DOI: 10.1093/jxb/erf024

REVIEW ARTICLE

Signs of the time: environmental input to the circadianclock

Paul F. Devlin1

Division of life Sciences, King's College London, Franklin-Wilkins Building, 150 Stamford Street,London SE1 8WA, UK

Received 25 September 2001; Accepted 18 April 2002

Abstract

The circadian clock forms one of the most fascinat-

ing adaptations to life on earth. Organisms can not

only anticipate the day/night cycle but can make use

of an internal clock to measure daylength as an indi-

cator of the changing of the seasons. The innate per-

iod of the clock is not exactly equal to 24 h, but is

reset each day by environmental signals at dawn and

dusk, most notably by changes in light and tempera-

ture. This ability to re-entrain also ensures that the

clock is synchronized with the day/night cycle which

in turn is crucial for anticipation of dawn and dusk.

Recent advances in the ®eld have identi®ed the

photoreceptors involved in resetting the clock in sev-

eral systems. This has revealed surprising similar-

ities, but also key differences in the circadian

systems of plants, fungi, insects, and mammals. One

recurring feature emerging from this research is that

the photoreceptors themselves are under the control

of the clock with transcript abundance being tightly

regulated. Furthermore, elements of a feedback path-

way whereby the clock modulates the activity of the

light input pathway are now being identi®ed.

Key words: Circadian clock, environmental input,

photoreceptors, temperature.

Introduction

All organisms properly tested show a rhythm of metabol-ism, of physiological processes or even of behaviour intune with the day/night cycle of the earth. In plants, thecomponents of the photosynthetic machinery accumulateeach day in anticipation of dawn; leaves or ¯owers oftenclose up just before dusk to provide protection for more

delicate tissues from the lower temperatures of night(Darwin, [1895] 1981; Enright, 1982). In insects, the adult¯ies eclose from their pupae in synchrony with dawn toprovide the maximum chances of survival (Pittendrigh,1954), in mammals such as ourselves our body has arhythm of alertness and of temperature that causes us to beactive during the day and to sleep at night (Moore-Edeet al., 1982). These phenomena are not purely responses tothe external environment and will continue even in theabsence of any external cues. A plant maintained inconstant light will still show a cyclic production of itsphotosynthetic machinery (Millar and Kay, 1991), ahuman deep underground will continue to wake andsleep with an approximately 24 h rhythm as if stillexperiencing dawn and dusk (Luce, 1971; Sulzman, 1983).Such processes are controlled by an endogenous oscillatorknown as the circadian clock which continues to cycle,maintaining its own, approximately 24 h rhythm, as aresult of the oscillation of the levels or activity ofmolecules within the cells of each organism. The mol-ecules making up this core clock and their biochemicalinteractions have been well studied and in insects andmammals the components of this basic `clock mechanism'are now known. These are discussed in more detail byWager-Smith and Kay (2000). In plants, the clockmechanism remains more elusive, but several recentadvances have begun to shed light on this (Carre andKing, 2002).

The circadian clock does not run in isolation from thecycle of day and night. It must, itself, be set to the correcttime. The clock must include a resetting mechanism bywhich it can ®rst be synchronized with the day/night cycleso that the organism can correctly anticipate dawn anddusk. This clock resetting is a phenomenon very familiar toany travellers on long-haul ¯ights. When a person travelsthrough several time-zones, initially jet-lag is experienced

1 Fax: +44 (0)20 7848 4500. E-mail: paul.devlin@kcl.ac.uk

ã Society for Experimental Biology 2002

Journal of Experimental Botany, Vol. 53, No. 374, pp. 1535±1550, July 2002

whereby his/her circadian clock remains set to the timingof dawn and dusk in the place of departure though,gradually, over the course of a few days he/she ®nds his/her rhythm of sleep and wake has adjusted to the newtiming of dawn and dusk.

The two most prevalent environmental cues which act asZeitgebers (time givers) are the changes in light andtemperature occurring at dawn and dusk and both arecapable of resetting the clock (Edmunds, 1988;Roenneberg and Foster, 1997). The circadian clock canbe thought of in terms of three major components: an`input' pathway by which environmental cues act tosynchronize the clock; the endogenous `oscillator' itself;and the `output' pathway whereby the rhythmic metabolicsystems controlled by the clock are co-ordinated (Fig. 1).This review focuses on the input to the clock. More detailson clock output can be found in three recent papers byHarmer et al. (2000), McDonald et al. (2001) and Duf®eldet al. (2002), focusing on plants, insects and mammals,respectively. However, one caveat should be added to thisseparation in that it is now apparent that this traditionalthree component model of the circadian system is over-simplistic and it is becoming increasingly more dif®cult toconsider any one of these components in isolation. For

example, it has been clear for some time that the outputpathway modulates (gates) the sensitivity of the inputpathway (Fig. 1). Consequently, this review takes a fairlyholistic approach in describing the input pathway.

Light input to the circadian clock

Light forms the dominant signal in resetting the clock anda close link between the circadian photoreceptors and theclock itself has been demonstrated for several systems(Devlin and Kay, 2001). In fact, the distinction betweeninput, oscillator and output is becoming more and moreblurred as more is discovered about the mechanism itself.

In resetting, the clock exhibits a change in phase. Thatis, the hands of the circadian clock are moved forward orbackwards. In terms of the molecular mechanism of theclock this would represent a change in the level or activityof a clock component to a level or activity that wouldnormally be found at a different point in the cycle(Crosthwaite et al., 1995). The clock then continues asbefore. For most organisms the period of the circadiancycle is not quite 24 h and they show a slight resetting witheach dawn and/or dusk. This plasticity allows an organismto adjust continually to changing daylength as the seasonsof the year progress. The response of the clock to light isdifferent at different times of day. The onset of light priorto the expected dawn will generally cause an advance inthe phase of the rhythm whilst extension of the light periodafter the expected dusk will generally cause a delay in thephase of the rhythm. It is possible to produce a `phaseresponse curve' (PRC) illustrating this effect (Johnson,1990). The effects of light pulses at different times of dayon the phase of the rhythm are examined for organismsotherwise maintained in darkness. An example of a typicalphase response curve is shown in Fig. 2. The curve shows astrong phase-advance response around subjective dawn

Fig. 1. The circadian system allows an organism to anticipate the day/night changes in its surrounding environment. The central oscillatormaintains an endogenous approximately 24 h rhythm even underconstant environmental conditions and this controls a series of overtrhythms within an organism via a series of output pathways. To beuseful to an organism, its clock must be set to the right time. Thecentral oscillator is entrained to the external day/night cycle via inputpathways by which environmental signals of dawn and dusk aretransmitted to the oscillator. It is now apparent that these inputpathways are, themselves, subject to regulation by the clock, in thatlight signalling is modulated (gated) by the clock output pathway overthe course of the day.

Fig. 2. Typical phase response curve. Light signals prior to dawncause phase advances (positive values), setting the clock to an `earliertime'. Light signals after dusk cause phase delays (negative values)setting the clock to a `later time'.

1536 Devlin

and a strong phase-delay response around subjective dusk.The response to light is greatly reduced during thesubjective day, as might be expected, when the organismwill normally be `seeing' light. For many organisms thePRC exhibits a `dead-zone' during the subjective daywhere no response at all is seen to light (Johnson, 1990).

Entrainment of the circadian clock by such pulses oflight is known as non-parametric entrainment as opposedto parametric entrainment to day/night cycles. Althoughthe clock is sensitive to pulses of light for resetting, it isimportant that the clock should not be unduly in¯uencedby aberrations in the environment. It must not, forexample, be reset by a ¯ash of lightning at night. Ingeneral, a prolonged pulse of irradiation is required to resetthe clock (Nelson and Takahashi, 1991).

The photoreceptors involved in the perception of lightleading to resetting of the circadian clock have been thesubject of several recent advances in the ®eld of circadianbiology. Both plants and animals have a complex array of

photoreceptors (Roenneberg and Foster, 1997; Whitelamand Devlin, 1998; Foster, 1998; Hall, 2000). Thephotoreceptors in plants are particularly well character-ized. Light plays a key role in the development of a plant,controlling processes from germination, through seedlingestablishment, to determining the whole architecture of theplant (Kendrick and Kronenberg, 1994). Plants areexquisitely sensitive to small changes in the light envir-onment in order to be able to adapt to take maximumadvantage of the available light for photosynthesis. The®rst identi®cation of the photoreceptors involved inresetting of the circadian clock came from research inthe model plant, Arabidopsis thaliana where two familiesof photoreceptors, the red-absorbing phytochromes and theblue-absorbing cryptochromes combine to relay dawn anddusk signals to the endogenous oscillator (Somers et al.,1998a). In insects and mammals, the blue light photo-receptor, cryptochrome was also discovered and demon-strated to play a key role in the circadian clock. In insects,

Fig. 3. Time-course showing the circadian rhythm of bioluminescence in a population of Arabidopsis thaliana seedlings expressing the ®re¯yluciferase gene under the control of the Arabidopsis CAB2 promoter. The time-course follows the rhythm of transcription from the CAB2 promoterover 48 h. Two circular 10 cm Petri dishes of seedlings are represented for each time-point. Twenty-®ve images are shown, taken at 2 h intervalsover the course of 2 d and should be followed as if reading a book, from left to right, one line at a time. The seedlings, germinated on growthmedium were ®rst entrained 12/12 h light/dark cycles for 6 d then transferred to constant light. The sequence commences during the morning ofthe ®rst day after transfer to constant light. The CAB2 gene encodes part of the photosynthetic machinery and appropriately shows a peak ofexpression during the subjective day and a trough of expression during the subjective night. Images were taken using a NightOWL cooled CCDcamera, Berthold Technologies, Cambridge, UK.

Input to the clock 1537

cryptochrome is the main circadian photoreceptor althoughit has also been proposed to play a role in the oscillatormechanism in peripheral tissue (Emery et al., 1998, 2000;Stanewsky et al., 1998; Krishnan et al., 2001). Inmammals, the main role of cryptochrome is as acomponent of the central clock mechanism itself (Kumeet al., 1999; (van der Horst et al., 1999). The photo-receptors involved in the perception of light in resetting theclock remain elusive. It is known, however, that the eyesare essential for entrainment in mammals (Foster, 1998)yet several pieces of research have ruled out the involve-ment of the visual opsins (Freedman et al., 1999).

Circadian photoperception in plants

The earliest recorded observation of a circadian rhythmwas made by Androsthenes (historian of Alexander theGreat) around 400 bc. He observed that leaves of severaltree species exhibit a horizontal position during the dayand a more vertical position at night. However, it was in1729 that the French astronomer, Jean Jacques d'Ortous deMairan ®rst demonstrated that this was the result of theaction of an endogenous circadian clock. Mimosa, thesensitive plant, folds up its leaves at night and opens themagain in the day. De Mairan showed that this rhythm,continued even when the plants were placed in deep shadedemonstrating that bright sunlight was not required totrigger this response (de Mairan, 1729).

In plants, a large range of physiological processes arecontrolled by the circadian clock including rhythmic leafmovements (Engelmann et al., 1994; Millar et al., 1995a),photoperiodic induction of ¯owering (Devlin and Kay,2000b) and stomatal opening (Somers et al., 1998b).However, much of the work studying the circadian clock inplants has involved the analysis of the rhythm of expres-sion of the clock-controlled gene, light-harvesting chloro-phyll a/b protein (better known as the chlorophyll a/bbinding protein, CAB). CAB forms part of the photosyn-thetic machinery of the plant and, as would be predicted,the CAB transcript begins to accumulate prior to dawn,shows a peak of expression in the early part of the day then

decreases again towards dusk (Millar and Kay, 1991).Millar et al. (1992) attached a ®re¯y luciferase reportergene (LUC) to the CAB promoter and transformed thisCAB::LUC construct into plants. Using a highly sensitivephoton-counting camera they were able to follow therhythmic expression pattern of CAB in living seedlings byfollowing the transient bioluminescence of the ®re¯yluciferase produced as a result (Fig. 3).

This system allowed the isolation of several circadianclock mutants in A. thaliana (Millar et al., 1995a). One ofthese was the result of a mutation in the gene, TIMING OFCAB 1 (TOC1), that encodes a pseudo-response regulatorinvolved in the central clock mechanism in A. thaliana(Strayer et al., 2000). The toc1 mutation affects all knownobservable rhythms in A. thaliana even in constantdarkness and the TOC1 message, itself, displays circadianoscillation with a peak at subjective dusk, thus demon-strating its credentials as a central clock component(Somers et al., 1998a, b; Strayer et al., 2000). Othermethods looking at the transcription factors responsible forregulation of the CAB gene identi®ed two other centralclock components, LATE ELONGATED HYPOCOTYL(LHY) and CIRCADIAN CLOCK ASSOCIATED 1(CCA1), both MYB-type transcription factors (Wang andTobin, 1998; Schaffer et al., 1998). Mutations in cca1 andlhy, likewise affect all known observable rhythms in A.thaliana and the CCA1 and LHY transcripts oscillate with acircadian rhythm showing a peak of expression atsubjective dawn (Wang and Tobin, 1998; Schaffer et al.,1998). It was recently demonstrated that TOC1 is respon-sible for the positive regulation of CCA1 and LHYexpression, whilst both LHY and CCA1 bind to theTOC1 promoter for the negative regulation of TOC1expression (Alabadi et al., 2001). This loop (represented inthe `Oscillator' section of Fig. 4) is critical for clockfunction in A. thaliana and is proposed to form thefundamentals of the central circadian oscillator in A.thaliana.

The CAB::LUC reporter system was also used to analysethe light input pathway to the plant circadian clock. Millaret al. (1995b) demonstrated that the A. thaliana circadian

Fig. 4. A detailed plan of the circadian system in Arabidopsis thaliana. Critical to the maintenance of a circadian rhythm is a feedback loop madeup of three proteins: TIMING OF CAB 1 (TOC1), CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) AND LATE ELONGATED HYPOCOTYL(LHY). Levels of TOC1 oscillate with an approximately 24 h rhythm, peaking in the evening. Levels of CCA1 and LHY oscillate with anapproximately 24 h rhythm, peaking in the morning. CCA1 and LHY proteins bind to an `evening' element (AAAATATCT) in the TOC1promoter negatively regulating transcription of the TOC1 gene. Conversely, TOC1 protein positively regulates transcription of the CCA1 and LHYgenes. As levels of TOC1 protein rise, they promote transcription of LHY and CCA1. Then, as levels of CCA1 and LHY proteins subsequentlyrise, they inhibit transcription of TOC1, thereby negatively feeding back on their own transcription. Light input to the A. thaliana circadian clockis via phytochrome (PHY) and cryptochrome (CRY) photoreceptors. PHYB acts directly to promote transcription of CCA1 and LHY byassociating with PHYTOCHROME INTERACTING FACTOR (PIF3) bound to a `G-box' element (CACGTG) in the CCA1 and LHY genepromoters. Light signalling via PHYB and CRY1 involves direct association with a common circadian light input signal transduction component,ZEILUPE (ZTL). Output from the clock potentially regulates these input pathways via two methods. Transcription of the PHYB, CRY1 and CRY2genes is directly clock regulated, whilst sensitivity to light input signals is periodically regulated by the clock controlled gating factor EARLYFLOWERING 3 (ELF3). The CCA1 and LHY proteins are proposed to act in the output pathway regulating a range of clock controlled outputgenes. CCA1 and LHY positively regulate CHLOROPHYLL A B BINDING PROTEIN (CAB) and other `morning genes' by binding to`morning' promoter elements (AAAAATCT). CCA1 and LHY negatively regulate `evening genes' by binding to `evening' promoter elements(AAAATATCT).

1538 Devlin

Input to the clock 1539

clock ran faster in continuous light than in darkness. Theresponse of an endogenous oscillator to continuous light isthe sum of the phase advances and phase delays occurringthroughout the circadian cycle. As light intensity increases,the magnitude of these phase shifts increases (Aschoff,1979). The overall effect depends on the shape of the phaseresponse curve. In diurnal organisms, such as A. thaliana,phase advances predominate over phase delays with theresult that, in continuous light, increasing light intensitytends to lead to a shortening of the circadian period length.In nocturnal organisms, phase delays predominate overphase advances with the result that increasing lightintensity tends to lead to a lengthening of period length.This phenomenon has become known as Aschoff's rule(Aschoff, 1979). Millar et al. (1995b) showed that both redand blue light were capable of causing a shortening of theCAB::LUC rhythm in A. thaliana.

Photoreceptors involved in light input to thecircadian clock in plants

In addition to chlorophyll, which carries out light harvest-ing for photosynthesis, plants possess several informationgathering photoreceptors that allow them to regulate theirdevelopment to take maximum advantage of their lightenvironment. These fall into three families, the phyto-chromes, absorbing red and far red light and thecryptochromes and the phototropins, absorbing blue andUV wavelengths (Whitelam and Devlin, 1998; Christieand Briggs, 2001). The phytochrome family consists of®ve members in A. thaliana, phyA to phyE (Sharrock andQuail, 1989; Clack et al., 1994). All share the same basicstructure, comprising a protein moiety of about 124 kDaand a covalently attached linear tetrapyrrole chromophore(Quail, 1991). Phytochromes exist in two photo-inter-convertible forms, a red-absorbing, Pr form and a far-red-absorbing Pfr form (Quail, 1991). Phytochrome issynthesized in the inactive Pr form and upon absorptionof a photon of light is converted to the active Pfr form thatis responsible for regulating a range of physiological anddevelopmental responses. PhyA is light labile and accu-mulates to high levels in etiolated seedlings (Quail, 1991).PhyA acts most prominently in germination and inseedling establishment and triggers a response to verylow levels of light (Whitelam et al., 1993; Johnson et al.,1994). It is rapidly degraded in high ¯uences of red lightand hence phyA is thought of as an `antenna' detecting thesmall amount of light penetrating through the upper layerof soil indicating that a seedling is about to emerge intolight (Yanovsky et al., 1995). PhyB±phyE are light stableand are generally involved in responses to higher ¯uencesof red light (Hirschfeld et al., 1998; Whitelam et al., 1998).It is noticeable that the phytochromes also show a peak ofabsorption in the blue region of the spectrum and,consistent with this, phytochrome A has been implicatedin some blue light responses (Whitelam et al., 1993).Suf®cient phyA Pfr is formed in blue light to trigger aresponse.

The cryptochrome family in A. thaliana comprises twomembers, cry1 and cry2 (Devlin and Kay, 1999).Cryptochromes show a strong resemblance to the photo-lyases involved in the blue/UV light-dependent repair ofDNA damage (Cashmore et al., 1999). The N-terminalsection of each of the A. thaliana cryptochromes shares astrong homology with type II photolyase and, like thephotolyases, cryptochromes bind two chromophores, alight-harvesting pterin and a catalytic ¯avin (Malhotraet al., 1995; Lin et al., 1995). Cry1 is light stable whilstcry2 is light labile and is degraded under high ¯uences ofblue light (Lin et al., 1998). Cry2 has been demonstrated tobe involved in seedling establishment in low ¯uence bluelight whilst cry1 shows an involvement in such responsesat both low and high ¯uences of blue (Lin et al., 1998).

Fig. 5. (A) The action of cryptochromes in light input to theArabidopsis thaliana circadian clock. The response of the cry1 mutantto increasing ¯uence rates of blue light reveals a triphasic action forcry1. At low ¯uence rates of blue light cry1 acts as a signaltransduction component downstream of phytochrome A. Atintermediate ¯uence rates the light-labile phyA is degraded and playsno role and cry1 and cry2 act redundantly as photoreceptors for bluelight input to the clock. At high ¯uence rates the light-labile cry2 isdegraded and no longer plays a role and cry1 acts alone as aphotoreceptor for blue light input. This plasticity in recruitment ofdifferent photoreceptors over a range of ¯uence rates may allow theplant to regulate its sensitivity to light depending on the external lightenvironment. (B) Light input to the Arabidopsis thaliana circadianclock. The red light photoreceptors, the phytochromes and the bluelight photoreceptors the cryptochromes mediate distinct red and bluelight signalling pathways to the clock. The light labile phyA detectslow ¯uence rates of both red and blue light, showing a functionaldependence on cry1 which does not itself act as a photoreceptor underthese conditions. The light stable phytochromes, phyB, phyD andphyE detect higher ¯uence rates of red light whilst the cryptochromesact as photoreceptors for higher ¯uence rates of blue light.

1540 Devlin

Somers et al. (1998a) demonstrated that both phyto-chromes and cryptochromes contribute to light input to thecircadian clock in A. thaliana. By crossing the CAB::LUCtransgene into null mutants for phyA, phyB, cry1 or cry2,roles for phyA and cry1 in the perception of blue light, andphyA and phyB in the perception of red light weredemonstrated (Somers et al., 1998a). The assay made useof Aschoff's rule whereby increasing ¯uence rate (lightintensity) causes a decreasing period length in constantlight (Aschoff, 1979). PhyA mutant seedlings showed ade®ciency in the perception of low ¯uence rates of red andblue light consistent with the role of phyA as a low ¯uencephotoreceptor in seedling establishment. PhyB mutantseedlings showed a de®ciency in the perception of high¯uence rate red light (Somers et al., 1998a). The responsesof the phyA and phyB mutant seedlings nicely demon-strates the plasticity of recruitment of photoreceptors bythe plant to allow meaningful detection of a range of¯uence rates. Red light ¯uence rates above which the phyAresponse would saturate fall within the active range ofphyB. At these ¯uence rates, phyA is degraded and phyBbecomes the dominant red light photoreceptor. Studieswith phyA phyB double mutants con®rm these ®ndingsshowing an additivity between the phyA and phyBmonogenic phenotypes (Devlin and Kay, 2000a).

Cry1 mutants showed a de®ciency in response to low¯uence rates of blue light, a wild-type response tointermediate ¯uence rates of blue light and a de®ciencyin response to high ¯uence rates of blue light. This clearlyindicated a role for cry1 in blue light input to the clock.The cry2 monogenic mutant showed a wild-type responsefor blue light input to the clock (Somers et al., 1998a), butanalysis of the cry1 cry2 double mutant showed aredundancy between cry1 and cry2 at intermediate ¯uencerates (Devlin and Kay, 2000a). This is consistent with theway in which cry1 and cry2 act in seedling establishmentat these ¯uence rates. Both cry1 and cry2 act in seedlingestablishment in lower ¯uence rates of blue light, but cry2is degraded at higher ¯uence rates leaving cry1 as the soleblue light photoreceptor for de-etiolation (Lin et al., 1998).

Signi®cantly, for light input to the clock the cry1 mutantwas also found to show a de®ciency in the perception oflow ¯uence rate red light in a similar manner to that seen inthe phyA mutant (Devlin and Kay, 2000a). Given the factthat the absorption spectrum for the cryptochromes showsno peak in the red region of the spectrum, it must beconcluded that cryptochrome is acting downstream of thephotoreceptor phyA, presumably as a signal transductioncomponent. Double mutant analysis of phyA cry1 in whitelight con®rms that there is no additivity between these twomutations for low ¯uence rate light input to the clock(Devlin and Kay, 2000a). Both phyA and cry1 are capableof acting as photoreceptors in white light, yet cry1 isepistatic to phyA. This suggests that in low ¯uence rates ofboth red and blue light cry1 acts purely as a signal

transduction component downstream of phyA. This rolefor cry1 appears unique to circadian photoperception. Noevidence was found for a role for cry1 in seedling de-etiolation in low ¯uence rates of red light (Devlin and Kay,2000a). A scheme for the action of the cryptochromes inblue light input to the clock is shown in Fig. 5a.

Evidence for the action of phyD and phyE in red lightinput to the clock was also demonstrated. In the regulationof physiological development in A. thaliana phyD andphyE show a conditional redundancy with phyB in that theroles of phyD and phyE only become apparent in theabsence of phyB (Whitelam et al., 1998). This also appearsto be the case for the action of phyD and phyE in light inputto the clock. The phyD and phyE monogenic mutants showa wild-type response to red light, but when the phyA phyBphyD triple mutant was compared to the phyA phyB doublemutant, the phyA phyB phyD triple mutant showed arelative de®ciency in the perception of high ¯uence ratered light. Likewise the phyA phyB phyE triple mutantshowed a de®ciency in the perception of high ¯uence ratered light relative to the phyA phyB double mutant (Devlinand Kay, 2000a). The redundancy between phyD or phyEand phyB is, to some extent, understandable, given thefairly close similarity between the DNA sequences ofphyB, phyD and phyE. In particular, phyB and phyD of A.thaliana are thought to have arisen as the result of a veryrecent gene duplication subsequent to the divergence of theCruciferae and the Solanaceae (Mathews and Sharrock,1997). Signi®cantly the phyA phyB phyD triple mutant andthe phyA phyB phyE triple mutant both still showed asigni®cant shortening of period length as red light ¯uencerate increased, indicating, in each case, the action of theremaining phytochromes in red light input to the clock(Devlin and Kay, 2000a). Whether this represents theaction of phyC awaits the creation of the phyA phyB phyDphyE quadruple mutant. One important point to note is thatthe photoreceptor mutants show no effect on the periodlength of the clock in darkness indicating that they arepurely affecting the light input pathway rather thandisrupting the clock mechanism itself (Devlin and Kay,2000a). The roles of the phytochromes and the crypto-chromes in light input to the circadian clock can besummarized as in Fig. 4b.

The third class of plant photoreceptors, the phototropins,are involved in the perception of blue wavelengths of light,leading to phototropism (Huala et al., 1998; Jarillo et al.,1998; Kagawa et al., 2001). Two phototropins are presentin A. thaliana, nph1 (named after the mutant phenotype,non-phototropic hypocotyl) (Liscum and Briggs, 1995)and npl1 (nph1-like) (Kagawa et al., 2001). Phototropinsconsist of a protein moiety with two ¯avin chromophoresheld within PAS or LOV domains within the proteinmolecule (Christie et al., 1998). Analysis of the responseof the nph1 mutant to a range of ¯uence rates of blue lightshowed no evidence for the involvement of nph1 in light

Input to the clock 1541

input to the clock (Harmer et al., 2000; supplemental data),however, it remains possible that nph1 could act redun-dantly with other photoreceptors.

Oscillating photoreceptor expression

In several systems, a close association between light inputand the clock itself has been established with the result thatthe traditional distinction of input, oscillator and output isbecoming blurred. Transcription of the PHYB gene and ofthe CRY1 and CRY2 genes in A. thaliana was recentlyshown to oscillate with a circadian rhythm in continuouslight (Bognar et al., 1999; Harmer et al., 2000). PHYBshows a peak around subjective dawn whilst the CRYsshow a peak in the later part of the subjective day (Harmeret al., 2000). The circadian photoreceptors, phyB, cry1 andcry2 are therefore part of both the input and outputpathways from the clock and in this way the outputpathway may feed back on the input pathway, adding anadditional level of ®ne tuning to the clock resettingmechanism (Fig. 4). It is possible that the expressionpattern of the circadian photoreceptors may, to someextent, contribute to the shape of the phase response curvegiving increased sensitivity to light at the dawn and dusktransitions.

Such modulation of light responsiveness over the courseof the circadian cycle has been termed gating (Millar andKay, 1996). Gating is not only observed for clock resettingbut also for several clock outputs which can also bedirectly regulated by light. Control of CAB gene expres-sion is directly light responsive as well as being clockregulated. Following a period of darkness, CAB geneexpression will show an acute spike in transcription inresponse to a light pulse. However, a strong circadianrhythm in this light responsiveness of CAB induction isobserved, indicating that light-signalling to CAB isstrongly gated by the clock (Millar and Kay, 1996). Veryrecently, the ®rst candidate for a protein involved in thisgating process was identi®ed in A. thaliana. The elf3mutant shows an arrhythmic phenotype in constant whitelight (Hicks et al., 1996) but does show a circadian rhythmin darkness. McWatters et al. (2000) demonstrated that theelf3 mutant fails to show any gating of the acute responseof CAB in darkness. They concluded that the arrest of theoscillator in constant light is also a result of a loss of gatingof light input, demonstrating that gating of phototransduc-tion is an important part of the resetting response(McWatters et al., 2000) (Fig. 4).

Candidates for factors involved in the light inputpathway downstream of the photoreceptors themselveshave also been identi®ed in plants. The zeitlupe (ztl)mutant of A. thaliana was identi®ed as showing a longperiod length for the rhythm of CAB::LUC expression inconstant light (Somers et al., 2000). It also shows a longperiod length for expression of other circadian clock

regulated genes and for the rhythm of cotyledon movement(Somers et al., 2000). The ztl phenotype is stronglydependent on ¯uence rate, displaying a greater periodlengthening effect at lower ¯uence rates (Somers et al.,2000). This suggests that ztl speci®cally disrupts lightinput to the clock. ZTL forms part of a small family ofthree closely related proteins along with FKF1 and LKP2(Nelson et al., 2000; Kiyosue and Wada, 2000). FKF1 alsoaffects circadian regulated gene expression and mutationsin both ZTL and FKF1 both cause late ¯owering (Somerset al., 2000; Nelson et al., 2000). Recently ZTL wasdemonstrated to bind to both phyB and cry1 using yeasttwo-hybrid and in vitro binding studies, further suggestinga close association with the light input pathway to thecircadian clock (Jarillo et al., 2001) (Fig. 4).

Another candidate signal transduction componentdownstream of phytochrome in light input to the clock isPHYTOCHROME INTERACTING FACTOR 3 (PIF3).PIF3 was identi®ed in a yeast two-hybrid screen forphytochrome interacting factors (Ni et al., 1998). The PIF3gene encodes a bHLH-type transcription factor that bindsto G-box sequences present in the promoters of the genesencoding the critical clock components, LHY and CCA1(Martinez-Garcia et al., 2000). When a phytochromemolecule is converted to the Pfr form it moves from thecytoplasm to the nucleus and there binds to the promoter-bound PIF3 transcription factor (Kircher et al., 1999; Niet al., 1999). CCA1 and LHY gene expression is light-regulated and PIF3 has been shown to be essential fornormal light regulation (Ni et al., 1999). This has led to theconclusion that binding of phytochrome to PIF3 acts as theswitch to trigger an increase in expression of the CCA1 andLHY genes. Such a system would allow a pulse of light toreset the clock by triggering a change in CCA1 and LHYmessage levels (Fig. 4).

Photoreceptors involved in light input to thecircadian clock in insects

Following the discovery of cryptochromes in plants(Ahmad and Cashmore, 1993), molecules showing strongsimilarity to photolyases, but which failed to show anyphotolyase function, were identi®ed in animals (Sancar,2000). Because of this similarity to the plant crypto-chromes, these molecules were also termed crypto-chromes. As with the plant cryptochromes, the N-terminus of the animal cryptochromes shows stronghomology to the photolyases whilst the C-terminal formsa unique extension (Cashmore et al., 1999; Devlin andKay, 1999; Sancar, 2000). However, the animal crypto-chromes show a stronger homology with the 6-4 photo-lyases than with the type II photolyases (Cashmore et al.,1999). It is thought that the animal cryptochromes arosesubsequent to the plant cryptochromes by divergence fromthe 6-4 photolyases. The phylogeny of the plant crypto-

1542 Devlin

chromes, however, suggests that they arose much earlier.In fact, they are predicted to have arisen prior to thedivergence of plants and animals (Cashmore et al., 1999)suggesting that plant-like, type II photolyase-derivedcryptochrome has been subsequently lost in animals.None-the-less, the same blue-light-sensing mechanismseems to have been adapted once again to provide acircadian photoreceptor in animals, particularly in insectswhere cryptochrome forms the primary circadian photo-receptor (Emery et al., 2000; Devlin and Kay, 2001). Thecore clock components have been well characterized in theinsect, Drosophila melanogaster, where a transcriptionalfeedback loop involving four key molecules generatessustained oscillation at the molecular level with arelatively tight circadian period (Allada et al., 1998;Rutila et al., 1998; Darlington et al., 1998). Transcriptionof the genes, period (per) and timeless (tim) leads to a risein PER and TIM protein levels in the afternoon and earlyevening. PER±TIM dimers then enter the nucleus andrepress their own transcription by inhibiting the activity ofa positively acting transcriptional activation complex

consisting of the proteins CLOCK (CLK) and CYCLE(CYC). Consequently, levels of PER and TIM protein fallagain as these proteins are degraded until they reach a levelat which they no longer inhibit their own transcription andthe cycle begins again (Allada et al., 1998; Rutila et al.,1998; Darlington et al., 1998) (Fig. 6).

A second, interlocked negative feedback loop causesCLK to cycle in antiphase with PER and TIM. In this, CLKfeeds back as a repressor of its own transcription whilstPER and TIM act as derepressors (Glossop et al., 1999).

In D. melanogaster it was known that visual photo-receptors alone were not responsible for light input to theclock. The norpAp41 mutation causes the compound eyesand ocelli to be completely unresponsive to light, althoughmonogenic norpAp41 mutant ¯ies show normal entrainment(Wheeler et al., 1993; Yang et al., 1998). The discovery ofcryptochrome provided a strong candidate for the circadianphotoreceptor. A D. melanogaster mutant lacking crypto-chrome, named crybaby (cryb) has been most informative inrevealing the role of cryptochrome in the D. melanogasterclock. In the cryb mutant, rhythms of PER and TIMexpression in the body of the ¯y fail to entrain to light/darkcycles (Stanewsky et al., 1998). Curiously, cryb mutant¯ies were still observed to display behavioural locomotorrhythms. It was subsequently discovered that the rhythm ofper and tim expression within lateral neuron cells remainsentrainable by light suggesting that photoreceptors otherthan D. melanogaster cryptochrome (dCRY) are able toentrain these rhythms. When the norpAp41 mutation wascombined with the cryb mutation, norpAp41 cryb doublemutant ¯ies now failed to show normal entrainment ofbehavioural rhythms (Stanewsky et al., 1998) suggestingthat behavioural entrainment in D. melanogaster is medi-ated by a combination of signals from cryptochrome andfrom the visual photoperception pathway.

Phase shifting is also compromized in the cryb mutant¯ies. In response to a pulse of light that will induce a phaseshift in behavioural rhythms in wild-type ¯ies, the cryb

mutant fails to show any phase shift (Stanewsky et al.,1998). Responses of overexpressors of dcry, however,vary. Emery et al. (1998) observed an enhanced responseto light pulses for phase shifting in overexpressors of dcry.Ishikawa et al. (1999) observed a decreased response. Anassay based on Aschoff's rule (Aschoff, 1979) has alsobeen used de®nitively to demonstrate the role ofcryptochrome as a circadian photoreceptor. In D. melano-gaster, period length in constant light increases withincreasing ¯uence rate to the extent that, at very highintensities of light, wild-type ¯ies become arrhythmic(Aschoff, 1979; Konopka et al., 1989). Mutant cryb ¯ies donot respond to this increasing ¯uence rate and continue todisplay strong rhythmicity even in intense constantillumination. This indicates that circadian photoperceptionis completely impaired in the absence of cryptochrome(Emery et al., 2000). A wild-type phenotype can be

Fig. 6. The action of cryptochrome in light input to the circadianclock in the fruit ¯y, Drosophila melanogaster. The circadian clockconsists of a negative feedback loop whereby the TIMELESS (TIM)and PERIOD (PER) proteins mediate a rhythmic suppression of theirown transcription via repression of a transcriptional activationcomplex made up of the CLOCK (CLK) and CYCLE (CYC). Inresponse to light CRYPTOCHROME (CRY) binds to TIM negatingthe action of the TIM±PER dimers and resetting the clock to a point atwhich tim and per transcription is high. The D. melanogaster cry geneshows a circadian rhythm of expression with a peak in the later part ofthe day. It is proposed that this oscillation may modulate lightresetting of the clock.

Input to the clock 1543

restored in the cryb mutant ¯ies by expression of wild-typedCRY in lateral neuron cells. dCRY is thus demonstratedto be the only circadian photoreceptor that impingesdirectly on the clock in D. melanogaster (Emery et al.,2000). This leaves a question as to how the visualphotoreceptors can entrain the clock in a cryb mutant. Itwas proposed by Emery et al. (2000) that visual stimulimay act to drive a rhythm of physical activity in light/darkcycles and that this, in turn, somehow feeds back to entrainthe behavioural rhythm that can subsequently be observedin cryb ¯ies under constant conditions.

The mode of action of dCRY in resetting the clock at themolecular level was identi®ed by Ceriani et al. (1999).Using a yeast two-hybrid assay, dCRY was shown tointeract directly with TIM. This interaction was observedin the light, but not in darkness, pointing to a light-dependent interaction of dCRY with the PER-TIM dimer(Ceriani et al., 1999). Furthermore, in a D. melanogastercell culture system, dCRY was found to negate the actionof the PER-TIM dimer in feeding back on per and timtranscription in light but not in darkness, effectivelyresetting the clock to a point at which per and timtranscription are high (Ceriani et al., 1999).

The D. melanogaster cryptochrome dcry shows acircadian rhythm of expression with a peak in the laterpart of the day. As is proposed to be the case for the plantphotoreceptors, this oscillation may modulate light reset-ting of the clock and affect the shape of the phase responsecurve (Ishikawa et al., 1999) (Fig. 6).

In a recent study by Krishnan et al. (2001), a further rolefor dCRY in the D. melanogaster clock was proposed. Inaddition to the role of dCRY as the circadian photoreceptorin D. melanogaster brain, Krishnan et al. (2001) showedthat dCRY may be part of the oscillator mechanism itselfin peripheral tissue. A circadian rhythm of olfactoryresponses can be observed in the antennae of D.melanogaster. In wild-type ¯ies maintained in constantdarkness, this rhythm could be entrained by temperaturecycles, but such temperature cycles were unable to entrainthe rhythm in cryb ¯ies. This suggests either that dCRY isessential for temperature entrainment in D. melanogasteror that it forms part of the oscillator mechanism in theantenna. As there is no precedent of dCRY having athermoreceptive role, the latter seems more likely. Thisresult stresses the care that needs to be exercized whendenoting a particular gene/protein as a photoreceptor orclock component.

Photoreceptors involved in light input to thecircadian clock in mammals

In mammals, cryptochrome plays an integral role in thecentral clock mechanism. Two cryptochromes, CRY1 andCRY2 exist in mammals (Devlin and Kay, 1999), though itis uncertain whether they play a role within light input to

the clock. In mice, the clock mechanism is made up of asimilar transcriptional feedback loop to that discovered ininsects, though CRY replaces TIM within the mammaliansystem and there are three per genes present (Kume et al.,1999; Field et al., 2000; Reppert and Weaver, 2001).Transcription of the mCry and mPer genes begins in thelate afternoon and levels of the proteins in the cytoplasmrise steadily. mCRY±mCRY dimers or mCRY±mPERdimers then form, enabling their entry into the nucleuswhere these dimers inhibit the action of their owntranscriptional activators, CLK and CYC, and thus exerta negative feedback on their own transcription so thatlevels of cry and per subsequently fall again. It appears thatinhibition of the CLK±CYC complex is effected largely bythe CRY proteins, thus the role of CRY has become centralto the clock mechanism itself (Reppert and Weaver, 2001).Consistent with this, mCry1±/± mCry2±/± mice are arrhyth-mic in constant light (van der Horst et al., 1999). Again asecond, interlocking feedback loop exists. In this casemPER2 acts as a promoter of CYC transcription, whilstmCRY plays a role in stabilizing mPER2 (Shearman et al.,2000).

Some evidence points to a role for cry in light input tothe mammalian clock, though this is not de®nitive. It isdif®cult to prove a role for CRY in light input to themammalian clock based on studies of the mCry1±/±

mCry2±/± double mutant mice as the clock itself no longerruns in these mutants (van der Horst et al., 1999). Evidencefor altered response to phase-shifting pulses of light hasbeen demonstrated for the mCry2±/± monogenic mutants.mCry2±/± mutant mice show an enhanced response to aphase-shifting light pulse during the subjective night(Thresher et al., 1998). One of the early events associatedwith phase resetting in response to a light pulse is a rapidinduction of mPer1. mCry1±/± and mCry2±/± monogenicmutants display aberrant mPer1 induction (Vitaterna et al.,1999). Studies of light-induced mPer1 transcription inmCry1±/± mCry2±/± double mutant mice also suggest a rolefor mCRY as a circadian photoreceptor. mCry1±/± mCry2±/±

double mutant mice fail to show mPer1 induction inresponse to a light pulse given during the night. However,if mCry1±/± mCry2±/± mutants are left in darkness for 52 hprior to a light pulse, then light-induction of mPer1 isobserved, suggesting the action of another, possibly light-labile, photoreceptor (Okamura et al., 1999). In addition,mPer2 is still induced in response to a light pulse inmCry1±/± mCry2±/± mutants and, whilst induction of mPer2has not been shown to lead to clock resetting, this alsoindicates the involvement of photoreceptors other thancryptochrome acting upon the clock system (Vitaternaet al., 1999).

mCRYs do play a role as photoreceptors for behaviouralmodi®cation in mice. As well as showing a circadianregulation of locomotor activity, whereby they are activeduring the subjective night and inactive during the

1544 Devlin

subjective day, wild-type mice show a simple responsivebehaviour to light/dark cycles whereby they remaininactive whenever the light is on. The cryptochromeshave been shown to act redundantly with the classicalretinal photoreceptors in the perception of light in thecontrol of this response (Selby et al., 2000). This resultindicates that they can act as photoreceptors and that theydo play a dual role, both as part of the clock mechanismand as photoreceptors in their own right. Whether theycontribute to circadian photoperception will be moredif®cult to ascertain.

It is certain, however, that the mammalian circadianphotoreceptor is located in the eye as enucleation results inloss of circadian entrainment (Foster, 1998). Experimentswith rodless and coneless rats, however, have demon-strated that neither rods nor cones are essential for clockresetting (Freedman et al., 1999). mCRY is found in theretinal ganglion layer of the eye, thus cryptochrome is notruled out as the ocular photoreceptor (Miyamoto andSancar, 1998). However, it is probable that the action of anovel mammalian photoreceptor located in the eye is beingwitnessed. This is also suggested by the action spectrumfor mammalian clock resetting which shows a peak ofaction at ~500 nm, corresponding better to an opsin(Provencio and Foster, 1995; Yoshimura and Ebihara,1996) than to cryptochrome (Sancar, 2000). Novel opsinshave, indeed, recently been found in Xenopus and insalmon (Soni et al., 1998; Provencio et al., 1998b).

Differences in circadian photoperceptionbetween plants and animals

It is notable that the A. thaliana cry1 cry2 double mutant isstill rhythmic, clearly indicating that, in A. thaliana, thecryptochromes are not essential for the running of theclock as they are in mammals but are purely involved inlight input (Devlin and Kay, 2000a). The clock itself seemsto have evolved independently in plants and in animalsand, although much of the clock `mechanism' appearsconserved between insects and mammals (Willaims andSehgal, 2001), these components have not been found inplants. It is, therefore, no surprise that the circadianphotoreceptors also arose independently in plants andanimals, but it remains an interesting observation that, inboth cases, molecules derived from the blue-light sensingphotolyases are closely involved with the clock.

Another distinct feature of plant and animal clocks isthat, whilst in animals there is a strong central co-ordination of circadian oscillations throughout the organ-ism mediated via the brain, in plants there appears to be nosuch co-ordination. Within mammals a region of thehypothalamus known as the suprachiasmatic nucleus(SCN) acts as a central clock regulating rhythms through-out the rest of the body (Yamazaki et al., 2000). The SCNshows a direct synaptic connection to the eyes, thought to

be the route of clock resetting light signals (Provencioet al., 1998a). Whilst rhythms can be maintained inisolated mammalian tissues, clock resetting by light hasnot been demonstrated in these peripheral tissues andappears to require central control via signals from the SCN(Sakamoto et al., 1998; Yamazaki et al., 2000). Within¯ies the central control appears to be exhibited by thelateral neurones (Stanewsky et al., 1998). Rhythmicexpression of the clock genes within the lateral neuronesalone has been demonstrated to be suf®cient to maintainlocomotor rhythms of the whole ¯y (Stanewsky et al.,1998). Isolated body parts of D. melanogaster can alsomaintain rhythmic clock gene expression and, furthermore,can be entrained by light signals indicating that, althoughthe clock is controlled centrally in ¯ies, clock resettingthroughout the organism is not absolutely dependent onsignals from the lateral neurones (Plautz et al., 1997). Inplants, it has been demonstrated the intact cotyledons onthe same plant can be stably entrained to different light/dark cycles independently and will continue to cycle out-of-phase when transferred back to constant conditions(Thain et al., 2000). This indicates that not only is eachpart of the plant capable of independent clock resetting butthat there is no systemic co-ordination of the clockthroughout the plant as a whole. All co-ordination ofrhythms throughout the plant is therefore dependent oneach part of the plant detecting the same environmentalsignals.

Photoreceptors involved in light input to thecircadian clock in fungi

Yet another independently evolved clock mechanism isfound in the fungus, Neurospora crassa. N. crassa formeda model organism in which many of the principles of anoscillator based on a transcriptional feedback loop wereestablished (Dunlap, 1999). The frequency (frq) gene of N.crassa forms part of a self-sustaining oscillator wherebythe FRQ protein feeds back on frq transcription (Aronsonet al., 1994). In N. crassa two proteins, WHITECOLLAR1 and 2 (WC-1 and WC-2), that are required for all knownphotoresponses, have been shown to be essential for lightinput to the clock (Harding and Melles, 1983; Russo, 1988;Sommer et al., 1989; Lauter and Russo, 1991; Arpaia et al.,1993). WC-1 and WC-2 act as positive regulators thatmaintain robust FRQ cycling (Crosthwaite et al., 1997).WC-1 and WC-2 form a transactivating complex known asthe whitecollar complex (WCC) which activates frqtranscription (Ballario et al., 1998; Talora et al., 1999;Denault et al., 2001). As FRQ protein subsequently rises,the FRQ protein interacts directly with this complex tonegate its action on the frq promoter so that FRQ protein isno longer produced and the levels of FRQ subsequentlyfall again (Denault et al., 2001). WC-1 is responsible forthe light induction of the frq transcript that results in a

Input to the clock 1545

phase shift in the rhythm (Crosthwaite et al., 1997).Interestingly, the WC-1 is also rhythmically expressedunder the control of FRQ protein (Denault et al., 2001). Aswith the oscillating light input components in A. thalianaand D. melanogaster this may form a mechanism of gatingof the light responsiveness of the clock to appropriatetimes of day. Very recently a further putative gatingcomponent was identi®ed in N. crassa. The protein, VIVID(VVD) regulates the light response in N. crassa (Heintzenet al., 2001). VVD forms an independent autoregulatorytranscriptional feedback loop. Transcription of vvd isregulated by the WCC and VVD protein is rapidly inducedby light. The VVD protein then acts to negate the WCC,effectively reducing the light responsiveness of N. crassa.Consequently, vvd mutants of N. crassa show severelydampened gating and an alteration in the phase responsecurve (Heintzen et al., 2001) (Fig. 7).

Temperature input to the clock

The response of the circadian clock to temperature showssomething of a paradox. The clock exhibits a temperaturecompensation whereby, under constant conditions, theperiod length of the rhythm is almost unaffected by the

external temperature. The phenomenon of temperaturecompensation was ®rst observed by Pittendrigh (1954)studying the eclosion rhythm of Drosophila pseudoobs-cura pupae. He found that this rhythm maintains the sameperiod length over a wide range of temperature. Similarly,the rhythms studied in plants show very low sensitivity totemperature (Salisbury et al., 1968; Somers et al., 1998b).For most biochemical reactions the rate will increase 2±3-fold for a 10 °C increase in temperature (Q10=2±3).CAB::LUC rhythms in A. thaliana show a slight shorteningof period length in response to increasing temperature witha Q10 of 1.0±1.1 (Somers et al., 1998b).

However, despite the temperature compensation whichis built into the endogenous clock, a sudden temperaturechange is able to act as a Zeitgeber and entrain the clock toa new phase. Clocks can, in fact, be entrained to atemperature rhythm such as that which occurs with theday/night cycle in the natural environment (BuÈnning,1973; Somers et al., 1998b; Merrow et al., 1999), thoughmuch less is known about the temperature input to theclock in entrainment.

Temperature entrainment has been used experimentallyto examine the position of mutations within the clocksystem. The frq null mutant of N. crassa shows anarrhythmic phenotype in constant conditions. The frq nullfails to entrain to light/dark cycles, but will stably entrainto temperature cycles, suggesting that frq is a critical clockcomponent essential for light-mediated entrainment ofcircadian rhythms in N. crassa (Merrow et al., 1999). Itappears that another frq-independent oscillator is respon-sible for the rhythms observed in frq mutants in response totemperature entrainment (Roenneberg and Merrow, 2000).Such a frq-independent oscillator had also previously beenproposed on the basis that the frq null mutant willeventually display a (very long period) rhythm after anextended period in constant conditions (Loros andFeldman, 1986). Curiously, wc-1 and wc-2 mutants cannotbe entrained by either light or temperature indicating awider involvement in the clock (Crosthwaite et al., 1997).

Similarly temperature entrainment was used to demon-strate that the elf3 mutant of A. thaliana speci®cally affectslight input to the oscillator. As discussed earlier, the elf3mutant has lost gating light signalling causing the action ofthe oscillator to be masked in continuous light (McWatterset al., 2000). Temperature entrainment can restorerhythmicity in the elf3 mutant demonstrating that themutation speci®cally affects the light input pathway to theclock (McWatters et al., 2000).

Temperature cycles were used to demonstrate that thetoc1 mutation in A. thaliana lies within the centraloscillator as opposed to being in an input pathway(Somers et al., 1998b). The toc1 mutant displays a shortperiod length for a number of rhythmic outputs even indarkness indicating that it does not lie in the light inputpathway (Somers et al., 1998b). toc1 was shown not to

Fig. 7. The circadian system in the fungus, Neurospora crassa. TheFREQUENCY (FRQ) protein feeds back on frq transcription. Twoproteins, WHITECOLLAR 1 and 2 (WC-1 and WC-2) act as positiveregulators that maintain robust FRQ cycling. WC-1 and WC-2 are alsoessential for light input to the clock. Sensitivity to light is modulated(gated) by a fourth protein, VIVID. VIVID oscillates with a circadianrhythm under the control of the WHITECOLLAR complex (WCC)and, in turn, acts to negate the WCC reducing light sensitivity.

1546 Devlin

affect temperature compensation and following tempera-ture entrainment the toc1 mutant also showed a shortperiod length. Furthermore, in cycles of alternating highand low temperature, the peak of CAB::LUC expression inthe toc1 mutant led that of the wild type (Somers et al.,1998b). A feature of mutants affecting the period length ofthe clock is that the timing of the peak of the rhythmrelative to the entraining signal will differ from that seen inthe wild type. In a short period mutant the peak will leadthat of the wild type and in a long period mutant the peakwill lag behind that of the wild type. Thus the toc1 mutantdoes not affect the process of temperature entrainment ofthe clock and so does not lie within the temperature inputpathway. toc1 is, therefore, proposed to affect the oscil-lator mechanism itself (Somers et al., 1998b).

Conclusions

A common feature of circadian systems in all speciesexamined is the close association between light input andthe oscillator itself. Several systems display an oscillationin expression of the circadian photoreceptors themselves.Phytochrome B in A. thaliana shows a circadian rhythm oftranscription with a peak in the late part of the night/earlymorning (Bognar et al., 1999; Harmer et al., 2000). The A.thaliana cryptochromes oscillate with a peak in the latepart of the day (Harmer et al., 2000). D. melanogastercryptochrome also shows a circadian oscillation with apeak in the late part of the day (Ishikawa et al., 1999).Interestingly, whilst it is uncertain whether the mammaliancryptochromes are circadian photoreceptors their expres-sion also oscillates under circadian control. mCry1 oscil-lates with a circadian rhythm in the SCN, also showing apeak in the late part of the day, whilst both mCry1 andmCry2 oscillate with a circadian rhythm in peripheralskeletal tissue (Kume et al., 1999).

Clearly the output from the clock feeds back to regulateinput. As well as direct regulation of photoreceptor levels agating or modulation of responsivity to signals from thephotoreceptors is an integral part of many circadiansystems. This can regulate the degree of clock resettingin response to the signals from the light environment atdifferent times of day (Roenneberg and Foster, 1997). Thisfeature of these clocks means that they will respond moststrongly to environmental conditions at crucial times ofday, namely dawn and/or dusk when the acquisition ofinformation about the light environment is most important.Recently, elements involved in gating responses havebegun to be characterized (McWatters et al., 2000;Heintzen et al., 2001). The elucidation of such regulatorygating or modulating mechanisms will be important for thefull understanding of the central oscillator itself. It mayeven be hypothesized that a feedback loop modulating theactivity of a photoreceptor could have been at the veryorigin of circadian clocks.

References

Ahmad M, Cashmore AR. 1993. HY4 gene of A. thaliana encodesa protein with characteristics of a blue-light photoreceptor.Nature 366, 162±166.

Alabadi D, Oyama T, Yanovsky MJ, Harmon FG, Mas P, KaySA. 2001. Reciprocal regulation between TOC1 and LHY/CCA1within the Arabidopsis circadian clock. Science 293, 880±883.

Allada R, White NE, So WV, Hall JC, Rosbash M. 1998. Amutant Drosophila homolog of mammalian Clock disruptscircadian rhythms and transcription of period and timeless. Cell93, 791±804.

Aronson BD, Johnson KA, Loros JJ, Dunlap JC. 1994. Negativefeedback de®ning a circadian clock: autoregulation of the clockgene frequency. Science 263, 1578±1584.

Arpaia G, Loros JJ, Dunlap JC, Morelli G, Macino G. 1993. Theinterplay of light and the circadian clock. Plant Physiology 102,1299±1305.

Aschoff J. 1979. Circadian rhythms: in¯uences of internal andexternal factors on the period measured in constant conditions.Zeitschrift fuÈr Tierpsychologie 49, 225±249.

Ballario P, Talora C, Galli D, Linden H, Macino G. 1998. Rolesin dimerization and blue light photoresponse of the PAS and LOVdomains of Neurospora crassa white collar proteins. MolecularMicrobiology 29, 719±729.

Bognar LK, Hall A, Adam E, Thain SC, Nagy F, Millar AJ.1999. The circadian clock controls the expression pattern of thecircadian input photoreceptor, phytochrome B. Proceedings of theNational Academy of Sciences, USA 96, 14652±14657.

BuÈnning E. 1973. The physiological clock, New York: Springer.Carre IA, Kim J-Y. 2002. MYB transcription factors in the

Arabidopsis circadian clock. Journal of Experimental Botany 53,1551±1557.

Cashmore AR, Jarillo JA, Wu YJ, Liu D. 1999. Cryptochromes:blue light receptors for plants and animals. Science 284, 760±765.

Ceriani MF, Darlington TK, Staknis D, Mas P, Petti AA, WeitzCJ, Kay SA. 1999. Light-dependent sequestration of TIMELESSby CRYPTOCHROME. Science 285, 553±556.

Christie JM, Briggs WR. 2001. Blue light sensing in higher plants.Journal of Biological Chemistry 276, 11457±11460.

Christie JM, Reymond P, Powell GK, Bernasconi P, RaibekasAA, Liscum E, Briggs WR. 1998. Arabidopsis NPH1: a¯avoprotein with the properties of a photoreceptor forphototropism. Science 282, 1698±1701.

Clack T, Mathews S, Sharrock RA. 1994. The phytochromeapoprotein family in Arabidopsis is encoded by ®ve genes: thesequences and expression of PHYD and PHYE. Plant MolecularBiology 25, 413±427.

Crosthwaite SK, Dunlap JC, Loros JJ. 1997. Neurospora wc-1and wc-2: Transcription, photoresponses, and the origins ofcircadian rhythmicity. Science 276, 763±769.

Crosthwaite SK, Loros JJ, Dunlap JC. 1995. Light-inducedresetting of a circadian clock is mediated by a rapid increase infrequency transcript. Cell 81, 1001±1012.

Darlington TK, Wager-Smith K, Ceriani MF, Staknis D,Gekakis N, Steeves TDL, Weitz CJ, Takahashi JS, Kay SA.1998. Closing the circadian loop: CLOCK-induced transcriptionof its own inhibitors per and tim. Science 280, 1599±1603.

Darwin C. [1895] 1981. The power of movement in plants. NewYork: D Appleton and Co.

de Mairan J. 1729. Observation botanique. Histoire de l'AcademieRoyale des Sciences, 35±36.

Denault DL, Loros JJ, Dunlap JC. 2001. WC-2 mediates WC-1-FRQ interaction within the PAS protein-linked circadianfeedback loop of Neurospora. EMBO Journal 20, 109±117.

Input to the clock 1547

Devlin PF, Kay SA. 1999. CryptochromesÐbringing the blues tocircadian rhythms. Trends in Cell Biology 9, 295±298.

Devlin PF, Kay SA. 2000a. Cryptochromes are required forphytochrome signaling to the circadian clock but not forrhythmicity. The Plant Cell 12, 2499±2510.

Devlin PF, Kay SA. 2000b. Flower arranging in Arabidopsis.Science 288, 1600±1602.

Devlin PF, Kay SA. 2001. Circadian photoperception. AnnualReview of Physiology 63, 677±694.

Duf®eld GE, Best JD, Meurers BH, Bittner A, Loros JJ, DunlapJC. 2002. Circadian programs of transcriptional activation,signalling and protein turnover revealed by microarray analysisof Mammalian cells. Current Biology 12, 551±557.

Dunlap JC. 1999. Molecular bases for circadian clocks. Cell 96,271±290.

Edmunds LN. 1988. Cellular and molecular bases of biologicalclocks. New York: Springer-Verlag.

Emery P, Stanewsky R, Hall JC, Rosbash M. 2000. A uniquecircadian-rhythm photoreceptor. Nature 404, 456±457.

Emery PT, So WV, Kaneko M, Hall JC, Rosbash M. 1998. CRY,a Drosophila clock and light-regulated cryptochrome, is a majorcontributor to circadian rhythm resetting and photosensitivity.Cell 95, 669±679.

Engelmann W, Simon K, Phen CJ. 1994. Leaf movement inArabidopsis thaliana. Zeitschrift fuÈr Naturforschung 47, 925±928.3

Enright JT. 1982. Sleep movements of leaves: in defense ofDarwin's interpretation. Oecologia 54, 253±259.

Field MD, Maywood ES, O'Brien JA, Weaver DR, Reppert SM,Hastings MH. 2000. Analysis of clock proteins in mouse SCNdemonstrates phylogenetic divergence of the circadian clockworkand resetting mechanisms. Neuron 25, 437±447.

Foster RG. 1998. Shedding light on the biological clock. Neuron20, 829±832.

Freedman MS, Lucas RJ, Soni B, von Schantz M, Munoz M,David-Gray Z, Foster R. 1999. Regulation of mammaliancircadian behaviour by non-rod, non-cone ocular photoreceptors.Science 284, 502±504.

Glossop NR, Lyons LC, Hardin PE. 1999. Interlocked feedbackloops within the Drosophila circadian oscillator. Science 286,766±768.

Hall JC. 2000. Cryptochromes: sensory reception, transduction,and clock functions subserving circadian systems. CurrentOpinion in Neurobiology 10, 456±466.

Harding RW, Melles S. 1983. Genetic analysis of phototropism ofNeurospora crassa perithecial beaks using white collar and albinomutants. Plant Physiology 72, 996±1000.

Harmer SL, Hogenesch JB, Straume M, Chang HS, Han B, ZhuT, Wang X, Kreps JA, Kay SA. 2000. Orchestrated transcriptionof key pathways in Arabidopsis by the circadian clock. Science290, 2110±2113.

Heintzen C, Loros JJ, Dunlap JC. 2001. The PAS protein VIVIDde®nes a clock-associated feedback loop that represses lightinput, modulates gating, and regulates clock resetting. Cell 104,453±464.

Hicks KA, Millar AJ, Carre IA, Somers DE, Straume M,Meeks-Wagner R, Kay SA. 1996. Conditional circadiandysfunction of the Arabidopsis early-¯owering 3 mutant.Science 274, 790±792.

Hirschfeld M, Tepperman JM, Clack T, Quail PH, SharrockRA. 1998. Coordination of phytochrome levels in phyB mutantsof Arabidopsis as revealed by apoprotein-speci®c monoclonalantibodies. Genetics 149, 523±535.

Huala E, Oeller PW, Liscum E, Han IS, Larsen E, Briggs WR.1998. Arabidopsis NPH1: a protein kinase with a putative redox-sensing domain. Science 278, 2120±2123.

Ishikawa T, Matsumoto A, Kato Jr T, Togashi S, Ryo H,Ikenaga M, Todo T, Ueda R, Tanimura T. 1999. DCRY is aDrosophila photoreceptor protein implicated in light entrainmentof circadian rhythm. Genes to Cells 4, 57±65.

Jarillo JA, Ahmad M, Cashmore AR. 1998. A second member ofthe NPH serine/threonine kinase family of Arabidopsis. PlantPhysiology 117, 719.

Jarillo JA, Capel J, Tang RH, Yang HQ, Alonso JM, Ecker JR,Cashmore AR. 2001. An Arabidopsis circadian clock componentinteracts with both CRY1 and phyB. Nature 410, 487±490.

Johnson CH. 1990. PRC Atlas. http://johnsonlab.biology.vanderbilt.edu/prcatlas/prcatlas.html

Johnson E, Bradley M, Harberd NP, Whitelam GC. 1994.Photoresponses of light-grown phyA mutants of Arabidopsis.Phytochrome A is required for the perception of daylengthextensions. Plant Physiology 105, 141±149.

Kagawa T, Sakai T, Suetsugu N, Oikawa K, Ishiguro S, Kato T,Tabata S, Okada K, Wada M. 2001. Arabidopsis NPL1: aphototropin homolog controlling the chloroplast high-lightavoidance response. Science 291, 2138±2141.

Kendrick RE, Kronenberg GHM. (eds.) 1994.Photomorphogenesis in plants. Dordrecht: Kluwer AcademicPlublishers.

Kircher S, Kozma-Bognar L, Kim L, Adam E, Harter K,Schafer E, Nagy F. 1999. Light quality-dependent nuclearimport of the plant photoreceptors phytochrome A and B. ThePlant Cell 11, 1445±1456.

Kiyosue T, Wada M. 2000. LKP1 (LOV kelch protein 1): a factorinvolved in the regulation of ¯owering time in Arabidopsis. ThePlant Journal 23, 807±815.

Konopka RJ, Pittendrigh CS, Orr D. 1989. Reciprocal behaviorassociated with altered homeostasis and photosensitivity ofDrosophila clock mutants. Journal of Neurogenetics 6, 1±10.

Krishnan B, Levine JD, Lynch MK, Dowse HB, Funes P, HallJC, Hardin PE, Dryer SE. 2001. A new role for cryptochrome ina Drosophila circadian oscillator. Nature 411, 313±317.

Kume K, Zylka MJ, Sriram S, Shearman LP, Weaver DR, JinX, Maywood ES, Hastings MH, Reppert SM. 1999. mCRY1and mCRY2 are essential components of the negative limb of thecircadian clock feedback loop. Cell 98, 193±205.

Lauter FR, Russo VE. 1991. Blue light induction of conidiation-speci®c genes in Neurospora crassa. Nucleic Acids Research 19,6883±6886.

Lin C, Robertson DE, Ahmad M, Raibekas AA, Jorns MS,Dutton PL, Cashmore AR. 1995. Association of ¯avin adeninedinucleotide with the Arabidopsis blue light receptor CRY1.Science 269, 968±970.

Lin C, Yang HY, Guo HW, Mockler T, Chen J, Cashmore AR.1998. Enhancement of blue-light sensitivity of Arabidopsisseedlings by a blue light receptor cryptochrome 2. Proceedingsof the National Academy of Sciences, USA 95, 2690.

Liscum E, Briggs WR. 1995. Mutations in the NPH1 locus ofArabidopsis disrupt the perception of phototropic stimuli. ThePlant Cell 7, 473±485.

Loros JJ, Feldman JF. 1986. Loss of temperature compensation ofcircadian period length in the frq-9 mutant of Neurospora crassa.Journal of Biological Rhythms 1, 187±198.

Luce GG. 1971. Biological rhythms in human and animalphysiology. New York: Dover Publications.

Malhotra K, Kim S-T, Batschauer A, Dawut L, Sancar A. 1995.Putative blue-light photoreceptors from Arabidopsis thaliana andSinapis alba with a high degree of sequence homology to DNAphotolyase contain the two photolyase cofactors but lack DNArepair activity. Biochemistry 34, 6892±6899.

Martinez-Garcia JF, Huq E, Quail PH. 2000. Direct targeting of

1548 Devlin

light signals to a promoter element-bound transcription factor.Science 288, 859±863.

Mathews S, Sharrock RA. 1997. Phytochrome gene diversity.Plant, Cell and Environment 20, 666±671.

McDonald MJ, Rosbash M. 2001. Microarray analysis andorganization of circadian gene expression in Drosophila. Cell107, 567±578.

McWatters HG, Bastow RM, Hall A, Millar AJ. 2000. The ELF3zeitnehmer regulates light signalling to the circadian clock.Nature 408, 716±720.

Merrow M, Brunner M, Roenneberg T. 1999. Assignment ofcircadian function for the Neurospora clock gene frequency.Nature 399, 584±586.

Millar AJ, Carre IA, Strayer CA, Chua N-H, Kay SA. 1995a.Circadian clock mutants in Arabidopsis identi®ed by luciferaseimaging. Science 267, 1161±1163.

Millar AJ, Kay SA. 1991. Circadian control of cab genetranscription and mRNA accumulation in Arabidopsis. ThePlant Cell 3, 541±550.

Millar AJ, Kay SA. 1996. Integration of circadian andphototransduction pathways in the network controlling CABgene transcription in Arabidopsis. Proceedings of the NationalAcademy of Sciences, USA 93, 15491±15496.

Millar AJ, Short SR, Chua N-H, Kay SA. 1992. A novel circadianphenotype based on ®re¯y luciferase expression in transgenicplants. The Plant Cell 4, 1075±1087.

Millar AJ, Straume M, Chory J, Chua N-H, Kay SA. 1995b. Theregulation of circadian period by phototransduction pathways inArabidopsis. Science 267, 1163±1166.

Miyamoto Y, Sancar A. 1998. Vitamin B2-based blue-lightphotoreceptors in the retinohypothalamic tract as thephotoactive pigments for setting the circadian clock inmammals. Proceedings of the National Academy of Sciences,USA 95, 6097±6102.

Moore-Ede CM, Sulzman FM, Fuller CA. 1982. The clocks thattime us. Cambridge: Harvard University Press.

Nelson DC, Lasswell J, Rogg LE, Cohen MA, Bartel B. 2000.FKF1, a clock-controlled gene that regulates the transition to¯owering in Arabidopsis. Cell 101, 331±340.

Nelson DE, Takahashi JS. 1991. Sensitivity and integration in avisual pathway for circadian entrainment in the hamster(Mesocricetus auratus). Journal of Physiology 439, 115±145.

Ni M, Tepperman JM, Quail PH. 1998. PIF3, a phytochrome-interacting factor necessary for normal photoinduced signaltransduction, is a novel basic helix±loop±helix protein. Cell 95,657±667.

Ni M, Tepperman JM, Quail PH. 1999. Binding of phytochromeB to its nuclear signalling partner PIF3 is reversibly induced bylight. Nature 400, 781±784.

Okamura H, Miyake S, Sumi Y, Yamaguchi S, Yasui A,Muijtjens M, Hoeijmakers JH, van der Horst GT. 1999.Photic induction of mPer1 and mPer2 in cry-de®cient micelacking a biological clock. Science 286, 2531±2534.

Pittendrigh CS. 1954. On temperature independence in the clocksystem controlling emergence time in Drosophila. Proceedings ofthe National Academy of Sciences, USA 40, 1018±1029.

Plautz JD, Kaneko M, Hall JC, Kay SA. 1997. Independentphotoreceptive circadian clocks throughout Drosophila. Science278, 1632±1635.

Provencio I, Cooper HM, Foster RG. 1998a. Retinal projectionsin mice with inherited retinal degeneration: implications forcircadian photoentrainment. Journal of Comparative Neurology395, 417±439.

Provencio I, Foster RG. 1995. Circadian rhythms in mice can beregulated by photoreceptors with cone-like characteristics. BrainResearch 694, 183±190.

Provencio I, Jiang G, De Grip WJ, Hayes WP, Rollag MD.1998b. Melanopsin: an opsin in melanophores, brain, and eye.Proceedings of the National Academy of Sciences, USA 95, 340±345.

Quail PH. 1991. Phytochrome: a light-activated molecular switchthat regulates plant gene expression. Annual Review of Genetics25, 389±409.

Reppert SM, Weaver DR. 2001. Molecular analysis of mammaliancircadian rhythms. Annual Review of Physiology 63, 647±676.

Roenneberg T, Foster RG. 1997. Twilight times: light and thecircadian system. Photochemistry and Photobiology 66, 549±561.

Roenneberg T, Merrow M. 2000. Circadian clocks: Omnes viaeRomam ducunt. Current Biology 10, R742±R745.

Russo VE. 1988. Blue light induces circadian rhythms in the bdmutant of Neurospora: double mutants bd,wc-1 and bd,wc-2 areblind. Journal of Photochemistry and Photobiology 2, 59±65.

Rutila JE, Suri V, Le M, So WV, Rosbash M, Hall JC. 1998.CYCLE is a second bHLH-PAS clock protein essential forcircadian rhythmicity and transcription of Drosophila period andtimeless. Cell 93, 805±814.

Sakamoto K, Nagase T, Fukui H, Horikawa K, Okada T,Tanaka H, Sato K, Miyake Y, Ohara O, Kako K, Ishida N.1998. Multitissue circadian expression of rat period homolog(rPer2) mRNA is governed by the mammalian circadian clock,the suprachiasmatic nucleus in the brain. Journal of BiologicalChemistry 273, 27039±27042.

Salisbury FB, Spomer GG, Sobral M, and Ward RT. 1968.Analysis of an alpine environment. Botanical Gazette 129, 16±32.

Sancar A. 2000. Cryptochrome: the second photoactive pigment inthe eye and its role in circadian photoreception. Annual Review ofBiochemistry 69, 31±67.

Schaffer R, Ramsay N, Samach A, Corden S, Putterill J, CarreÂIA, Coupland G. 1998. The late elongated hypocotyl mutation ofArabidopsis disrupts circadian rhythms and the photoperiodiccontrol of ¯owering. Cell 93, 1219±1229.

Selby CP, Thompson C, Schmitz TM, Van Gelder RN, SancarA. 2000. Functional redundancy of cryptochromes and classicalphotoreceptors for non-visual ocular photoreception in mice.Proceedings of the National Academy of Sciences, USA 97,14697±14702.

Sharrock RA, Quail PH. 1989. Novel phytochrome sequences inArabidopsis thaliana: structure, evolution, and differentialexpression of a plant regulatory photoreceptor family. Genesand Development 3, 1745±1757.

Shearman LP, Sriram S, Weaver DR, Maywood ES, Chaves I,Zheng B, Kume K, Lee CC, van der Horst GT, Hastings MH,Reppert SM. 2000. Interacting molecular loops in themammalian circadian clock. Science 288, 1013±1019.

Somers DE, Devlin PF, Kay SA. 1998a. Phytochromes andcryptochromes in the entrainment of the Arabidopsis circadianclock. Science 282, 1488±1490.

Somers DE, Schultz TF, Milnamow M, Kay SA. 2000.ZEITLUPE, a novel clock associated PAS protein fromArabidopsis. Cell 101, 319±329.

Somers DE, Webb AAR, Pearson M, Kay S. 1998b. The short-period mutant, toc1-1, alters circadian clock regulation ofmultiple outputs throughout development in Arabidopsisthaliana. Development 125, 485±494.

Sommer T, Chambers JA, Eberle J, Lauter FR, Russo VE. 1989.Fast light-regulated genes of Neurospora crassa. Nucleic AcidsResearch 17, 5713±5723.

Soni BG, Philp AR, Foster RG, Knox BE. 1998. Novel retinalphotoreceptors. Nature 394, 27±28.

Stanewsky R, Kaneko M, Emery P, Beretta B, Wager-Smith K,Kay SA, Rosbash M, Hall JC. 1998. The cryb mutation

Input to the clock 1549

identi®es cryptochrome as a circadian photoreceptor inDrosophila. Cell 95, 681±692.

Strayer C, Oyama T, Schultz TF, Raman R, Somers DE, Mas P,Panda S, Kreps JA, Kay SA. 2000. Cloning of the Arabidopsisclock gene TOC1, an autoregulatory response regulator homolog.Science 289, 768±771.

Sulzman FM. 1983. Primate circadian rhythms. Bioscience 33,445±450.

Talora C, Franchi L, Linden H, Ballario P, Macino G. 1999.Role of a white collar-1-white collar-2 complex in blue-lightsignal transduction. EMBO Journal 18, 4961±4968.

Thain SC, Hall A, Millar AJ. 2000. Functional independence ofcircadian clocks that regulate plant gene expression. CurrentBiology 10, 951±956.

Thresher RJ, Vitaterna MH, Miyamoto Y, Kazantsev A, HsuDS, Petit C, Selby CP, Dawut L, Smithies O, Takahashi JS,Sancar A. 1998. Role of mouse cryptochrome blue-lightphotoreceptor in circadian photoresponses. Science 282, 1490±1494.

van der Horst GT, Muijtjens M, Kobayashi K, Takano R,Kanno S, Takao M, de Wit J, Verkerk A, Eker AP, vanLeenen D, Buijs R, Bootsma D, Hoeijmakers JH, Yasui A.1999. Mammalian Cry1 and Cry2 are essential for maintenance ofcircadian rhythms. Nature 398, 627±630.

Vitaterna MH, Selby CP, Todo T, Niwa H, Thompson C,Fruechte EM, Hitomi K, Thresher RJ, Ishikawa T, MiyazakiJ, Takahashi JS, Sancar A. 1999. Differential regulation ofmammalian period genes and circadian rhythmicity bycryptochromes 1 and 2. Proceedings of the National Academyof Sciences, USA 96, 12114±12119.

Wager-Smith K, Kay SA. 2000. Cicadian rythym genetics: from¯ies to mice to humans. Nature Genetics 26, 23±27.

Wang ZY, Tobin EM. 1998. Constitutive expression of theCIRCADIAN CLOCK ASSOCIATED 1 (CCA1) gene disruptscircadian rhythms and suppresses its own expression. Cell 93,1207±1217.

Wheeler DA, Hamblen-Coyle MJ, Dushay MS, Hall JC. 1993.Behavior in light-dark cycles of Drosophila mutants that arearrhythmic, blind, or both. Journal of Biological Rhythms 8, 67±94.

Whitelam GC, Devlin PF. 1998. Light signalling in Arabidopsis.Plant Physiology and Biochemistry 36, 125±133.

Whitelam GC, Johnson E, Peng J, Carol P, Anderson ML, CowlJS, Harberd NP. 1993. Phytochrome A null mutants ofArabidopsis display a wild-type phenotype in white light. ThePlant Cell 5, 757±768.

Whitelam GC, Patel S, Devlin PF. 1998. Phytochromes andphotomorphogenesis in Arabidopsis. Philosophical Transactionsof the Royal Society of London, Series B: Biological Sciences353, 1445±1453.

Willaims JA, Sehgal A. 2001. Molecular components of thecircadian system in Drosophila. Annual Review of Physiology 63,729±755.

Yamazaki S, Numano R, Abe M, Hida A, Takahashi R, Ueda M,Block GD, Sakaki Y, Menaker M, Tei H. 2000. Resettingcentral and peripheral circadian oscillators in transgenic rats.Science 288, 682±685.

Yang Z, Emerson M, Su HS, Sehgal A. 1998. Response of thetimless protein to light correlates with behavioral entrainment andsuggests a non-visual pathway for circadian photoreception.Neuron 21, 215±223.

Yanovsky MJ, Casal JJ, Whitelam GC. 1995. Phytochrome A,phytochrome B and HY4 are involved in hypocotyl growthresponses to natural radiation in Arabidopsis: weak de-etiolationof the phyA mutant under dense canopies. Plant, Cell andEnvironment 18, 788±794.

Yoshimura T, Ebihara S. 1996. Spectral sensitivity ofphotoreceptors mediating phase-shifts of circadian rhythms inretinally degenerate CBA/J (rd/rd) and normal CBA/N (+/+)mice. Journal of Comparative Physiology A-Sensory Neural andBehavioral Physiology 178, 797±802.

1550 Devlin