apparent peaks. Two mutants are defective in genes that may be part of output pathways, but are likely to be closely allied to the central clockwork. One is null mutant of sasA, which encodes a histidine protein kinase whose N terminus resembles KaiB and that interacts physically with KaiC. Amplitude of expression from all downstream genes is greatly reduced in the mutant, as are the levels of kai transcripts and proteins. The mutant also shows a short period phenotype and impaired growth in LD cycles.
Mutation of another group 2 sigma factor gene, sigC, causes a long period phenotype (27.5 h vs 25 h) in the circadian oscillation of many genes, including the psbA genes and kaiA. However, circadian period of kaiBC and purF is unaffected in the mutant. The 27.5 h rhythm of psbA genes in a sigC background is evident even when kai alleles that normally cause arrhythmia or other periods (33 h, 40 h) are present, suggesting that this mutation unmasks a kai-independent rhythm.
FRQ, WC-1 and WC-2 in Neurospora entrainment
Jennifer Loros, Jay Dunlap, Mike Collett, Deanna Denault, Kwangwon Lee, Yi Liu, Allan Froehlich, Christian Heintzen, Susan Crosthwaite, Deborah Bell Pedersen, Dartmouth Medical School, Hanover, USA
An important property of circadian oscillators is their ability to remain in appropriate phase, via entrainment, to the daily period of the earth’s rotation. Research concerning the functioning of the clock mechanism itself, through genetic and molecular identification of the genes and proteins that comprise the daily oscillator mechanism, has established the products of the frequency (frq) locus as components of the Neurospora clock, comprising a molecular negative feedback loop wherein the frq gene encodes two FRQ proteins which, in turn, feed back to turn down expression of the gene. Light delivered at any point within the circadian cycle acts rapidly to increase the level of frq transcript. The magnitude of the lightinduced increase in frq mRNA and the extent of clock resetting are correlated and the threshold, kinetics and magnitude of this response indicate elevation of the level of frq transcript in the cell is the initial clock-specific event involved in resetting of the clock by light (Crosthwaite et al., Cell 81, 1003, 1995).
This light induction requires the white-collar [WC] proteins, both of which are additionally required for positive progression of the clock in the dark (Crosthwaite et al., Science 276, 763, 1997), a demonstration at the molecular level of the extent to which the light input pathway to the clock is merged with the clock cycle itself. In a manner that appears to be almost identical in timing and kinetics to that described for Neurospora, light induction of the mper1 gene is associated with light resetting of the mouse circadian system (Shigeyoshi et al., Cell 91, 1043, 1997).
In additional to the general importance of this information for understanding how clocks are reset by light, the specific importance of understanding these genes has been affirmed with the identification by others of several functional and sequence homologue’s associated with either light or clock function. The wc-1 and wc-2 genes encode DNA binding proteins that contain PAS domains and, as transcriptional activators, correspond to the generic positive elements in the feedback loop now seen among all eukaryotic clocks. The presence of PAS domains in the white-collar proteins – involved in both signal transduction and in clock function -suggests that photoreception and circadian rhythmicity are linked not only at the physiological level but also at the molecular level, and suggests that circadian oscillators may have evolved from ancient proteins involved in signal transduction and photoresponsivity (Crosthwaite et al., Science 276, 763, 1997). Current work to be discussed will include the analysis of the protein products of these transcription factors to understand their exact role in photic entrainment of the Neurospora clock. We are examining WC protein regulation, their interactions with each other and with FRQ to understand their effect on frq gene expression and how environmental light releases the WC proteins from FRQ repression. Partially functional wc-2 alleles display several effects on the clock, including increased period length, reduced temperature compensation and reduced levels of frq mRNA and FRQ protein in constant darkness suggesting that wc-2’s function in frq regulation is essential and limiting in constant darkness proteins for light responsiveness of frq gene expression.
Clocks in nearly all non-homeothermic organisms interpret temperature stops as cues for dawn and steps down as cues for dusk. The absolute amount of FRQ protein is quite dependent on the ambient temperature so that even the low point in the daily FRQ cycle at 28°C is higher than the high point in the cycle at 20°C: this fact explains how temperature steps reset the clock. Following a step for 20°C to 28°C the amount of FRQ in the cell is always at or below the low point in the 28°C cycle, so the clock is reset to the tune of day corresponding to the low point in the cycle, approximately subjective dawn. Conversely, a high to low temperature step sets the clock to the time of day corresponding to the high point in FRQ levels, approximately dusk (Liu et al., Science 281, 825-29, 1998).
Regulators of FRQ expression and FRQ levels in Neurospora
Jay C. Dunlap, Jennifer Loros, Yi Liu, Michael Collett, Christian Heintzen, Allan Froehlich, Deanna Denault, Kwangwon Lee, Hildur Colot, Dartmouth Medical School, USA
The circadian system in Neurospora comprises an autoregulatory feedback cycle, wherein the White Collar (WC1 and WC2) proteins promote the expression of the frequency (frq) gene which encodes two forms of the FRQ protein, each of which can feed back via interactions with the WCs to depress the level of transcript arising from the frq gene. As predicted, both frq RNA and protein cycle in abundance showing peaks in the subjective day with frq mRNA peaking about 4 hrs after subjective dawn (CT4) and FRQ peaking about 4 hrs later at CT 8-10 (Aronson et al., Science 263, 1578, 1994; Garceau et al., Cell 89, 469, 1997; Merrow et al., PNAS 94, 3877, 1997). The two forms of FRQ arise from alternative translation initiation sites; ambient temperature influences the clock by determining both the absolute amount of FRQ and the ratio between the two forms (Garceau et al., Cell 89, 469, 1997; Liu et al., Cell 89,477, 1997). FRQ is phosphorylated as soon as it is made and is processively re-phosphorylated over the course of the day (Garceau et al., Cell 89, 469, 1997), a modification that appears to play a role in regulating turnover. FRQ spends the early part of each day immediately following its synthesis in the nucleus, and the timing of its localization there appears to be regulated (Luo et al., EMBO J. 17, 1228, 1998). Nuclear localization is required for FRQ function. Within a few hours after FRQ acts to depress the level of its transcript, FRQ levels in the nucleus begin to fall although they continue to rise in the cytoplasm for a few hours before beginning to fall there also. Some FRQ persists in the cytoplasm till the mid-subjective night when positive acting factors, encoded by the wc-1 and wc-2 genes, turn on synthesis of frq and begin the cycle again (Crosthwaite et al., Science 276, 763, 1997). Interestingly, the transcription of neither wc gene appears to be strongly rhythmic although WC1 levels appear to cycle and mutations in wc-2 can lead to period length effects. The WC-1 and WC-2 proteins have PAS domains similar to clock-associated proteins from the mouse (mPER1,2,3, CLOCK, BMAL) and the fly (PER), and also similar to light-response associated proteins from a number of systems, suggesting that clock molecules may have arisen from ancient proteins involved in light responsivity (Crosthwaite et al., Science 276, 763, 1997).
There are several necessary expectations and consequences of such a negative feedback loop the we have been testing. First, there should be interactions between WC1, WC2, and FRQ. This is clearly the case: WC2 antiserum immunoprecipitates WC1 and FRQ, WC2 interacts with FRQ in a GST pull-down assay, and sucrose gradient centrifugation of extracts from cells grown in the dark reveals a complex(es?) containing WC1, WC2, and FRQ. A second prediction is that mutations in the positive elements should alter clock function, and indeed a partial loss-of-function allele in wc-2 shows both a long period length and altered temperature compensation. Finally, if the phosphorylation of FRQ is important in regulating the turnover of this molecule, then mutations that abrogate the phosphorylation of specific sites on FRQ should reduce turnover (increase the stability) of FRQ and thereby lengthen the circadian period length. To assess this, several sites of phosphorylation on FRQ have been identified; mutation to nonphosphorylatable amino acids both increases FRQ stability and increases the period of the clock to approximately 35 hrs.
The role of feedbacks in circadian systems
Till Roenneberg and Martha Merrow, University of Munich, Germany
The circadian system is implemented at the cellular level. Its outputs control many metabolic functions and, at the same time, the rhythm generating loop appears to be exquisitely shielded from metabolic variations. Although the recent decade of circadian research has brought insights into how circadian periodicity may be generated at the molecular level, still little is known about the relationship between this molecular feedback loop and metabolism both at the cellular and the organismic level.
The marine unicell Gonyaulax and the fungus Neurospora are extremely different organisms in regard to phylogenetic position, habitat, and temporal ecology. Both have been models for the study of circadian system over several decades, but the experimental approaches used for this research have been very different. The Gonyaulax circadian system has mainly been studied using biochemistry and physiology, while Neurospora has been used for molecular and genetic investigations. The former model has provided insights into circadian ecology and metabolism as well as into the physiological complexity of the circadian clock (of a single cell). The latter was one of the first systems in which a negative feedback loop bas been described which fulfills the traditional criteria for a circadian oscillator.
In theoretical modeling and in experiments, the specific qualities and characteristics of these two systems have been combined, as well as methods of the physiological and the molecular eras of circadian research. The aim of these studies is to answer conceptual questions of how circadian systems are organized at the molecular level beyond forming a transcriptional feedback. First results show that a cryptic oscillator with circadian qualities is present in Neurospora strains that do not produce functional FRQ protein. The role of feedback loops within the circadian system (those that generate the rhythmicity and those that function as control loops) and the integration of these loops into cellular metabolism will be discussed.
Molecular genetics of circadian rhythms in Drosophila
Michael Rosbash, Ravi Allada, Patrick Emery, Carolyn Kotarski, Myai Le, Li Liu, Michael McDonald, Joan Rutila, Lea Sarov-Blat, Venus So, Vipin Suri, Howard Hughes Medical Institute, Brandeis University, Waltham, and NSF, Center for Biological Timing, USA
Genetic and Molecular analyses of the Drosophila circadian system identified the period and timeless proteins (PER and TIM) as clock molecules that contribute to circadian pacemaker function. Both genes show robust circadian rhythms of transcription, mRNA and protein expression. Furthermore, the two proteins interact to form a heterodimeric complex, and TIM levels respond to light, thereby tying the circadian pacemaker to photic stimuli. More recent work has identified four new Drosophila clock genes: Clock, cycle, doubletime, and cry. Clk and cycle are bHLH-PAS transcription factors. These heterodimeric partners drive PER and TIM expression, in the mammalian as well as the Drosophila system. doubletime is a kinase implicated in PER phosphorylation, and cry is a photoreceptor that connects the molecular clock components to the major environmental entraining stimulus, light. There is evidence that cry contributes to circadian light perception in mammals and plants as well as Drosophila and is thus the first clock molecule to cross the plant-animal boundary.
Despite this substantial recent progress, there are a large number of questions that remain unanswered. These include the feedback effect of PER and TIM on CLK and CYC activity, the post-transcriptional regulation of clock proteins, identification of the cells that drive locomotor activity rhythms, the multiple inputs that entrain these rhythms, and the outputs required for their detection. In the latter case, some of these operate far downstream of the transcriptional events that have been a major focus of the field.
New features of these issues, especially on the input side, will be considered. In particular, new experiments that address cry function will be presented. Constant light has been known for decades to alter clock properties in a wide range of organisms. Aschoff’s rule summarizes many of these observations and indicates that there are systematic period effects and arrhytmicity at high intensities. Drosophila is no exception, as shown by Konopka and colleagues many years ago. Remarkably, cry mutant flies are insensitive to intense constant light and manifest rhythmic activity essentially indistinguishably from wild-type flies in constant darkness. The wild-type arrhythmic phenotype can be rescued by expressing CRY only in the lateral neurons, indicating that these key brain pacemaker cells are important photosensitive centers for circadian rhythms. Finally, some evolutionary issues will be considered, in an attempt to provide a framework for considering together different circadian systems.
Molecular cycling of the Drosophila transcription factor vrille is required for behavioral rhythmicity
J. Blau and Michael W. Young, Rockefeller University, USA
Using differential display of expressed genes, we isolated vrille (vri), a transcription factor previously identified with a role in development. Intriguingly, vri shows homology to DBP, mammalian transcription factor that oscillates in the SCN. We show that vri is expressed in clock cells in both adult and developing flies and that vri RNA levels oscillate in synchrony with the period (per) and timeless (tim) genes in heads and bodies of wild type flies. vri RNA levels are constitutively low in Clock and cycle mutant files, suggesting that vri is a direct target of the of the dCLK/CYC heterodimer. The vri promoter contains a single functional dCLK/CYC binding site. Therefore, the same transcriptional loop that maintains 24 hour cycling of the per and tim genes also regulates other genes.
Flies with only one copy of the vri gene have a shortened period of behavioral rhythms. This suggested that vri may function as a clock component(s). To further examine the role of vri in clock function, a system was established for noncycling vri expression in Drosophila pacemaker cells. This constitutive expression eliminated cycling of per and tim expression in these tissues, and blocked behavioral rhythmicity. Our observations extend the similarities between Drosophila and mammalian clocks, and demonstrate the usefulness of alternative molecular approaches to the study of circadian control.
Putative role of the neuropeptide PDF in the circadian system of Drosophila
Charlotte Helfrich-Förster, Universität Tübingen, Jae Park, Brandeis University, Marcus Täuber, Universität Regensburg, Max Mühlig-Versen, Universität Regensburg, Stephan Schneuwly, Universität Regensburg, Alois Hofbauer, Universität Regensburg and Jeffrey C. Hall, Brandeis University
Neuropeptides are major components of the circadian systems of mammals and insects and may be involved in input or output pathways of the pacemaker cells. The most abundant neuropeptide in the insect pacemaker system is the pigment dispersing factor PDF —an ortholog of the crustacean pigment-dispersing hormone (PDH) family which regulates light-adaptive pigment responses. PDF is expressed in the putative circadian pacemaker neurons of all insects studied so far (reviewed by Helfrich-Förster et al. 1998, Chronobiol. Internat. 15, 567-594). There is also direct evidence for a role of PDF in insect circadian rhythmicity: In the cockroach Leucophaea maderae, injections of synthetic PDF into one optic lobe cause phase-shifts in the locomotor activity rhythm indicating that PDF acts as an input factor to the clock (Petri and Stengl 1997, J. Neurosci. 17, 4087-4093). In Drosophila melanogaster, PDF is so far the only neuropeptide found in the fly’s putative pacemaker cells. It is co-localized with the clock proteins PER and TIM in the small and large ventral Lateral Neurons (sLNv and 1LNv) – and it was hypothesized to be an output factor of the clock mediating circadian signals to follower neurons in the dorsal central brain.
To study the function of PDF in the circadian system of Drosophila we misexpressed PDF either from Drosophila or from the grasshopper Romalea in Drosophila’s central nervous system and investigated the effect of this on behavioral rhythmicity. With help of the GAL4 /UAS system PDF was either ectopically expressed in different numbers of neurons in the brain or the thoracical nervous system or it was overexpressed in the pacemaker neurons alone. Ectopic PDF-expression in the thoracical nervous system and in the neurosecretory cells of the pars intercerebralis had no effect on locomotor activity. On the other hand, ectopic PDF expression in neurons that projected into the dorsal central brain led to severe alterations in locomotor activity rhythms. Under LD-conditions such flies showed an advanced morning peak and a delayed evening peak. Under DD-conditions they were more active than the controls, had longer periods and frequently showed internal desynchronization into two freerunning components. These results indicate that PDF serves as circadian mediator and that it exerts its influence in the dorsal central brain.
Our study also revealed some important details of PDF regulation. Amazingly, overexpression of the pdf gene in the PER and TIM containing LNv alone, did not influence the rhythm of locomotor activity. This suggests that these clock proteins regulate PDF posttranscriptionally instead of transcriptionally and that additional pdf expression in these cells can not eliminate this rhythmic process. The following observations support this conclusion: (1) pdfmRNA levels did not cycle in the wildtype (Park and Hall 1998, J. Biol. Rhythms 13, 219-228) and were not altered in the per0 and tim0 mutants, (2) nevertheless, the amount of PDF varied rhythmically in the central brain terminals of the sLNv in the wildtype, (3) in the short period mutant pers, this PDF rhythm continued with a period about 20 hours under constant dark conditions (DD), (4) in the arrhythmic mutants per0 and tim0 the PDF rhythm was absent, (5) it was also lacking in the terminals of all ectopically pdf-expressing cells that do not express per and tim but it remained present in those of the sLNv after pdf overexpression. All these results indicate that PDF is controlled by the clock genes per and tim at the posttranscriptional level. In contrast to that, the clock genes clk, and cyc appear to control the pdf gene at the transcriptional level: almost no pdf mRNA could be detected in the sLNv of the arrhythmic mutant clkJrk and cyc0.
Interestingly, both clock controlled regulations are restricted to the sLNv which appear to be a particularly important component of Drosophila’s circadian system. In the 1LNv terminals, no PDF cycling was found and pdf mRNA levels were not altered in clkJrk and cyc0 mutants.
Functional and developmental analysis of the pacemaker neurons in the Drosophila brain
François Rouyer, Institut Alfred Fessard, Gif-sur-Yvette
The lateral neurons of the Drosophila brain are involved in the generation of circadian rhythms of locomotor activity. We have used the Gal4/UAS system to target the expression of several genes in these neurons and test their role in the clock function. The genetic ablation of the lateral neurons, induced by targeted expression of apoptosis genes, eliminates activity rhythms. A re-examination of the mostly arrhythmic disconnected mutant, using GFP as a cell marker, indicates that most of the flies do have lateral neurons that project in the dorsal brain but fail to express the period gene in these cells. The overexpression of the period gene in the lateral neurons abolishes activity and eclosion rhythms, demonstrating their requirement for both circadian functions. Conversely, a moderate, non-cycling expression of period in the lateral neurons is sufficient to restore PER protein cycling and 24 hrs behavioral rhythms in per0 flies.
Previous work has shown that a first set of PDF-expressing lateral neurons can be detected at the beginning of larval development and that a second set appears during pupation. We observe that the ablation of the larval neurons leads to strong defects in the morphology of the pupal neurons, suggesting that the larval cells are involved in the guidance of the projections of the adults cells. Surprisingly, we find that similar defects are induced by mutations in the dClock, and cycle genes, that encode transcriptional activators of period and timeless. These results indicate that dClock and cycle are involved in the differentiation of the pacemaker neurons during brain development. Clock cells alterations are also detected in the glass mutant. glass flies fail to develop photoreceptors and show a strong reduction of the dendritic arborization of lateral neurons. This defect could be responsible for the loss of rhythmicity that can be observed when the flies never see the light during their development.
Modeling the molecular regulatory mechanisms of circadian rhythms
Jean-Christophe Leloup, Didier Gonze, and Albert Goldbeter, Université Libre de Bruxelles
Thanks to genetic and biochemical advances on the molecular mechanism of circadian rhythms, theoretical models based on experimental observations can be considered for the regulatory mechanisms of circadian clocks. The organisms studied in most detail so far are Drosophila and Neurospora. The model for circadian oscillations in Drosophila takes into account the autoregulatory negative feedback exerted by a complex between the PER and TIM proteins on the expression of the per and tim genes. Sustained rhythmic variations in protein and mRNA levels occur in continuous darkness, in the form of limit cycle oscillations, with an endogenous period close to 24 h.
The effect of light on circadian rhythms is taken into account in the model by considering that light triggers TIM degradation. When incorporating the control exerted by light, the model accounts for the entrainment of circadian rhythms by light-dark cycles of various photoperiods and for the damping of the oscillations in constant light. The model also accounts for the phase shifts induced by light pulses and allows the construction of phase response curves. These compare well with experimental results obtained for Drosophila wild type and mutant files. Oscillations in the model can still occur when per expression remains constant, but stop when both per and tim expression escape regulation by the PER-TIM complex. The model also allows examination of the bases of temperature compensation.
Extensions of the model are discussed, which take into account the role of the clock and cyc genes, and the existence of post-transcriptional regulation. The model for Drosophila also shows that when applied at the appropriate phase, light pulses of appropriate duration and magnitude can suppress circadian rhythmicity in a transient or permanent manner. In the latter case, a second pulse of light can restore sustained oscillations. We investigate the effects of the duration and magnitude of light-induced biochemical changes to determine the conditions in which permanent suppression by a critical light pulse occurs.
The model for circadian rhythms in Neurospora is based on the negative feedback exerted by the FRQ protein on the expression of its gene frq. Here, in contrast to Drosophila, light controls the rhythm by inducing frq transcription. The model accounts for sustained oscillations in continuous darkness. It is possible, but more difficult than with the Drosophila model, to account for damped oscillations in constant light. The model yields phase relationships between total and nuclear FRQ and frq mRNA which can be compared with experimental observations.
Leloup J.C. & Goldbeter A. 1998. A model for circadian rhythms in Drosophila incorporating the formation of a complex between the PER and TIM proteins. J. Biol. Rhythms 13:70-87.
Leloup J.-C., Gonze D. & Goldbeter A. 1999. Limit cycle models for circadian rhythms based on transcriptional regulation in Drosophila and Neurospora. J. Biol. Rhythms (in press).
Gonze D., Leloup J.C. & Goldbeter, A. 1999. Theoretical models for circadian rhythms in Neurospora and Drosophila. C. 13 Acad. Sci. (Paris) III (in press).
The cellular organisation of the circadian system in higher plants
Andrew Millar, University of Warwick, UK
Organisms regulate light-dependent processes either by responding to the light environment, and/or by predicting the onset and duration of light with a timing mechanism. The circadian timing system controls biological rhythms with periods of approximately 24 hours. The expression of chlorophyll a/b-binding protein (CAB or LHCB) genes reflects both regulatory mechanisms: plants containing a cab2::luciferase fusion transgene exhibit both the ‘acute’ induction of luc activity by light signals (2-hour timescale) and a circadian rhythm. Circadian gating restricts the acute light response to the subjective day; other light-regulated processes in plants, such as hypocotyl elongation, may also be subject to circadian gating.
The phytochromes, a small gene family of red-light photoreceptors, mediate many developmental and cellular light responses in plants. We have recently shown that the gene encoding the major phytochrome in green tissue, phyB, is transcribed under circadian clock control in tobacco and Arabidopsis (Kozma-Bognar et al., submitted). PHYB protein synthesis is also rhythmic but the total abundance of PHYB protein bas at most a very weak rhythm.
Phytochromes including phyB control the period of the circadian clock (Millar et al., Science 1995; Somers et al., Science, 1998), so this work identifies an «outer loop». Its implications for the mechanism of the oscillator and of circadian gating will be discussed with respect to the light regulation of phytochrome nuclear transport.
We have used luciferase markers fused to CAB, CHS and PHYB promoters in order to monitor rhythms in several parts of an intact plant. Internal desynchronisation among organs can be initiated by light-dark cycles and is then maintained in a subsequent free-run, indicating that the oscillators are functionally independent in various plant organs. Both 12-hour and smaller phase differences can be maintained. Entrainment of one organ is ineffective in altering the phase of other, freerunning organs, indicating that the input signals are also not communicated. Physiological experiments have yielded conflicting results regarding the cell-autonomy of photoreceptor function in plants. Earlier data on the acute response of cab2::luc, however, indicated that an intercellular signal was transmitted from activated phytochromes within, but not between cotyledons (Bischoff et al., Plant Journal, 1997).
We have also exploited natural allelic variation to identify several novel loci that control the circadian period in Arabidopsis thaliana (Swarup et al., Plant Journal, Oct. 1999). We analysed the circadian rhythm of leaf movements in the accession Cvi from the Cape Verde Islands, and in the commonly-used, laboratory strains Columbia (Col) and Landsberg (erecta)(Ler), which originated in Northern Europe. The parental lines had similar rhythmic periods, but the progeny of crosses among them revealed extensive variation for this trait. An analysis of 48 Ler/Cvi recombinant inbred lines (RILs) and a further 30 Ler/Col RILs allowed us to locate four putative quantitative trait loci (QTLs) that controlled the period of the circadian clock. Near-isogenic lines (NILs) that contained a QTL in a small, defined chromosomal region allowed us to confirm the phenotypic effect and map positions of three period QTLs, designated ESPRESSO, NON TROPPO, and RALENTANDO. QTLs at the locations of RALENTANDO and of a fourth QTL, ANDANTE, were identified in both Ler/Cvi and Ler/Col RIL populations. Some QTLs for circadian period were closely linked to loci that control flowering time, including FLC. We showed that flc mutations shortened the circadian period, such that the known allelic variation in the MADS-box gene FLC can account for the ANDANTE QTL. The QTLs ESPRESSO and RALENTANDO identify new genes that regulate the Arabidopsis circadian system in nature, one of which may be the flowering-time gene GIGANTEA.
Genetic interactions between candidate components of the Arabidopsis circadian clock
Isabelle Carré, University of Warwick, UK
The molecular basis of the circadian clock of higher plants is still unknown. However, a number of loci that regulate its function have been identified by genetic approaches, in Arabidopsis thaliana. This abstract summarizes work in progress, to assign some of these genes to different components of the circadian clock (input pathway, central oscillator, and output pathways) and to attempt to order them into pathways.
The ELF3 (early flowering-3) gene is thought to regulate the light input pathway to the clock. Plants carrying null alleles of the elf3 mutation exhibit arrhythmic expression of a clockcontrolled reporter gene (cab:luciferase, or cab:luc) in constant light (LL), but not in constant darkness (DD) nor in light-dark cycles. The DET1 (deetiolated-1) and COP1 (constitutively photomorphogenic-1) encode negative regulators of lightregulated gene expression. Loss-of function mutations at these loci result in the constitutive activation of phototransduction pathways. The det1-1 and copl-6 mutants exhibit short-period rhythms of cab:luc expression in LL and DD, mimicking the effect of very bright light on the circadian light input pathway. To test whether the ELF3 gene product functions in the same phototransduction pathway as DET1 and COP1, we have constructed elf3-1 det1-1 and elf3-1 cop1-6 double mutants, carrying a cab:luc reporter gene. Luminescence rhythms were analysed by in vivo imaging using a photon-counting camera, under LL, DD, and under light-dark cycles. Both double mutants were rhythmic under light-dark cycles, with a phase distinct from the single mutant parents. The difference in phase persisted after transfer to DD, indicating that this alteration reflected a property of the circadian oscillator.
Surprisingly, the double mutants exhibited weak, but statistically significant rhythmicity under LL, suggesting that the effect of the det1 and cop1 mutations may partially suppress that of the elf3 mutation. The results indicated that the effects of the mutations were additive, therefore the ELF3 locus regulates clock function independently of DET1 and COP1.
The LHY (late elongated hypocotyl) gene encodes a component of an autoregulatory feed-back loop, which regulates circadian rhythms. LHY is part of a small family of rhythmically expressed, single domain MYB transcription factors, of which CCA1 (circadian clock-associated-1) is the most closely related. Plants overexpressing the LHY gene (LHY overexpressor mutant, lhyox) are arrhythmic in LL, and plants carrying null mutations at the LHY locus exhibit short period rhythms of CCAI and CCR2 gene expression. The short-period phenotype of lhy null mutants indicates that LHY does not simply function as a component of output pathways. LHY may function as a component of the light input pathway, since expression of its transcript is light-inducible. Our recent findings suggest that LHY activity may also be regulated post-transcriptionally in response to light-dark cycles. When grown under light-dark cycles, lhyox plants express constitutively high levels of LHY transcript. However, they exhibit cab:luc expression rhythms that are out-of phase with wild-type rhythms, and that appear to be driven (rather than entrained) by the light-dark cycle. We have also measured the expression of a clock-controlled gene that is not light-responsive: the CCR2 gene was expressed rhythmically, and 1800 out-of-phase with wild-type plants. Therefore, light-dark (and perhaps dark-light) transitions may resume partial clock function, and allow the completion of an incomplete cycle, in spite of the constant level of expression of LHY. The simplest hypothesis is that light-dark cycles cause the rhythmic posttranscriptional regulation of LHY activity, thus allowing the expression of overt rhythmicity.
To determine whether LHY functions within the same phototransduction pathway as either the DET1 or ELF3 gene, we have constructed lhyox det1-1 and lhyox elf3-1 double mutants. The lhyox det1-1 double mutant was arrhythmic in LL (as the lhyox mutant) indicating that a functional DET1 gene product is not required for the disruption of rhythmicity by LHY overexpression. The elf3-1 mutant has previously been shown to exhibit dramatically reduced levels of LHY transcript in LL. If the transcriptional down-regulation of LHY transcript levels mediates the effect of the elf3-1 mutation on the clock, the elf3-1 mutation should have no effect in plants that constitutively overexpress LHY. However, temporal patterns of cab:luc expression in the lhyox elf3-1 double mutant were distinct from both single mutant parents, suggesting that the effects of the two mutations were additive. Therefore, the effect of the elf3-1 mutation on LHY expression levels must be posttranscriptional, or else indirect.
In summary, our results suggest that the DET1, COP1, ELF3 and LHY genes do not function in a simple linear circadian light input pathway. Phototransduction pathways downstream of phytochrome have been shown to branch out into Ca2+ and cGMP dependent pathways. Furthermore, there is evidence for feed-back from the circadian oscillator onto its input pathway. Therefore, the analysis of genetic interactions between components of the circadian regulatory system may be more complex than initially imagined.
The roles of the transcription factor CCA1 and the protein kinase CK2 in circadian rhythms of plants
Elaine M. Tobin, Christos Andronis, Rachel M. Green, M.S. Ong, Shoji Sugano, U.C.L.A., USA
CCA1 is a Myb-related transcription factor that is involved both in phytochrome regulation of Lhcb gene expression (1) and in circadian regulation of gene expression and various physiological processes (2). It has been proposed that it may, along with the closely related LHY protein (3), be closely associated with a central oscillator in plants. We have isolated a line of Arabidopsis that has a T-DNA insertion in the CCA1 gene, and produces no CCA1 protein (4). These plants show an alteration of the circadian rhythm of gene expression for at least four circadian regulated genes and also show reduced phytochrome induction of Lhcb gene expression. The results are consistent with CCA1 being part of the oscillatory system and also suggest that there is another protein, possibly LHY, whose function is at least partially redundant with that of CCA1. We have found that CCA1 and LHY can interact in vitro, and that LHY can also bind to the Lhcbl*3 promoter at the same sites as CCA1.
Additional experiments suggested that the binding activity of CCA1 involves phosphorylation by the protein kinase CK2 (5), and we have recently found that overexpression of a CK2 subunit in transgenic plants affects the circadian rhythms of gene expression, demonstrating that CK2 has a function in the regulation of circadian rhythms in vivo.
Wang et al. (1997) Plant Cell 9: 491-507.
Wang and Tobin (1998) Cell 93: 1207-1217.
Schaffer et al. (1998) Cell 93: 1219-1229.
Green and Tobin (1999) Nat/. Acad. S ci. USA 96: 4176-4179
Sugano et al. (1998) Nati _Acad. Ki. USA 95: 11020- 11025.
mPerl, mPer2 and mPer3 in mammalian clock oscillations
Hitoshi Okamura, Kobe University School of Medicine
The mouse period and timeless genes mPer1, mPer2, mPer3 and mTim are structural homologues of the drosophila period or timeless, and all four were strongly expressed in the mammalian circadian center, the suprachiasmatic nucleus (SCN). The pattern of expression of mPer1, mPer2 and mPer3 are characterized by a day time peak and night time trough in both light-dark and constant dark conditions. mTim, which product mTIM binds to mPER1 inside the nucleus, did not show clear clear circadian rhythm in the SCN. Light elicited induction was also observed in mPer1 and mPer2, and this phenomenon was one of the difference between Drosophila and mammals.
In Drosophila, nuclear entry of negative elements of circadian feedback loop is a key step for the generation of 24 hour cycle. We have examined the factors regulating nuclear entry of mammalian period gene family in COS7 cells, and here report that the physical interactions with mPER3 are critical for the nuclear translocation of mPER1 and mPER2, and this movement was accelerated by a brief exposure of high concentrations of horse serum. mPER3 has a functional cytoplasmic localization domain (CLD) in its middle portion and nuclear localization signal (NLS) in its C-terminal portion, suggesting that the interplay of a CLD and NLS that regulates nuclear entry of PER in Drosophila may conserved in mammals, but with the novel twist that mPER3 may perform as a dimerizing partner.
Among three mammalian per homologues, evidences are accumulating that mPer1 is a clock oscillating gene. Studies of the positive and negative regulatory mechanisms of mPer1 gene in mammalian cells are likely to further elucidate the nature of the circadian feedback loop at the molecular level and deepen our understanding of mammalian circadian biology.
Control of clock function in zebrafish
Nicholas S. Foulkes and Paolo Sassone-Corsi, Institut de génétique et de biologie moléculaire et cellulaire (IGBMC), Strasbourg
The cloning of the first circadian clock gene in a vertebrate, the mouseclock gene, was the result of a large scale mutant screen. This work confirms the power of the genetic screening approach to identify the molecular components of the circadian clock. A potential alternative vertebrate model system for genetic screening analysis is the zebrafish (Danio rerio). The zebrafish has become a powerful model system to study early vertebrate development. For these types of studies, the zebrafish has the advantage that its early development is extremely rapid (from fertilization to hatching from the egg in 3 days) and that the embryos are transparent so they can be easily observed. However, a variety of techniques such as transgenesis and genetic screening are now being developed rapidly for zebrafish, so making it a useful model to explore many other biological processes.
The zebrafish potentially provides two advantages for studying the circadian clock. Firstly, it constitutes an alternative genetic model system to the mouse to identify new clock molecules via mutant screening. Secondly, it represents an ideal tool to examine the origin and role of the clock during early, embryonic development. Originally, it was demonstrated that the eyes and the pineal gland of this fish contain an endogenous circadian clock capable of directing rhythmic synthesis of the hormone, melatonin, in organ culture. In addition, circadian rhythms in the locomotor activity of zebrafish adults and larvae have been described. However, little is known about clock function in the zebrafish at the molecular level.
As a first step to develop the zebrafish as a model to study the circadian clock, we chose to characterize the zebrafish homologue of the mouse clock gene. Low stringency screening of a zebrafish embryo cDNA library resulted in the cloning of a zebrafish homolog of the clock gene, that shows a high degree of similarity to the mouse counterpart. RNase protection analysis revealed that the clock transcript oscillates with a pronounced circadian rhythm in the eye and pineal gland of the adult fish. This is in sharp contrast to the situation in the mouse, where clock expression shows no day-night changes. Further examination revealed that clock mRNA also oscillates with a circadian rhythm in many tissues within the adult fish. This observation raises the possibility that each tissue may contain its own circadian oscillator, or that these peripheral oscillations are driven from a central «master» clock. To answer this question, we placed several tissues into culture in constant darkness, and, at a number of phases, RNA was extracted and the level of clock transcript was assayed. The oscillation of clock observed in vivo was also clearly apparent in vitro, in the absence of any timing cues. We can conclude that the heart and kidney of the zebrafish do, in fact, contain an endogenous circadian oscillator.
Now, in possession of an oscillating clock component in the zebrafish, we envisage many possible lines of study. Defining the mechanisms driving rhythmic clock expression as well as the partners which interact with the clock protein may potentially lead to the identification of new clock components or regulatory systems. Furthermore, exploiting the clock promoter to drive expression of a fluorescent reporter gene in the context of a stable transgenic fish, could provide a valuable tool for a full-scale mutant screen.
Whitmore, D., Foulkes, N.S., Strähle, U. and Sassone-Corsi,
P., Zebrafish Clock rhythmic expression reveals independent peripheral circadian oscillators. Nature neurosci. 1, 701-707 (1998).
King, D.P. et al., Positional cloning of the mouse circadian clock gene. Cell 89, 641-653 (1997).
Antoch, M.P. et al., Functional identification of the mouse circadian Clock gene by transgenic BAC rescue. Cell 89, 655667 (1997).
Vitaterna, M.H. et al., Mutagenesis and mapping of a mouse gene, Clock, essential for circadian behavior. Science 264, 719725 (1994).
Cahill, G.M., Circadian regulation of melatonin production in cultured zebrafish pineal and retina. Brain Res. 708, 177-181 (1996).
Cahill, G.M., Hurd, M.W. and Batchelor, M.M., Circadian rhythmicity in the locomotor activity of larval zebrafish. Neuroreport. 9, 3445-3449 (1998).
Relevance of circadian rhythms for the pharmacology of anticancer drugs. In vitro rhythms of bone marrow proliferation
Francis Lévi, Université de Paris-Sud, Hôpital Paul Brousse
Anticancer drugs are toxic for both normal and cancer cells. This lack of specificity can be improved by selecting the dosing time of these agents along the 24-h time scale. Thus, the tolerability of mice or rats for over 30 anticancer drugs vary 2 to 10-fold as a function of the circadian time when they are given. In murine tumor models synchronized with LD 12:12, the circadian time of best tolerability usually coincides with that of best antitumor efficacy. The circadian rhythm dependency of antitumor efficacy most often results from ability to increase the maximum tolerated dose at the least toxic time, as shown for vinorelbine in P388 leukemia-bearing mice and with docetaxel, a taxane derivative, in P03 pancreatic adenocarcinoma-bearing mice. Nevertheless, some tumor models have retained circadian rhythmicity, yet best efficacy is still achieved following the administration of non-toxic doses near the time of best tolerability as shown in mice for oxaliplatin against Glasgow osteosarcoma and for docetaxel or doxorubicin against MA13/C mammary adenocarcinoma.
Several cellular and biochemical rhythms contribute to anticancer drug chronopharmacology. For instance, the least toxicity of the antimetabolite drug 5-fluorouracil (5-FU) results from the circadian rhythms which modulate. 1) dehydropyrimidine dehydrogenase (DPD) activity, an enzyme which catabolizes the drug, hence protects the cells against cytotoxic damage, 2) thymidine kinase, which leads to the incorporation of 5-FU into DNA and 3) thymidilate synthase, a target enzyme of 5-FU, which is required for DNA synthesis. Recent data indicate that both DPD and TS activity rhythms are associated with transcription rhythms, yet the relationship between both of these rhythms need clarification. Other relevant rhythms which modulate drug cytotoxicity include reduced glutathione, which protects cells against the toxicity from alkylating agents and Platinum complexes, O6- alkylguanine transferase activity, a protein which repairs DNA lesions resulting from nitrosourea exposure. Finally, most anticancer drugs are toxic for dividing cells, yet the proportions of them which are in a given phase of the tell cycle (G1, S, G2, M) also vary according to a circadian rhythm. The latter also contribute to the changes in anticancer drug toxicity. Both enzymatic and cell cycle related rhythms have been found in human tissues and display a phase relationship with the restactivity cycle similar to that found in most mice or rats.
Since bone marrow granulomonocytic (GM) progenitors represent a target tissue for most anticancer drugs, we have investigated the ability of these cells to maintain an in vitro circadian rhythm in the extent of response to growth factors. A circadian rhythm in the count of GM-colony forming units was demonstrated over the 72-96 h which followed liquid culture initiation, with a maximum occurring near the projected circadian time of the in vivo maximum. Other data in rats kept in constant light, with suppressed rest-activity and body temperature rhythms and preliminary data in SCN lesioned rats further support the persistence of coordinated cellular rhythms in the absence of an effective central pacemaker.
The clinical relevance of the chronotherapy principle has been tested and validated in clinical trials involving nearly 1500 patients with metastatic colorectal cancer. In these studies, all the patients received the same chronomodulated chemotherapy regimen, which was compared to constant rate infusion or to schedules with different peak times of drug delivery. Tolerability was improved up to 5-fold, concurrently with a near doubling of antitumor activity. Ongoing trials in Europe and in Canada (European Organization for Research and Treatment of Cancer – E.O.R.T.C.) are currently assessing the relevance of chronotherapy for enhancing survival. Nevertheless, individual rest-activity cycle monitoring and cortisol rhythm estimation in 200 patients with colorectal cancer indicate that nearly 1/3 of them display rhythm alterations, which are independent prognostic factors of both tumor response and survival. Specific treatment measures may result from a better understanding of the mechanisms of cancer-associated circadian alterations.
Peripheral mammalian clocks and their entrainment
Ueli Schibler, Aurelio Balsalobre, Juergen Ripperger, Steve Brown, Francesca Damiola, Nicholas Preitner, Minh Nguyet Le, University of Geneva
DBP, TEF, and HLF are related transcription factors of the PAR basic leucine zipper (PAR bZip) protein family. All three of these proteins as well as the mRNAs encoding them accumulate according to robust circadian rhythms in liver and other peripheral tissues. PAR bZip mRNAs also oscillate in the suprachiasmatic nucleus (SCN) of the hypothalamus, thought to contain the central circadian pacemaker. Furthermore, genetic loss-of-function experiments in transgenic mice suggest that all threee PAR bZip proteins contribute to the determination of circadian period length (tau). However, mice homozygous for null alleles for these proteins display rhythmic locomotor activity and robust circadian gene expression in peripheral cells. Therefore, PAR bZlP family members are players of circadian output pathways, rather than central clock components. Since the expression of PAR bZip and mPer genes oscillates with a similar phase angle, we considered that the transcription or these genes is coordinately regulated. This has been verified for dbp by both genetic and biochemical studies. The mapping of DNAse 1 hypersensitive sites at different day times has revealed four putative enhancer regions, of which two are located downstream of the cap site. These putative downstream enhancers are likely to play important roles, since a mutant allele in which most intragenic sequences had been deleted and replaced by a lacZ reporter gene is essentially silent. By using a two-dimensional electrophoretic mobility shift assay (2d-EMSA) we have demonstrated that Clock, a transcriptional activator protein essential for pacemaker function, binds to an E-box in the first intron. Moreover, DBP mRNA accumulation is essentially abolished in mice expressing a dominant negative Clock mutant protein. Together with previously published data, these experiments suggest that circadian transcription of dbp and mper1 is controlled by the same mechanisms and hence that dbp expression is directly hardwired to the negative feedback loop generating daily oscillations in the expression of central clock genes.
Rev-Erba, a nuclear orphan receptor of the thyroid hormone receptor family binds avidly to two recognition sequences within the dbp promoter. Since the accumulation of Rev-Erba parallels circadian transcription of dbp, we considered this orphan receptor as an additional regulator of circadian dbp expression. To examine this conjecture, the rev-erba gene was inactivated by homologous recombination, and dbp expression was monitored in rev-erba mutant mice. Unexpectedly, these experiments demonstrated that Rev-Erba contributes little if anything to circadian dbp transcription under LD or DD conditions. However, Rev-Erba turned out to be an important regulator of dbp expression in severely starved animals. Thus, in wild type mice kept for more than 36 hours without food, DBP mRNA and protein accumulation is reduced to diurnal trough levels in wild type mice, but remains high in homozygous rev-erba mutant animals. Hence, Rev-Erba is required for the repression of dbp expression in starving animals. As DBP increases energy expenditure by about 15 to 20% (L. Lopez-Molina and U.S., unpublished observation), the repression of dbp during starvation may be physiologically meaningful.
The mRNAs issued by the putative mammalian clock genes per1, per2, per3, and cry-1, as well as those encoded by the clockcontrolled genes dbp, tef, hlf and rev-erba, all oscillate in SCN neurons with an about four-hour phase advance as compared to peripheral cell types (e.g. hepatocytes). We thus suspect that the SCN determines the phase-setting of peripheral clocks via humoral cues rather than through neuronal connections. The existence of true peripheral clocks has recently been demonstrated unambiguously. Thus, we have established rat and mouse tissue culture systems that mimic the circadian expression of genes observed in intact animals. After treatment of immortalized fibroblasts with high concentrations of serum, the levels of the mRNAs encoding Rev-Erba DBP, TEF, Per1 and Per2, cycle for several consecutive days in a circadian fashion. Biochemical fractionation of serum factors revealed multiple activities that can induce circadian rhythms in tissue culture cells. Likewise, several chemicals known for inducing different signal transduction pathways are capable of triggering circadian gene expression in vitro. These include forskolin, butyryl cAMP (both inducing the cAMP signaling pathway), the tumor promoter PMA (activating protein kinase C), EGF (activating a membrane tyrosine kinase receptor), and glucocorticoid hormones (directly operating through a nuclear receptor). We considered glucocorticoids to be particularly attractive candidates for resetting peripheral clocks, since their serum levels oscillate during the day and since the glucocorticoid receptor is expressed in virtually all cell types (except SCN neurons). Our preliminary experiments show that glucocorticoids can indeed influence several parameters (phase, amplitude, magnitude) of circadian gene expression in vivo, but that they are not required for determining the phasing in peripheral organs under normal LD and DD conditions. The large variety of signal transduction pathways capable of inducing circadian gene expression in tissue culture cells may indicate that the resetting of peripheral clocks can be controlled by many different chemical cues. Moreover, it is conceivable that several different signals may act in a synergistic fashion.