The clock in the cell:Entrainment of the circadian clock in Neurospora crassa

All experiments described in this thesis were carried out at the Institute for MedicalPsychology, Ludwig-Maximilians-Universität (University of Munich), Germany, and at theDepartment of Chronobiology, Rijksuniversiteit Groningen, The Netherlands.

The research presented in this thesis was funded by the Deutsche Forschungsgemeinschaft(DFG), the School of Behavioural and Cognitive Neurosciences (BCN) and the University ofGroningen

Layout: Dick VisserPhoto credits for photos at the introduction of chapters:Pamarthi Maruthi Mohan, Osmania University, Hyderabad: Chapter 1, 3, 4, 7 David Jacobson, Stanford University: Chapter 2Naboori B. Raju, Stanford University, California: Chapter 8, 9Print: Van Denderen bv, Groningen

ISBN: 978-90-367-3616-9ISBN digitaal: 978-90-367-3617-6

RIJKSUNIVERSITEIT GRONINGEN

The clock in the cell:Entrainment of the circadian clock in Neurospora crassa

PROEFSCHRIFT

ter verkrijging van het doctoraat in deWiskunde en Natuurwetenschappenaan de Rijksuniversiteit Groningen

op gezag van de Rector Magnificus, dr. F. Zwarts,in het openbaar te verdedigen op

maandag 10 november 2008om 16.15 uur

door

Cornelia Madeti Jyothi, geb. Boesl

geboren op 21 mei 1978te Muenchen, Duitsland

Promotores: Prof. dr. M. MerrowProf. dr. T. Roenneberg

Beoordelingscommissie: Prof. dr. Menno GerkemaProf. dr. Rolf Hoekstra Prof. dr. Michael Brunner

Chapter 1 Introduction 7

Chapter 2 New findings of Neurospora in Europe and comparisons of 27diversity in temperate climates on continental scales

Chapter 3 A population study: Latitudinal clines and chronotype 45of Neurospora wild type strains

Chapter 4 Using entrainment to discover clock genes: A QTL-analysis in 61Neurospora crassa

Chapter 5 Entrainment reveals the photoreceptor gene cryptochrome 79as a clock gene in Neurospora crassa

Chapter 6 Entrainment of the Neurospora circadian clock 97

Chapter 7 Circadian entrainment: the rules of daily synchronization of 109Neurospora crassa in temperature cycles

Chapter 8 Time bandit: the band mutant holds up the wild type 125

Chapter 9 Cellular clocks: circadian rhythms in primary human fibroblasts 135

Chapter 10 Summary 141

Nederlandse samenvatting 151

Dankwoord/Dankword/Acknowledgements 157

Contents

Introduction

1CHAPTER

Circadian clocks

Organisms on earth are influenced by the alternation of day and night caused bythe rotation of the earth. Many – if not all - species have developed strategies tocope with daily changes in the environment, which are to a large extent system-atic and predictable. One of the evolved coping strategies is an endogenous,temporal program that regulates daily rhythms of physiology and behavior. Thisprogram functions as a clock to control endogenous daily rhythms with highprecision. The so-called circadian clock (from Latin “circa diem”, about a day) isendogenous, as was proven by experiments that released plants, animals, fungiand cyanobacteria into constant conditions yielding oscillations with an approxi-mately 24 h period. These are called circadian rhythms.

The internal clock, which regulates aspects of molecular biology, physiologyand social interactions across phyla, generates a temporal program that opti-mizes the sequence of daily events and prepares the organism for upcomingevents. Locomotor activity (Pittendrigh & Daan, 1976a) and photoreception areregulated by the circadian system (Freedman et al, 1999) in animals. Leaf move-ment (Darwin, 1880), growth (McClung, 1992; McClung et al, 1992; Quail,2002), opening of stomata, photosynthesis, cell metabolism (Lüttge, 2000) andgene regulation (Bognar et al, 1999) in plants or spore release in fungi (Merrow& Dunlap, 1994; Roenneberg & Merrow, 2001a) are examples of circadian‘behaviour’ in sessile organisms. Even unicellular organisms, e.g. cyanobacteria(Kondo et al, 1994) or the alga Gonyaulax polyedra, show circadian rhythmicity(Roenneberg & Morse, 1993): in a daily repeating cycle Gonyaulax travels fromthe ocean’s surface (during the day to gather photosynthetic energy) to greatdepths (during the night to harvest nutrients).

That the circadian clock also confers an adaptive advantage has beendescribed experimentally (DeCoursey et al, 2000; Johnson & Golden, 1999; Yanet al, 1998). Therefore selective pressure might be the main driving force for theevolution of circadian rhythms.

One of the reasons that circadian biology is a relevant question is that itconcerns our lives in immediately recognizable ways. There are hundreds ofexamples for how circadian rhythms control physiology (and hence potentiallylead to pathology) in humans by influencing circadian variations in hormonelevels, body temperature, mental and physical performance and pharmaco-kinetics (McFadden, 1988; Moore-Ede et al, 1982a; Moore-Ede et al, 1982b;Rocco et al, 1987).

The huge impact that circadian rhythmicity has on human biology and humansociety (Moore, 1997) can be seen in the example of shift work: about 20% of allemployees in developed countries work in night shifts. The problem here arises

CHAPTER 1

8

from the fact that during a night shift, the circadian clock-regulated physiology ofthe worker usually remains entrained to local time instead of adjusting to thenight shift. So, while physiology and psychology are saying “sleep!”, shift workersare forced to be active and alert (Scott, 2000; Waterhouse et al, 1997).

Clock mechanisms

Given the pervasive effect of circadian biology on life on earth, there is muchinterest in understanding molecular mechanisms using genetic tools. Though themolecular components of the circadian clock are significantly different amonganimals, plants, fungi and cyanobacteria, important features are common acrossphyla: According to Eskin’s model – which very simply describes circadiansystems for essentially all living things – circadian clocks consist of three majorfunctional components: input pathways, a rhythm generator (central oscillator)and output pathways (Fig. 1.1) (Eskin, 1979). Zeitgeber (German for time-giver)signals (the most important of which are light or/and temperature) are trans-duced to the central oscillator via the input pathway.

One of the most important zeitgebers for the circadian system is light. Whilephotoreception for vision requires high time- and high spatial- resolution, circa-dian photoreception must integrate the amount of light over the course of a day,comparable to a scintillation counter (Roenneberg & Foster, 1997; Roenneberg &Merrow, 2000). The use of additional light qualities to tell time-of-day (e.g.,wavelength/color) has not yet been described, even though both blue and redlight have been shown to feed into the circadian clock of plants as inputs.

INTRODUCTION

9

Zeitgebersignal

Inputpathway

Rhythmgenerator

Outputpathway

Figure 1.1 Cartoon of a basic circadian system, that can be described as input pathways thatperceive and transduce external, entraining signals from a zeitgeber (e.g. light, tempera-ture), and a rhythm generator or central oscillator that generates rhythms and regulatesvarious output pathways creating overt rhythms. (re-drawn after Roenneberg and Merrow,1998; Eskin, 1979).

Photoreceptors are the best characterized components of circadian inputpathways. They will either deliver a signal directly (e.g., WC-1 in Neurosporacrassa) or indirectly (e.g., melanopsin in mice) to the rhythm generator/centraloscillator. Concerning cellular clocks (the former case), there will be diversestrategies to carry signals that hold exogenous time information to the endoge-nous circadian program. Plants and fungi and some examples of transparentanimals (like Drosophila) can harvest light intracellularly, allowing for an effi-cient (i.e. comprising only few steps) transfer of information downstream.Concerning complex, hierarchical clocks (the latter example), animals receivelight exclusively through the eyes, and must send signals via the retino-hypothal-amic tract to the circadian pacemaker in the brain, called the suprachiasmaticnucleus (SCN) (Berson et al, 2002; Menaker, 2003). These examples show howbroadly the Eskin model applies to circadian systems.

The SCN is made up of neurons that display a circadian rhythm in geneexpression and neurophysiology, even when dissociated into single cells (Welshet al, 1995). Thus, the complexity of the system is revealed: the SCN functions asone oscillator that orchestrates others such as liver and lung clocks, but it itself ismade up of individual cells that show self-sustained free running and entrainablecircadian rhythms. Hence, the cellular system (input-oscillator-output) parallelsthat of the organism.

Concerning the molecular oscillator mechanism, animals with altered circa-dian properties were generated in mutagenesis experiments, eventually leadingto the discovery of clock genes. The first clock genes were discovered inDrosophila (Konopka & Benzer, 1971), with Neurospora following soon after(Feldman & Hoyle, 1973). Mammalian clock genes were finally revealed, too(King et al, 1997; Lowrey et al, 2000; Ralph & Menaker, 1988). The discoveredcomponents were modeled into a network based on genetic experiments. Thecartoons drawn of circadian networks (Reppert & Weaver, 2002; Schwartz et al,2001) reveal the complexity of circadian systems.

The central oscillator generates self-sustained rhythmicity (see below) of theclock, and then it is the job of output pathways to transduce this oscillatorysignal downstream. One mechanism is via gene expression (transcriptional regu-lation). This was first shown for Neurospora, specifically with the discovery ofclock controlled genes (ccg’s) (Loros et al, 1989). Microarray studies helped toidentify 145 ccg’s, whose predicted or known functions in development, metabo-lism, cell signaling and stress responses suggest a contribution of the circadianclock in a wide range of cell processes (Correa et al, 2003). An example of anoutput pathway in mammals is the one that leads to induction of the vip (vasoac-tive intestinal peptide)-gene (Hurst et al, 2002; Silver et al, 1999). This genetogether with the genes coding for other neuropeptides (vasopressin, cholecys-

CHAPTER 1

10

tokinin and substance P) are used as molecular readouts for circadian rhythmsand represent examples of ccg’s in mammals. Furthermore, more than a hundredgenes have been shown to be under direct clock control (Oishi et al, 2003), andmany more might be, as many microarray analyses in mammals (Delaunay &Laudet, 2002; Duffield, 2003), but also in Arabidopsis (Schaffer et al, 2001), andDrosophila (McDonald & Rosbash, 2001) suggest.

Clock properties

By analyzing the behavior of organisms, properties of their clocks can bededuced, shared features defined and characteristics of circadian clocks ingeneral described (Gwinner, 1986; Pittendrigh, 1960; Roenneberg & Merrow,1998). These include at least the following:● Rhythmicity.

There must be a quantifiable ‘up’ and ‘down’.● Circadian range.

The oscillation has a free-running period (FRP) in the circadian range inconstant conditions, i.e. one full cycle takes approximately 24 hours

● Robustness of the amplitude.The amplitude of the oscillation has to be sufficiently robust to drive outputrhythms

● Self-sustainment. In constant conditions (without zeitgebers) the rhythmicity continuesunabated, and is therefore self-sustained and endogenous. In some organismsthe endogenous circadian rhythmicity can continue over years. (Gwinner,1986; Richter, 1978)

● Entrainability. Circadian systems must be synchronizable to zeitgeber cycles, a propertycalled entrainment (Roenneberg et al, 2003). Hereby, the organism entrainswith a specific relationship, the phase angle, to external cues (like naturallight and temperature cycles, but also food or certain chemicals) keeping itsphysiological functions synchronized with the environment. Circadiansystems are able to entrain to cycle lengths different from 24 hours, but onlywithin a certain range. This property is called the range of entrainment,defined by the minimum and the maximum cycle length (called ‘T’) to whichthe system is still able to entrain. Being exposed to very short or long cyclelengths, an organism can show a frequency demultiplication (e.g., only oneconidial band every two 12 h cycles in Neurospora (Merrow et al., 1999) or afrequency multiplication ((Pittendrigh & Daan, 1976b), e.g., two conidial

INTRODUCTION

11

bands per cycle)). Entrainment differs from driven-ness (a reaction to a zeit-geber stimulus that is uniform in different zeitgeber conditions and does notnecessarily require a circadian clock) in being an active process where theinfluence of timing information on the circadian clock depends on the state ofthe circadian clock at the time of exposure.

● Temperature compensation.Circadian rhythms are highly temperature compensated, i.e. the period isroughly unaltered even when the (constant) temperatures applied vary over arather wide (10°C difference or more) range (Pittendrigh, 1954). Thisphenomenon extends to other parameters like pH inside a cell, nutrition andsocial interaction, as well, and could therefore be termed noise compensation.

The TTO (Transcription-Translation-Oscillator) as a model todescribe molecular clock mechanisms

As mentioned above, clock genes have been identified through mutant screensand they have been constructed in various configurations based largely on molec-ular genetic and genetic experiments. The predominant theory explaining themolecular mechanism of circadian rhythms is that of a Transcription-Translation-regulated-Oscillator (TTO). According to this theory some of the so-called“canonical clock genes” are rhythmically transcribed. Their protein productsnegatively feed back to regulate their own transcription (see Fig. 1.2). InNeurospora crassa, the negative element FREQUENCY (FRQ) feeds back via theWHITE-COLLAR-COMPLEX (WCC) to an element within the promoter region ofthe frq gene (Loros & Dunlap, 2001a). Via posttranslational protein modifica-tions, additional interlocked loops and nuclear import the molecular feedbackprocess is slowed down to occur once per circa 24h period (Lakin-Thomas,2006b). Recent modeling efforts (Roenneberg & Merrow, 2002) show that afreerunning period (FRP) of around 24 hours can be achieved by forming anetwork of several interconnected short-period TTOs. This model furthermoremimics all of the circadian clock properties mentioned above, suggesting that thisis one possibility for how molecular clocks are put together.

The TTO model, as it stands, still fails to explain much of the circadian mech-anism. Anomalies have been accumulating over the last years, including thedemonstration of rhythmicity in organisms with constant clock gene transcrip-tion, and rhythmicity in clock gene knock-out mutants (Bell-Pedersen et al, 2005;Loros & Feldman, 1986; Yang & Sehgal, 2001). It has been suggested thatrhythmic transcription may have other functions in the circadian system (e.g.participating in input and output pathways and providing robustness to the oscil-

CHAPTER 1

12

lations) and that circadian systems might use a non-circadian oscillator consistingof metabolic feedback loops, which acquires its circadian properties from addi-tional regulatory molecules such as the products of canonical clock genes (Lakin-Thomas, 2006b). Rhythmic de-/phosphorylation of clock components, a hypoth-esized ‘phoscillator’, might be common to all circadian systems, as suggested bythe pervasive and prominent role played by kinases and phosphatases in eukary-otic clocks (Merrow et al., 2006).

INTRODUCTION

13

PositiveElement

RNA

Clock protein B

Clock gene B

Clock gene A

NegativeElement

Clock protein A

RNA

Figure 1.2 A simplified circadian transcription/translation oscillator (TTO) model with 2interlocked loops: Clock gene A is transcribed into RNA and translated into protein. Clockprotein B positively regulates transcription of clock gene A. Clock protein A negatively regu-lates its own transcription by interfering with the positive effect of clock protein B. Clockprotein A also positively regulates production of clock protein 2, via either transcription ortranslation. Biosynthetic pathways are shown as solid lines with arrowheads. Positive influ-ence is shown as a circle with plus sign. Negative influence is shown as a dashed linetogether with a square filled with a minus sign. Nuclear/cytoplasmic compartmentation,phosphorylation, degradation pathways, environmental inputs, and outputs to clock-controlled genes and observed rhythms have been omitted. (re-drawn after Lakin-Thomas,2006)

The biology of Neurospora crassa

This thesis employs the model genetic organism Neurospora crassa as an experi-mental tool. Here, I describe basic features of Neurospora ecology.

The filamentous fungus N. crassa belongs to the phylum Ascomycota or ‘sacfungi’, due to the sac-like ascospore containers that are built during sexual propa-gation. Depending on environmental conditions this fungus can propagate asexu-ally or reproduce sexually. Most Neurospora species are haploid and spend mostof their life cycle in this state, because the diploid nuclei formed during thesexual phase are only transient (information taken from http://www.fgsc.net/Neurospora/sectionB2.htm). In its asexual stage, Neurospora forms a mycelium,a network of tubular filaments with multiple haploid nuclei (syncytial hyphae),whereas macroconidia (hereafter called conidia) are formed from aerial hyphae.Conidia do not survive for a long time in nature, but allow for rapid spreadingdue to their huge number. Upon environmental signals (e.g., hydration) conidiagerminate to form hyphae, which grow by tip extension and branch to formmycelia.

Historically, the genus Neurospora was thought to be predominant in moist,tropical and subtropical areas (information taken from http://www.fgsc.net/Neurospora/sectionB4.htm), but recent collection initiatives revealed thatNeurospora even habituates many temperate zones as far North as Alaska(Jacobson et al, 2004). In nature, Neurospora is one of the first colonists in areasof burnt-over vegetation (Jacobson et al, 2006; Perkins & Turner, 1988), and hasas such been described already in 1925 after the fire of Tokyo (Kitazima, 1925).It grows easily indoors on food or food waste, accounting for its commonly usedname “red bread mold”.

Evolving in and adapting to an exposed natural habitat, Neurospora hasdeveloped a variety of light responses including mycelial carotenoid production(Harding & Turner, 1981), formation of sexual structures (perithecia), theirphototropism (Degli-Innocenti et al, 1984; Harding & Melles, 1984), gene expres-sion (Arpaia et al, 1993; Collett et al, 2002; Crosthwaite et al, 1995a; Li & Schmid-hauser, 1995; Sommer et al, 1989) and entrainment of its circadian rhythm.

Neurospora crassa as a tool to study the circadian clock

Neurospora, which spawned the “One gene - One enzyme” hypothesis in the early1940s, is an excellent research tool for several reasons: ● It exists predominantly in a haploid state, e.g. no backcross is needed to

screen Neurospora progeny, which makes reverse and forward genetics easier.

CHAPTER 1

14

● It has a fully sequenced genome (Galagan et al, 2003) making molecularresearch systematic.

● A wealth of genetic and biochemical tools are available from the decades ofbasic genetics work that it has been used for.

● Additionally, Neurospora has a short generation time of a few weeks, andpotentially many progeny and much tissue can be grown in a few days.

However, the key for circadian research is the easily detectable circadian outputbehavior of Neurospora. The standard phenotypic assay to assess circadian rhyth-micity in Neurospora is the ‘race tube assay’, where cultures are grown on solidagar media in glass tubes. Thus grown, N. crassa shows a free running circadianrhythm in conidia formation (banding) of about 22h in darkness. The bands areeasily visualized on solid agar medium (Pittendrigh et al, 1959). Underentraining conditions (e.g., light or temperature cycles), the bands show adistinct phase relationship to external time (Chang & Nakashima, 1997; Lakin-Thomas & Brody, 2000; Merrow et al, 1999b); Roenneberg and Merrow, 2001).In constant light, discrete banding is mostly absent, with conidia being producedcontinuously (Roenneberg & Merrow, 2001; Pittendrigh et al, 1959).

Starting with mutant screening for strains with altered free running periods(FRP’s) in Neurospora, the first Neurospora clock gene, frequency (frq), was foundin the early 1970s (Feldman & Hoyle, 1973). Genetic analyses resulted in thedescription of more than 30 mutant alleles influencing the clock (Feldman &Dunlap, 1983; Lakin-Thomas et al, 1990). Further screening showed that sevenof these 30 mutants were alleles of the frq gene, conferring shorter (e.g. frq1,FRP=16.5 h) or longer period lengths (e.g. frq7, FRP=27 h) than the normal22h. Also, arrhythmic strains, like frq9, carrying a recessive, loss-of-functionmutation, were found (Loros & Feldman, 1986) or subsequently generated (e.g.,frq10, where almost the whole open reading frame (ORF) of the frq-gene isremoved). In the FRQ-deficient mutants light entrainment of the conidiationrhythm is impaired, indicating an additional role for the FRQ-protein in the lightinput system (Merrow et al, 1996; Merrow et al., 2003).

The expression of frq is regulated through transcriptional and posttranscrip-tional control mechanisms (Fig. 1.3). Briefly, the transcription of frq is positivelyregulated by the WHITE-COLLAR-1 (WC-1) and WHITE-COLLAR-2 (WC-2)proteins, and the FRQ protein feeds back negatively on its own transcription(Aronson et al, 1994; Crosthwaite et al, 1997a). Nuclear localization of FRQ isessential for rhythmicity. FRQ enters the nucleus as it is made and represses accu-mulation of frq mRNA (Luo et al, 1998). As mentioned, phosphorylation is acrucial player in the generation of circadian rhythms and FRQ is progressivelyphosphorylated throughout the day and controls the activity of WC-1 and WC-2

INTRODUCTION

15

by regulating their phosphorylation states (Schafmeier et al, 2005). If this phos-phorylation is inhibited experimentally, then the rate of FRQ turnover decreasesand period length increases (Liu et al, 2000). Several kinases, e.g CASEINKINASEs I and II or the calcium/calmodulin-dependent kinase (CAMK) as well asPROTEIN PHOSPHATASE 2A (PP2A) and PROTEIN PHOSPHATASE 1 (PP1),regulate the stability of the FRQ protein and the length of the free-running period(Görl et al, 2001; Liu et al, 2000; Yang et al, 2002).

CHAPTER 1

16

ccgs and otheroutput genes

frq

FRQWC-1 WC-2

VVD

WCC

FLO

wc-1 wc-2

vvd

Figure 1.3 The molecular circadian clock mechanism in Neurospora crassa. frq, wc-1, wc-2,vvd and their gene products form interconnected transcription/translation feedback loopsthat are essential for normal circadian behaviour in Neurospora. The levels of frequency (frq)RNA and FRQ protein depend on WHITE-COLLAR-1 (WC-1) and WHITE-COLLAR-2 (WC-2),which heterodimerize to form the White Collar Complex (WCC). WC-1 levels depend onFRQ. In constant darkness, expression of FRQ protein results in reduced frq RNA accumula-tion. The net effect is two interlinked regulatory loops. Light (shown as flashes) reaches thesystem through the WCC, which is essential for light responses in Neurospora. VVD gates thelight input to the system by interaction with WC-1. ccgs are clock-controlled-genes, some ofwhich are light-induced. The FLO (frq-less-oscillator) has been shown to exist, but compo-nents have not been described yet (Merrow et al., 1999; Loros & Feldman, 1986). Biosyntheticpathways are shown as solid lines with arrowheads. Positive influence is shown as a circle withplus sign. Negative influence is shown as a dashed line together with a square filled with aminus sign. (Re-drawn from (Merrow & Roenneberg, 2001; Heintzen et al, 2001).

Light reception in Neurospora

Several mutations have been reported to affect light responsiveness inNeurospora crassa (Linden et al, 1997). For example mutations in the whitecollar-1 and white collar-2 (wc-1 and wc-2) genes have been shown to impairlight-regulated carotenogenesis. Many other light responses are also abolished inthese mutants (Ninnemann, 1991; Perkins et al, 1982; Russo, 1988), whichmade the WC-1 and WC-2-proteins possible candidates for photoreceptors(Harding & Shropshire, 1980). That wc-1- and wc-2-mutants are also clockmutants was shown later (Crosthwaite et al., 1997) and makes any light inputpathway mutant potentially interesting to use for understanding the mechanismsof the circadian clock in Neurospora. An interesting case is the cytoplasmic bluelight photoreceptor and flavoprotein VIVID (VVD) (Schwerdtfeger & Linden,2003): although the mutant does not display a difference in free running periodin several conditions (Shrode et al, 2001), it regulates entrainment (Elvin et al,2005; Heintzen et al, 2001; Madeti, unpublished data), apparently through itsimpact on photoadaptation.

In addition to VVD, WC-1 and WC-2, the fully sequenced and annotatedgenome (Galagan et al, 2003) of Neurospora crassa (available at:http://www.broad.mit.edu/annotation/genome/neurospora/Home.html)provides additional photoreceptor candidates:● A possible green light photoreceptor protein with high homology to bacteri-

orhodopsin, novel opsin-1 (nop-1), was identified (Bieszke et al, 1999a;Bieszke et al, 1999b). NOP-1 binds retinal and forms a photochemically activepigment (Brown et al, 2001) but neither the physiological function, ingeneral, nor the involvement of this fungal opsin in the circadian system isknown.

● The same is true for the homolog to archaean rhodopsins, ORP-1 (OPSIN-RELATED-PROTEIN 1). Being regulated by heat-shock, it appears to beinvolved in responses to pH, organic solvents and stress (Nemcovic &Borkovich, 2003).

● The genome sequence also reveals a cryptochrome homologue (Daiyasu et al,2004) and two homologues of bacterial phytochromes (Catlett et al, 2003),possible candidates for Red/Far Red photoreceptor genes. phy1 mRNA levelshave been described to be under clock control (Froehlich et al, 2005),whereas knockouts of phy-1 and phy-2 are described not to have an effect onany -so far- known photoresponses. In the same publication, the putativeNeurospora blue light photoreceptor protein CRYPTOCHROME (nCRY) is saidto be photo-regulated by the WC-Complex (Froehlich et al, 2005), but itsfunction is not known yet.

INTRODUCTION

17

● Also, a homolog of the Aspergillus nidulans gene velvet is present inNeurospora. In Aspergillus, this gene is involved in the signal transduction ofboth red and blue light (Yager et al, 1998). The presence of three genes(velvet, phy-1 and phy-2) possibly involved in red light photoreception issurprising given the fact that red light responses have not been described yetin Neurospora – it is thought to be blind for red light (e.g.(Dharmananda,1980; Froehlich et al, 2005) - and suggests that the photobiology inNeurospora might be more complex than recognized, so far.

Action spectra show how much light of a given wavelength is required forsynthesis of mycelial carotenoids (De Fabo et al., 1976), phase shifting of theconidiation rhythm (Dharmananda, 1980), photosuppression of self-sustainedrhythmicity in conidial band formation (Sargent and Briggs, 1967) and the invitro light induced binding of the WC-1 protein to the promotor of the frq gene(Froehlich et al., 2002). In Neurospora crassa, the aforementioned responsescould not be stimulated by wavelengths longer than 520 nm and all of themshow a maximal activity around 460 nm with sensitivity extending into the UVAregion (De Fabo et al, 1976; Froehlich et al, 2002; Sargent & Briggs, 1967). Allthis indicates that flavins or carotenoids are involved as chromophores. However,since a triple albino mutant (al-1, al-2, al-3) containing less than 0.5% of wtcarotenoids is still able to exhibit normal sensitivity for other light responses,photoreception in N. crassa is probably not based on carotenoids (Russo, 1988).Furthermore, mutants deficient in biosynthesis of riboflavin exhibit a decreasedlight sensitivity, which makes flavin species (as have been identified cofactors forWC-1 and VVD) the best candidates for Neurospora photoreceptor chromophores.

Prospect of this thesis

As a young student I was fascinated by fungi and was inspired by a lecture ofProfessor Agerer from the Botanical Institute of the University of Munich. Hedescribed the complex factors interacting to make mushrooms grow in theautumn. Still, I would have never imagined writing a thesis on fungi (especiallythe „red bread mold“ Neurospora crassa) and the complex factors that influencetheir daily timing system. Even more improbable – even though my family tookme on mushroom collection trips as a child - was that one day I would be part ofan international team searching for the first wild Neurospora isolates intemperate climates in Europe.

In my first months as a doctoral student, I had the rare opportunity to see theobject of my further studies in nature after the huge and devastating fires inEurope in 2003. Neurospora crassa is a colonist found on burnt trees after forest

CHAPTER 1

18

fires and was, until 2003, apart from reports from French bakeries, not describedin temperate climates, but rather designated to be an inhabitant of the tropicsand subtropics. Chapter 2 of this thesis describes the findings of the 2003 collec-tion trip in Europe and compares the strain prevalence and growth patterns tothose of the previously known strains.

Chapter 3 describes chronobiological experiments done with a collection ofwild type Neurospora crassa strains from the whole world. These experimentsinclude the newly collected strains from Europe together with the older onesfrom the rest of the world. Strains collected from different latitudes were used toassess the correlations of latitude-of-origin and phase of entrainment or freerunning period.

Chapter 4 also characterizes clock properties in wild type strains, but thistime, with an eye to genetics. Two wild type strains were crossed and 200 to 500progeny were selected for a quantitative trait analysis experiment. While ourcollaborators at U. C. Berkeley generated genetic markers for the strains, weworked on phenotyping them for the circadian clock. Earlier QTL-studies in miceand Arabidopsis indicated that more genes than expected were involved in theclock quantitative traits phase and free-running period. Given this and Neurosporacrassa’s optimal prerequisites for a QTL-study, it is surprising that no earlierstudies existed.

The following chapters set up much of the remainder of the thesis. Whereasthe work with wild type strains, especially using QTL, is one approach to findnovel clock genes, using entrainment for phenotyping a mutant is another way toreach the same end. This approach is ongoing in the lab. For my thesis, however,I took this a step further, combining a functional genetics approach with usingentrainment to reveal new clock genes.

Chapter 5 describes assaying a cryptochrome mutant in Neurospora crassa. Itdisplays the same free running period as the wild type/background strain bdA,but it shows differences in the phase of entrainment in blue or white light cycles.What does this tell us about entrainment and the dogma of „longer period-laterentrainment and shorter period –earlier entrainment“?

In Chapter 6 formal entrainment properties of (a lab strain of) Neurosporacrassa are discussed.

In Chapter 7, I describe the entrainment of N. crassa in a ‚circadian surface’using temperature as a zeitgeber.

Chapter 8 reviews a recent publication describing the cloning of the ‚band’gene. Having been used as a basically universal standard and background strainfor circadian experiments, the discovery of its identity as a ras-1-mutant raisesmany questions about its applicability. The most important –and still not solved-question is: is bd/ras-1 a clock mutant itself?

INTRODUCTION

19

Chapter 9, a review on circadian clocks in human fibroblasts gives an insightinto relatively new circadian research in human tissues. Do in vitro experimentsusing human fibroblasts together with questionnaires have the potential to makethe challenging bunker or constant routine experiments unneccessary? IsNeurospora a parallel relative to human fibroblasts, with respect to circadiansystems? Much of the work that was built up using Neurospora was done beforetissue culture systems were developed for mammalian cells.

References

CHAPTER 1

20

Aronson BD, Johnson KA, Dunlap JC (1994) The circadian clock locus frequency: a singleORF defines period length and temperature compensation. Proc Natl Acad Sci, USA 91:7683-7687

Arpaia G, Loros JJ, Dunlap JC, Morelli G, Macino G (1993) The interplay of light and thecircadian clock. Plant Physiol 102: 1299-1305

Bell-Pedersen D, Cassone VM, Earnest DJ, Golden SS, Hardin PE, Thomas TL, Zoran MJ(2005) Circadian rhythms from multiple oscillators: lessons from diverse organisms. NatRev Genet 6(7): 544-556

Berson DM, Dunn FA, Takao M (2002) Phototransduction by retinal ganglion cells that setthe circadian clock. Science 295: 1070-1073

Bieszke JA, Braun EL, Bean LE, Kang S, Natvig DO, Borkovich KA (1999a) The nop-1 gene ofNeurospora crassa encodes a seven transmembrane helix retinal-binding protein homolo-gous to archaeal rhodopsins. Proc Natl Acad Sci USA 96(14): 8034-8039

Bieszke JA, Spudich EN, Scott KL, Borkovich KA, Spudich JL (1999b) A eukaryotic protein,NOP-1, binds retinal to form an archaeal rhodopsin-like photochemically reactivepigment. Biochem 38(43): 14138-14144

Bognar LK, Adam AH, Thain SC, Nagy F, Millar AJ (1999) The circadian clock controls theexpression pattern of the circadian input photoreceptor, phytochrome B. Proc Natl AcadSci USA 96: 14652-14657

Brown LS, Dioumaev AK, Lanyi JK, Spudich EN, Spudich JL (2001) Photochemical reactioncycle and proton transfers in Neurospora rhodopsin. J Biol Chem 276(35): 32495-32505

Catlett NL, Yoder OC, Turgeon BG (2003) Whole-genome analysis of two-component signaltransduction genes in fungal pathogens. Eukaryot Cell 2(6): 1151-1161

Chang B, Nakashima H (1997) Effects of light-dark cycles on the circadian conidiationrhythm in Neurospora crassa. J Plant Res 110: 449-453

Collett MA, Garceau N, Dunlap JC, Loros JJ (2002) Light and clock expression of theNeurospora clock gene frequency is differentially driven by but dependent on WHITECOLLAR-2. Genetics 160: 149–158

Correa A, Lewis ZA, Greene AV, March IJ, Gomer RH, Bell-Pedersen D (2003) Multiple oscil-lators regulate circadian gene expression in Neurospora. Proc Natl Acad Sci USA100(23): 13597-13602

Crosthwaite SK, Dunlap JC, Loros JJ (1997) Neurospora wc-1 and wc-2: Transcription,photoresponses, and the origin of circadian rhythmicity. Science 276: 763-769

Crosthwaite SK, Loros JJ, Dunlap JC (1995) Light-induced resetting of a circadian clock ismediated by a rapid increase in frequency transcript. Cell 81: 1003-1012

Daiyasu H, Ishikawa T, Kuma K, Iwai S, Todo T, Toh H (2004) Identification of cryptochromeDASH from vertebrates. Genes Cells 9(5): 479-495

INTRODUCTION

21

Darwin C (1880) On the power of movement in plants, London: John Murray.De Fabo EC, Harding RW, Shropshire jr. W (1976) Action spectrum between 260 and 800

nanometers for the photoinduction of carotenoid biosynthesis in Neurospora crassa.Plant Physiol 57: 440-445

DeCoursey PJ, Walker JK, Smith SA (2000) A circadian pacemaker in free-living chipmunks:essential for survival? J Comp Physiol [A] 186(2): 169-180

Degli-Innocenti F, Chambers JA, Russo VE (1984) Conidia induce the formation ofprotoperithecia in Neurospora crassa: further characterization of white collar mutants. JBacteriol 159(2): 808-810

Delaunay F, Laudet V (2002) Circadian clock and microarrays: mammalian genome getsrhythm. Trends Genet 18(12): 595-597

Dharmananda S (1980) Studies of the circadian clock of Neurospora crassa: light-inducedphase shifting. University of California, Santa Cruz, Santa Cruz, CA,

Duffield GE (2003) DNA microarray analyses of circadian timing: the genomic basis ofbiological time. J Neuroendocrinol 15(10): 991-1002

Eskin A (1979) Identification and physiology of circadian pacemaker. Federation Proceedings38: 2570-2572

Feldman JF, Dunlap JC (1983) Neurospora grassa: a unique system for studying circadianrhythms. Photochem Photobiol 7: 319-368

Feldman JF, Hoyle MN (1973) Isolation of circadian clock mutants of Neurospora crassa.Genetics 75: 605-613

Freedman MS, Lucas RJ, Soni B, von Schantz M, Munoz M, David-Gray Z, Foster R (1999)Regulation of mammalian circadian behavior by non-rod, non-cone, ocular photorecep-tors. Science 284(5413): 502-504

Froehlich AC, Liu Y, Loros JJ, Dunlap JC (2002) White Collar-1, a circadian blue lightphotoreceptor, binding to the frequency promoter. Science 297(5582): 815-819

Froehlich AC, Noh B, Vierstra RD, Loros J, Dunlap JC (2005) Genetic and molecular analysisof phytochromes from the filamentous fungus Neurospora crassa. Eukaryot Cell 4(12):2140-2152

Galagan JE, Calvo SE, Borkovich KA, Selker EU, Read ND, Jaffe D, FitzHugh W, Ma LJ,Smirnov S, Purcell S, Rehman B, Elkins T, Engels R, Wang S, Nielsen CB, Butler J,Endrizzi M, Qui D, Ianakiev P, Bell-Pedersen D, Nelson MA, Werner-Washburne M,Selitrennikoff CP, Kinsey JA, Braun EL, Zelter A, Schulte U, Kothe GO, Jedd G, Mewes W,Staben C, Marcotte E, Greenberg D, Roy A, Foley K, Naylor J, Stange-Thomann N,Barrett R, Gnerre S, Kamal M, Kamvysselis M, Mauceli E, Bielke C, Rudd S, Frishman D,Krystofova S, Rasmussen C, Metzenberg RL, Perkins DD, Kroken S, Cogoni C, Macino G,Catcheside D, Li W, Pratt RJ, Osmani SA, DeSouza CP, Glass L, Orbach MJ, Berglund JA,Voelker R, Yarden O, Plamann M, Seiler S, Dunlap J, Radford A, Aramayo R, Natvig DO,Alex LA, Mannhaupt G, Ebbole DJ, Freitag M, Paulsen I, Sachs MS, Lander ES, NusbaumC, Birren B (2003) The genome sequence of the filamentous fungus Neurospora crassa.Nature 422(6934): 859-868

Görl M, Merrow M, Huttner B, Johnson J, Roenneberg T, Brunner M (2001) A PEST-likeelement in FREQUENCY determines the length of the circadian period in Neurosporacrassa. EMBO J 20(24): 7074-7084,

Gwinner E (1986) Circannual rhythms, Vol. 18, Berlin: Springer Verlag.Harding RW, Melles S (1984) Genetic analysis of the phototrophism of Neurospora crassa

parethecial beaks using white collar and albino mutants. Plant Physiol 72: 996-1000Harding RW, Shropshire W (1980) Photocontrol of carotenoid biosynthesis. Annu Rev Plant

Physiol 31: 217-238Harding RW, Turner RV (1981) Photoregulation of the carotenoid biosynthetic pathway in

albino and white collar mutants of Neurospora crassa. Plant Physiol 68: 745-749

CHAPTER 1

22

Hurst WJ, Mitchell JW, Gillette MU (2002) Synchronization and phase-resetting by gluta-mate of an immortalized SCN cell line. Biochem Biophys Res Commun 298(1): 133-143

Jacobson DJ, Dettman JR, Adams RI, Boesl C, Sultana S, Roenneberg T, Merrow M,Duarte M, Marques I, Ushakova A, Carneiro P, Videira A, Navarro-Sampedro L, OlmedoM, Corrochano LM, Taylor JW (2006) New findings of Neurospora in Europe andcomparisons of diversity in temperate climates on continental scales. Mycologia 98(4):550-559

Jacobson DJ, Powell AJ, Dettman JR, Saenz GS, Barton MM, Hiltz MD, Dvorachek WH,Glass NL, Taylor JW, Natvig DO (2004) Neurospora in temperate forests of westernNorth America. Mycologia 96: 66-74

Johnson CH, Golden SS (1999) Circadian programs in cyanobacteria: adaptiveness andmechanism. Annu Rev Microbiol 53: 389-409

Kitazima K (1925) On the fungus luxuriantly grown on the bark of trees injured by the greatfire of Tokyo on Sept. 1, 1923. Nihon Shokubutso Byori Gakkai Ho (Ann Phytopathol SocJpn) 1(6): 15-19

King DP, Zhao Y, Sangoram AM, Wilsbacher LD, Tanaka M, Antoch MP, Steeves TD, VitaternaMH, Kornhauser JM, Lowrey PL, Turek FW, Takahashi JS (1997) Positional cloning ofthe mouse circadian clock gene. Cell 89(4): 641-653

Kondo T, Golden SS, Ishiura M, Johnson CH (1994) Circadian rhythms of cyanobacteriaexpressed form a luciferase reporter gene. In Evolution of circadian clock, Hiroshige T,Honma K-I (eds). Sapporo: Hokkaido University Press

Konopka R, Benzer S (1971) Clock mutants of Drosophila melanogaster. Proc Natl Acad SciUSA 68: 2112-2116

Lakin-Thomas PL (2006) Transcriptional feedback oscillators: maybe, maybe not. J BiolRhythms 21(2): 83-92

Lakin-Thomas PL, Brody S (2000) Circadian rhythms in Neurospora crassa: Lipid deficienciesrestore robust rhythmicity to null frequency and white-collar mutants. Proc Natl Acad SciUSA 97: 256-261

Lakin-Thomas PL, Coté GG, Brody S (1990) Circadian rhythms in Neurospora crassa:Biochemistry and Genetics. Crit Rev Microbiol 17: 365-416

Li C, Schmidhauser TJ (1995) Developmental and photoregulation of al-1 and al-2, struc-tural genes for two enzymes essential for carotenoid biosynthesis in Neurospora. DevBiol 169(1): 90-95

Linden H, Rodriguez-Franco M, Macino G (1997) Mutants of Neurospora crassa defective inregulation of blue light perception. Mol Gen Genet 254: 111-118

Liu Y, Loros J, Dunlap JC (2000) Phosphorylation of the Neurospora clock proteinFREQUENCY determines its degradation rate and strongly influences the period lengthof the circadian clock. Proc Natl Acad Sci USA 97: 234-239

Loros JJ, Denome SA, Dunlap JC (1989) Molecular cloning of genes under control of thecircadian clock in Neurospora. Science 243(4889): 385-388

Loros JJ, Dunlap JC (2001) Genetic and molecular analysis of circadian rhythms inNeurospora. Annu Rev Physiol 63: 757-794

Loros JJ, Feldman JF (1986) Loss of temperature compensation of circadian period length inthe frq-9 mutant of Neurospora crassa. J Biol Rhythms 1(3): 187-198

Lowrey PL, Shimomura K, Antoch MP, Yamazaki S, Zemenides PD, Ralph MR, Menaker M,Takahashi JS (2000) Positional syntenic cloning and functional characterization of themammalian circadian mutation tau. Science 288: 483-491

Luo C, Loros JJ, Dunlap JC (1998) Nuclear localization is required for function of the essen-tial clock protein FRQ. EMBO J 17: 1228-1235

Lüttge U (2000) The tonoplast functioning as the master switch for circadian regulation ofcrassulacean acid metabolism. Planta 211: 761-769

INTRODUCTION

23

McClung CR (ed) (1992) The higher plant Arabidopsis thaliana as a model system for themolecular analysis of circadian rhythms. NY: Marcel Dekker

McClung CR, Davis CR, Page KM, Denome SA (1992) Characterization of the formate (for)locus, which encodes the cytosolic serine hydroxymethyltransferase of Neurosporacrassa. Mol Cell Biol 12(4): 1412-1421

McDonald MJ, Rosbash M (2001) Microarray analysis and organization of circadian geneexpression in Drosophila. Cell 107(5): 567-578

McFadden ER, Jr. (1988) Circadian rhythms. Am J Med 85(1B): 2-5Menaker M (2003) Circadian rhythms. Circadian photoreception. Science 299(5604): 213-

214Merrow M, Brunner M, Roenneberg T (1999) Assignment of circadian function for the

Neurospora clock gene frequency. Nature 399: 584-586Merrow MW, Dunlap JC (1994) Intergeneric complementation of a circadian rhythmicity

defect: phylogenetic conservation of structure and function of the clock gene frequency.EMBO J 13: 2257-2266

Moore RY (1997) Circadian rhythms: basic neurobiology and clinical applications. Ann RevMed 48: 253-266

Moore-Ede MC, Lydic R, Albers HE, Tepper B (1982a) Three-dimensional structure of themammalian suprachiasmatic nuclei: A comparative study of five species 204: 225-237

Moore-Ede MC, Sulzman FM, Fuller CA (1982b) The Clocks that time us, Cambridge:Harvard University Press.

Ninnemann H (1991) Participation of the molybdenum cofactor of nitrate reductase fromNeurospora crassa in light promoted conidiation. J Plant Physiol 137: 677-682

Oishi K, Miyazaki K, Kadota K, Kikuno R, Nagase T, Atsumi G, Ohkura N, Azama T, MesakiM, Yukimasa S, Kobayashi H, Iitaka C, Umehara T, Horikoshi M, Kudo T, Shimizu Y, YanoM, Monden M, Machida K, Matsuda J, Horie S, Todo T, Ishida N (2003) Genome-wideexpression analysis of mouse liver reveals CLOCK-regulated circadian output genes.J Biol Chem 278(42): 41519-41527

Perkins DD, Radford A, Newmeyer D, Björkman M (1982) Chromosomal loci of Neurosporacrassa. Microbiol Rev 46: 426-570

Perkins DD, Turner BC (1988) Neurospora from natural populations: Toward the PopulationBiology of a Hapoid Eukaryote. Experimental Mycology 12: 91-131

Pittendrigh CS (1954) On Temperature Independence in the Clock System ControllingEmergence Time in Drosophila. Proc Natl Acad Sci USA 40(10): 1018-1029

Pittendrigh CS (1960) Circadian rhythms and the circadian organization of living systems.Cold Spring Harb Symp Quant Biol 25: 159-184

Pittendrigh CS, Bruce VG, Rosensweig NS, Rubin ML (1959) Growth patterns in Neurosporacrassa. Nature 184: 169-170

Pittendrigh CS, Daan S (1976a) A functional analysis of circadian pacemakers in nocturalrodents: I. The stability and lability of circadian frequency. J Comp Physiol A 106: 223-252

Pittendrigh CS, Daan S (1976b) A functional analysis of circadian pacemakers in nocturnalrodents: V. Pacemaker structure: a clock for all seasons. J Comp Physiol A 106: 333-355

Ralph MR, Menaker M (1988) A mutation of the circadian system in golden hamsters.Science 241: 1225-1227

Reppert SM, Weaver DR (2002) Coordination of circadian timing in mammals. Nature 418:935-941

Richter CP (1978) Evidence for existence of a yearly clock in surgically and self-blindedchipmunks. Proc Natl Acad Sci USA 75(7): 3517-3521

Rocco MB, Nabel EG, Selwyn AP (1987) Circadian rhythms and coronary artery disease. AmJ Cardiol 59(7): 13C-17C

CHAPTER 1

24

Roenneberg T, Daan S, Merrow M (2003) The art of entrainment. J Biol Rhythms 18(3):183-194

Roenneberg T, Foster RG (1997) Twilight Times - Light and the circadian system. PhotochemPhotobiol 66: 549-561

Roenneberg T, Merrow M (1998) Molecular circadian oscillators: an alternative hypothesis.J Biol Rhythms 13(2): 167-179

Roenneberg T, Merrow M (2000) Circadian light input: omnes viae Romam ducunt. CurrBiol 10: R742-R745

Roenneberg T, Merrow M (2001) Seasonality and photoperiodism in fungi. J Biol Rhythms16: 403-414

Roenneberg T, Merrow M (2002) Life before the clock - modeling circadian evolution. J BiolRhythms 17(6): 495-505

Roenneberg T, Morse D (1993) Two circadian oscillators in one cell. Nature 362: 362-364Russo VE (1988) Blue light induces circadian rhythms in the bd mutant of Neurospora:

double mutants bd,wc-1 and bd,wc-2 are blind. Photochem Photobiol 2(1): 59-65Sargent ML, Briggs WR (1967) The Effects of Light on a Circadian Rhythm of Conidiation in

Neurospora. Plant Physiol 42(11): 1504-1510Schaffer R, Landgraf J, Accerbi M, Simon V, Larson M, Wisman E (2001) Microarray analysis

of diurnal and circadian-regulated genes in Arabidopsis. Plant Cell 13(1): 113-123Schafmeier T, Haase A, Kaldi K, Scholz J, Fuchs M, Brunner M (2005) Transcriptional feed-

back of Neurospora circadian clock gene by phosphorylation-dependent inactivation ofits transcription factor. Cell 122(2): 235-246

Schwartz WJ, de la Iglesia HO, Zlomanczuk P, Illnerova H (2001) Encoding le quattrostagioni within the mammalian brain: photoperiodic orchestration through the suprachi-asmatic nucleus. J Biol Rhythms 16(4): 302-311

Schwerdtfeger C, Linden H (2003) VIVID is a flavoprotein and serves as a fungal blue lightphotoreceptor for photoadaptation. EMBO J 22(18): 4846-4855

Scott AJ (2000) Shift work and health. Prim Care 27(4): 1057-1079Shrode LB, Lewis ZA, White LD, Bell-Pedersen D, Ebbole DJ (2001) vvd is required for light

adaptation of conidiation-specific genes of Neurospora crassa, but not circadian conidia-tion. Fungal Genet Biol 32(3): 169-181

Silver R, Sookhoo AI, LeSauter J, Stevens P, Jansen HT, Lehman MN (1999) Multiple regula-tory elements result in regional specificity in circadian rhythms of neuropeptide expres-sion in mouse SCN. Neuroreport 10(15): 3165-3174

Sommer T, Chambers JA, Eberle J, Lauter FR, Russo VE (1989) Fast light-regulated genes ofNeurospora crassa. Nucl Ac Res 17: 5713-5723

Waterhouse J, Minors D, Redfern P (1997) Some comments on the measurement of circa-dian rhythms after time-zone transitions and during night work. Chronobiol Int 14(2):125-132

Welsh DK, Logothetis DE, Meister M, Reppert SM (1995) Individual neurons dissociatedfrom rat suprachiasmatic nucleus express independently phased circadian firingrhythms. Neuron 14: 697-706

Yager LN, Lee HO, Nagle DL, Zimmerman JE (1998) Analysis of fluG mutations that affectlight-dependent conidiation in Aspergillus nidulans. Genetics 149(4): 1777-1786

Yan OY, Andersson CR, Kondo T, Golden SS, Johnson CH, Ishiura M (1998) Resonatingcircadian clocks enhance fitness in cyanobacteria. PNAS 95(15): 8660-8664

Yang Y, Cheng P, Liu Y (2002) Regulation of the Neurospora circadian clock by casein kinaseII. Genes Dev 16: 994-1006

Yang Z, Sehgal A (2001) Role of molecular oscillations in generating behavioral rhythms inDrosophila. Neuron 29(2): 453-467

INTRODUCTION

25

New findings of Neurospora in Europeand comparisons of diversity in temperate climateson continental scales

D. J. Jacobson, J. R. Dettman, R. I. Adams, C. Boesl, S. Sultana,T. Roenneberg, M. Merrow, M. Duarte, I. Marques, A. Ushakova, P.Carneiro, A.Videira, L. Navarro-Sampedro, M. Olmedo, L. M.Corrochano, J. W. Taylor

Published in Mycologia 98 (4), (2006): 550-559

2CHAPTER

ABSTRACT

The life cycles of the conidiating species of Neurospora are adapted to respond to fire, whichis reflected in their natural history. Neurospora is found commonly on burned vegetationfrom the tropic and subtropical regions around the world and through the temperate regionsof western North America. In temperate Europe it was unknown whether Neurospora wouldbe as common as it is in North America because it has been reported only occasionally. In2003 and 2004 a multinational effort surveyed wildfire sites in southern Europe. Neurosporawas found commonly from southern Portugal and Spain (37ºN) to Switzerland (46ºN).Species collected included N. crassa, N. discreta, N. sitophila and N. tetrasperma. The speciesdistribution and spatial dynamics of Neurospora populations showed both similarities anddifferences when compared between temperate Europe and western North America, bothregions of similar latitude, climate and vegetation. For example the predominant species inwestern North America, N. discreta phylogenetic species 4B, is common but not predominantin Europe, wherea species rare in western North America, N. crassa NcB and N. sitophila, aremuch more common in Europe. The meiotic drive element Spore killer was also common inEuropean populations of N. sitophila and at a higher proportion than anywhere else in theworld. The methods by which organisms spread and adapt to new environments are funda-mental ecosystem properties, yet they are little understood. The differences in regionaldiversity, reported here, can form the basis of testable hypotheses. Questions of phylogeog-raphy and adaptations can be addressed specifically by studying Neurospora in nature.

Introduction

The conidiating species of the ascomycete fungus Neurospora, as a group, havebeen considered to be primarily tropical or subtropical with a complete longitu-dinal distribution (Turner and Perkins 1988, Turner et al 2001). These particularNeurospora species are well adapted to grow and sporulate on the surface of fire-scorched vegetation. Recent field surveys, however, have found that Neurosporacommonly occupies an entirely different ecological niche, in dry and/or coldhabitats. Within this new geographic range, western North America from NewMexico (34ºN) to Alaska (64ºN) (Jacobson et al 2004), Neurospora was foundunder the bark of firedamaged trees. This discovery has raised questions aboutthe occurrence of Neurospora in other temperate regions. The purpose of thisstudy was to determine whether Neurospora is common in temperate regions ofEurope. We hypothesized that the niche under the bark of burned vegetation hadbeen overlooked in Europe as it was in North America. In autumn 2003 a multi-national effort searched for Neurospora in fire sites across southern Europe after asummer of unusually devastating wildfires. Additional collections were made in2004.

Most published accounts of Neurospora in temperate regions were anecdotal(see Jacobson et al 2004). In Europe Neurospora most often has been associatedwith bakeries, (Legan 1993, Perkins 1991, Perkins and Turner 1988, Yassin andWheals 1992). High temperatures and the presence of easily colonized substratesthat usually are associated with bakeries may allow Neurospora to grow in loca-tions that traditionally were considered outside the geographic distribution ofthis fungus. However observations of Neurospora in nature have been sporadic inEurope with no systematic surveys or descriptions of population on the scale ofstudies in temperate North America.

Individuals collected in Europe were identified with both biological and phylo-genetic species recognition methods that have been developed for the outbreedingspecies of Neurospora. Phylogenetic species recognition also provided a prelimi-nary indication of genetic diversity within species. The comparison of the isolatescollected in this study with those from North America and throughout the worldhighlights differences in the ecology of Neurospora and the diversity ofNeurospora populations in temperate climates on different continents.

CHAPTER 2

28

Material and methods

Collection, culturing and species identificationAn international consortium was formed to survey for Neurospora in Europe insummer 2003 during which an unusually intense heat wave led to devastatingwildfires.

Fire progression was followed and fire maps obtained from the Global FireMonitoring Center website, http://www.fire.uni-freiburg.de/current/globalfire.htm, and links therein.

Satellite photos were obtained when possible from the Moderate ResolutionImaging Spectroradiometer (MODIS) Rapid Response system, near-real- timeproduction website, http://rapidfire.sci.gsfc.nasa.gov/production.

These maps and photographs were used to locate easily accessible and widelydistributed sites across southern Europe. Initial surveys were made in Portugal inearly Sep 2003. Systematic field work was conducted in late Sep and early Oct2003 at sites in Switzerland, northern Italy, southern France and northern Spain.Additional collections from Sapiaos, Portugal, and Seville, Spain, were maderespectively in Sep and Oct 2004.

Methods of handling isolates, including collecting, initial culturing, subcul-turing of single conidia and storage, were exactly as described in Jacobson et al(2004). A field sample of conidia was collected from a sporulating colony ontosterile filter paper, which then was placed in a sterile envelope. One colony perplant was sampled for up to 45 isolates per site. In addition, where possible, twoto seven isolates from the same plant were collected from one or two plants persite. Representative isolates of each species found at each site (both mating typeswhen possible), and strains (Table 2.I see below) have been deposited in theFungal Genetic Stock Center (FGSC), Kansas city, Missouri 64110 (http://www.fgsc.net) under accession numbers 10010– 10059.

Recent taxonomic work (Cai et al 2006, García et al 2004) has not changedthe status of the conidiating species of Neurospora. Therefore biological speciesrecognition was used to identify isolates to species with a three-step processfollowing methods outlined in Perkins and Turner (1988):(i) Assessment of heterothallism:

A single conidium subculture from each isolate was allowed to grow onVogel’s minimal medium N (Davis 2000) at 25ºC for 7–10 d to test for self-fertility. Perithecia from each self-fertile isolate were dissected to determinethe number of ascospores per ascus. All isolates with four ascospores perascus were concluded to be N. tetrasperma.

(ii) Mating-type (mat) determination:Each self-sterile (heterothallic) isolate was crossed to both mat A and mat a

NEW FINDINGS

29

species tester strains of N. crassa that contain the fluffy mutation (FGSCstrains 6682 and 6683, respectively). Conidia of unknown isolates (males)were used to fertilize protoperithecia of the tester strains (females) growingon Petri dishes of Westergaard’s synthetic crossing medium (Davis 2000).Fertilization was successful when conidia were applied to a small region ofthe tester colony, so that a single female tester on a 9 cm diam plate could befertilized with up to 30 different isolates. A darkening and swelling ofprotoperithecia indicated a mating reaction after 2–4 d incubation at 25ºC. Apositive mating reaction on one female tester was obtained for each isolate,thus revealing mating type.

(iii)Mating with species testers: When crossed to the N. crassa tester any isolate that produced >50% blackascospores after 7–10 d postfertilization was classified as N. crassa (Perkinsand Turner 1988). Isolates that produced only hyaline, unviable ascospores orno spores at all were judged not to be N. crassa. This response also confirmedthat none of these isolates were N. intermedia, which routinely produces5–10% black ascospores with the N. crassa tester strains (Perkins and Turner1988). Each isolate was crossed, again as a male, to plates of tester strainfemales of the appropriate mating type for both N. sitophila (FGSC 5940 matA or 5941 mat a) and N. discreta (FGSC 3228 mat A or 4378 mat a). Fertilityto these testers was mutually exclusive. Production of black ascospores waslimited to crosses with one and only one of the species testers; no isolatemade black ascospores with more than one tester. In addition no isolate wasinfertile with all Neurospora species testers.

Characterization of the genetic diversity among N. crassaand N. discreta strainsPhylogenetic analyses of N. crassa and N. discreta have revealed geneticallydistinct clades within these species (Dettman et al 2003a, 2006). To assignEuropean isolates to these clades, or to discover other clades within these biolog-ical species, sequence was obtained for three diagnostic polymorphic DNAregions (Dettman et al 2003a). Sequences of the three polymorphic regions(unlinked, noncoding loci that flank microsatellites [TMI, TML, and DMG]) wereobtained with methods described by Dettman et al (2003a). Sequences werealigned manually, because of the presence of microsatellites and insertion/dele-tion gaps (indels) within these loci. Microsatellite sequences were omitted fromthe analyses. Fourteen of 17 European N. discreta isolates were analyzed(excluding multiple isolates of the same mating type from the same plant), aswere 22 N. crassa isolates from all sites where N. crassa was present, includingmultiple isolates of different mating type where available (Table 2.I). The only

CHAPTER 2

30

NEW FINDINGS

31

Species, clade and isolate numbers

FGSC number D number* Mating type Country, site

N. crassa NcB10049 mat A Spain, Platja d’Aro10050 mat A Spain, Platja d’Aro10033 mat a Spain, Macanet de la Selva10043 mat a Spain, Seros10044 mat A Spain, Seros10045 mat a Spain, Seros10046 mat A Spain, Seros10017 mat A Portugal, Troviscal Sertã10018 mat A Portugal, Penedo Furado10020 mat A Portugal, Tapada de Mafra10021 mat a Portugal, Tapada de Mafra10024 mat A Portugal, Tapada de Mafra10027 mat a Portugal, Monchique10028 mat a Portugal, Monchique10036 mat a Italy, Turchino Est.10037 mat a Italy, Turchino Est.10038 mat a Italy, Turchino Est.10040 mat a Italy, Turchino Est.10042 mat a Italy, Turchino Est.10051 mat A Italy, Genoa10054 mat a Italy, Genoa10056 mat A Italy, Genoa

N. discreta9991 D221 mat A Spain, Macanet de la Selva9990 D220 mat A Portugal, Monchique9989 D218 mat A Portugal, Monchique10025 mat a Portugal, Monchique10010 mat a Portugal, Boticas10011 mat a Portugal, Boticas9986 D215 mat a Portugal, Boticas9987 D216 mat A Portugal, Boticas9988 D217 mat a Portugal, Boticas10012 mat a Portugal, Boticas10013 mat A Portugal, Boticas10014 mat A Portugal, Boticas9992 D224 mat A Switzerland, Leuk9993 D225 mat A Switzerland, Leuk

*D numbers refer to isolate numbers given by Dettman et al (2006) in the phylogenetic study of N. discreta.

Table 2.1 European isolates of Neurospora used in phylogenetic analyses.

two N. crassa isolates obtained from western North America (Montana; FGSC8571 and W-864) (Jacobson et al 2004) also were included. The sequences havebeen deposited in GenBank under accession numbers DQ442288–DQ442377.The sequences of the three loci were combined into a single dataset becauseprevious use of the partition homogeneity test showed a lack of incongruence(Dettman et al 2003a, 2006). Separate maximum parsimony trees were calcu-lated for N. discreta and N. crassa with PAUP* (version 4.0b10, Swofford 2003).Analysis of European N. discreta isolates in relation to worldwide collections of N.discreta sensu lato has been reported by Dettman et al (2006). For comparativepurposes the N. crassa dataset included sequences of the three loci from a subsetof 37 of the N. crassa strains included in Dettman et al (2003a). No outgroupswere included, because Dettman et al (2003a) clearly showed that N. crassa is awell supported phylogenetic species. Maximum parsimony bootstrapping for N.crassa was performed with heuristic searches (1000 replicates, simple stepwiseaddition, tree bisection-reconnection branch swapping, MAXTREES 5 100).

CHAPTER 2

32

A B

Figure 2.1 Neurospora growing and sporulating on scorched vegetation in Europe.A. Extensive colonization of an unidentified shrub at Turchino Est., Italy.B. Localized sporulation limited to the node of cane-like grass at Seros, Spain.

Results

The occurrence of Neurospora in wildfire sitesThe yellow to orange colonies of Neurospora conidiating on the surface of woodyand herbaceous plants killed by fire were recognized easily (Fig. 2.1). Neurosporawas found to be common at some sites, while being relatively rare at others sites(Table 2.2, Fig. 2.2). The 14 sites surveyed extend over ca. 1650 km in a pathleading generally northeast from southern Portugal (37º18’N, 8º35’W) toSwitzerland (46º19’N, 7º38’E). The number of isolates collected totals 247 andincludes the species N. crassa, N. discreta, N. sitophila and N. tetrasperma. At fiveof the 14 sites collections were from a single species of Neurospora, whereascollections at the other nine sites yielded multiple species. Most of these isolates(195) were single colonies collected from an individual plant. Multiple colonies(2–7) were sampled from 13 plants across five sites. Eleven of these plantsyielded a single species, but two plants were colonized by two Neurospora specieseach. Although no systematic attempt was made to gauge the level of clonality ormeasure intraspecific genetic diversity among the isolates, multiple genotypes ofthe same species were found on five of the plants inhabited by a single species ofNeurospora (see below).

Spore killer in N. sitophilaIsolates identified as N. sitophila could be separated into two classes based oncrosses with the species tester strains. One class produced 90–95% blackascospores, whereas the other produced 50% black ascospores with theremaining spores being hyaline, significantly smaller and unviable. Whenperithecia from these crosses were dissected microscopically, nearly every ascusshowed a 4:4 black:hyaline ascospore pattern (Fig. 2.3). This pattern is the hall-mark of Spore killer meiotic drive in Neurospora (Raju 2002). Because the testerstrains of N. sitophila used (FGSC 5940 mat A and 5941 mat a) are known to besensitive to Spore killer, the killer component must be present in the European N.sitophila isolates.

A single spore killer element, Spore killer (Sk-1), has been described in N.sitophila (Raju 2002, Turner 2001). Research with Sk-1 has shown that only killerX sensitive heterozygous crosses show killing; both homozygous crosses, killer Xkiller and sensitive X sensitive, show normal 8:0, black:hyaline, ascospores ineach ascus.

Therefore, to determine if the Spore killer in European isolates is Sk-1 or anew element, each N. sitophila strain was crossed to Sk-1 testers strains (FGSC2216 mat A or 2217 mat a). All European isolates that showed 4:4 killing whencrossed to sensitive produced 8:0 asci when crossed to Sk-1. Likewise all

NEW FINDINGS

33

European isolates that produced 8:0 asci when crossed to sensitive showed 4:4killing when crossed to Sk-1. All killer European N. sitophila isolates, therefore,are Sk-1; no new killer elements were apparent in these samples.

Fifty-four percent of the N. sitophila isolates collected (45 of 83) expressedthe killer phenotype. The killer haplotype was present in six of the seven sitescontaining N. sitophila, and three of these contained both killer and sensitivehaplotypes. However killer and sensitive haplotypes were not found together onany of the five plants from which multiple isolates of N. sitophila were recovered.

CHAPTER 2

34

N. crassa N.sitophilab N. discreta N. tetra-sperma

Country Site Latitude Longitude matA mat a matA mat a matA mat a matA+a

Portugal Monchique 37°18’ 8°35’W 2 12 5 1(0:12)

Tapade de 38°58’ 9°17’W 2 1 3 1Mafra (3:0)Penedo Furado 39°38’ 8°10’W 1 1Troviscal Sertã 39°52’ 8°0’W 1 4Boticas 41°42’ 7°41’W 3 5 1Sapiãos 41°43 7°37’W 10

Spain Sevilla 37°24’ 5°59’W 3 23Seros 41°23’ 0°19’E 20 22Macanet de la 41°46’ 2°43’E 1 2 5 1Selva (2:0) (4:1)Platja d’Aro 411°50’ 3°4’E 6 4 2

(2:0)France Vidauban 43°24’ 6°28’E 15Italy Genova 44°26’ 8°45’E 4 20 22 8

(17:5) (8:0)Turchino Est. 44°27’ 8°44’E 5 14 9

(0:14) (3:6)Switzerland Leuk 46°19’ 7°38’E 6

(6:0)

Totals 37 78 49 34 11 6115 83 17 32

% of total (247) 46% 34% 7% 13%

aAll isolates are totaled here, including those collected from the same plant. No systematic attempt was madeto identify clones which may have been repeatedly sampled. Characterization of a small number of geneticmarkers, for a limited number of isolates, was conducted for phytogenetic clade identification and treeconstruction (see Table 2.1, Fig 2.4 and text).bThe ratio of SK-1 killer to sensitive isolates is in parentheses.

Table 2.2 The distribution of species of Neurospora across sites surveyed in Europe in2003–2004a.

NEW FINDINGS

35

time (days)

40°N

N. crassa

45°N

10°W 5°W 0 5°E 10°E 15°E

N. sitophilaN. discretaN. tetrasperma

Figure 2.2 Distribution of Neurospora biological species collected in Europe.

A

B

Figure 2.3 Asci from crosses of European N. sitophila Spore killer strains. A. A cross hetero-zygous for Sk-1 (killer X sensitive). Asci contain four normal size, maturing, viable Sk-1ascospores and four hyaline, aborted sensitive ascospores. B. A homozygous (sensitive 3senstive) cross for Sk-1. Asci contain eight normal size, maturing, viable ascospores.Homozygous Sk-1 killer X killer crosses also show asci containing eight viable ascospores(Photomicrographs courtesy of N.B. Raju, Stanford University).

N. crassaA single maximum parsimony tree was produced from combined sequences ofthe TMI, TML, and DMG loci (Fig. 2.4). Included in the tree were representativesof the three major clades in N. crassa, NcA, NcB and NcC (Dettman et al 2003a).Sequence was obtained for all three loci from 22 European isolates of N. crassa.All these isolates fell into the single, previously described clade NcB (Table 2.I,Fig. 2.4).

CHAPTER 2

36

D69 Ivory Coast

D90 Yucatan

D96 Ivory Coast

D113 Louisiana

D115 Louisiana

D118 Louisiana

D119 Louisiana

D140 Ivory Coast

D144 Panama

D117Louisiana

D61Haiti

D143 Louisiana

D110 Louisiana

D111 Louisiana

D116 Louisiana

D68 Ivory Coast

D91 Yucatan

D60 Haiti

D62 HaitiD27 Florida

D19 Florida

D28 Florida

D23,D30FloridaD85 Yucatan

D105 Tamil Nadu

D107 Tamil NaduD42 Tamil Nadu

D99 Tamil Nadu

D104Tamil Nadu

D103Tamil Nadu D100

Tamil Nadu

D98Tamil Nadu

D106Tamil Nadu

D70Ivory Coast

D11,D12Karnataka

W1265Portugal

W1247Portugal

W1331 ItalyPortugal, Italy, Spain

(9 isolates)Portugal, Italy, Spain

(10 isolates)W619, W864 Montana

NcBNcC

NcA

10083

1 change

Figure 2.4 The relationships among the three phylogenetic clades within N. crassa. Maximumparsimony, unrooted phylogram produced from the sequences of three combined loci.Numbers next to bold branches separating NcA, NcB and NcC clades indicate bootstrapsupport (1000 replicates). Taxon labels indicate strain number and geographic source;European strains from this study are shown within the box, and western North Americanstrain from Jacobson et al (2004) are shown in the shaded box. All strains labeled with Dnumbers were sequenced as part of Dettman et al (2003a). One locus (TMI) was sequencedfrom an additional 93 European N. crassa isolates because this locus is diagnostic for N.crassa clade. All 93 isolateswere placed definitively into NcB based on TMI sequence (datanot shown).

The sequence of the TMI locus subsequently was obtained from theremaining 93 isolates of European N. crassa to associate each with the appro-priate phylogenetic clade. TMI was chosen as a diagnostic locus because itssequence is clearly distinct between the NcB clade versus clades NcA and NcC.All 93 isolates fell within the NcB clade (data not shown). Of these 83 had TMIsequences that were essentially identical, including the number of microsatelliterepeats. Ten isolates, all from Seville, Spain, were exceptional in having a singlenucleotide polymorphism at base 119 in the microsatellite flanking sequence anda microsatellite with 5 rather than 12 repeats.

Although it was beyond the scope of this study to assess clonality of strainsfrom the same plant, we did investigate the genotypes of multiple N. crassaisolates collected from seven individual plants. When two polymorphic markers(mat and TMI) were combined, five plants from Seville, Spain, revealed morethan one genetically distinct individual of N. crassa per plant. In contrast themultiple isolates of N. crassa from the two other plants (from Seros, Spain, andGenoa, Italy) were monomorphic at both markers. This preliminary study indi-cated that more than one genetic individual could be present in very close spatialscales, as was reported by Powell et al (2004).

Phylogenetic species 4B within the N. discreta complexThe European isolates of N. discreta sensu lato, as defined by biological speciesrecognition, all were identified as belonging to phylogenetic species (PS) 4B (treenot shown, refer to Dettman et al 2006 Fig. 2 for relationships among phyloge-netic species within the N. discreta complex). Of the six isolates sequenced herethat were not analyzed by Dettman et al (2006), each had sequence identical toat least one isolate examined by Dettman et al (2006). Therefore no additionalgenetic diversity was found within PS 4B or the European population, and PS 4Bis the only species of the N. discreta complex found in Europe to date.

Discussion

Reports of the occurrence of Neurospora in Europe have been published sporadi-cally over the past 160 y, beginning with its earliest description from France in1843 (see Perkins 1991). Most of these descriptions have concentrated onNeurospora contamination of bakeries and their products; the most recent wasYassin and Wheals (1992). Not long after formal description of the genus byShear and Dodge (1927), however, Ramsbottom and Stephens (1935)mentioned that Neurospora was found on other natural substrates, most notablyburnt trees and gorse in Britain. Other anecdotal observations have suggested

NEW FINDINGS

37

that Neurospora is not uncommon in Europe (e.g. D. Zickler, University ParisSouth, personal communication with D.D. Perkins, Stanford University).However, to our knowledge, this is the first study that systematically sampledNeurospora from natural habitats in Europe.

All five classically described, conidiating, biological species of Neurosporahave now been identified in Europe: N. crassa, N. discreta, N. intermedia, N.sitophila and N. tetrasperma. This is the first study to report N. discreta, whereaswe did not find N. intermedia, which was reported by Ramsbottom and Stephens(1935). Recent work has further divided Neurospora into phylogenetic speciesand clades (Dettman et al 2003a, 2006). Of the eight phylogenetic species withinthe N. discreta complex only one (PS 4B) was identified among the Europeanisolates collected here. Two newly described phylogenetic species outside the N.discreta complex also were found to be distinct biological species (Dettman et al2003b); neither of these species were found among the European isolates. Of thethree distinct clades within N. crassa (NcA, NcB and NcC), all new European N.crassa isolates fell into NcB. Based on these finer scale measures of genetic diver-gence among members of Neurospora, similarities and differences were assessedbetween the newly sampled populations from Europe and populations fromother continents, including both temperate and tropical/subtropical climates.

The similarity of Neurospora between Europe and southeastern, subtropicalareas of the United States is also reflected in the overall species diversity anddistributions. The complement of species and their frequency of collection aresimilar in Europe and southeastern United States (Fig. 2.5). This distribution is instark contrast to populations of Neurospora in western North America, which are

CHAPTER 2

38

0

20

40

60

80

100

perc

enta

ge

western NorthAmerica

southeasternUnited States

Europe

N. crassaN. sitophilaN. discreta

N. tetraspermaN. intermedia

Figure 2.5 Frequency of Neurospora biological species by region. Data for western NorthAmerica taken from Jacobson et al (2004), data for southeastern United States taken fromTurner (2001).

composed pre dominantly of a single species in the N. discreta complex (PS 4B),with only rare occurrences of N. sitophila and N. crassa. Neurospora has beenfound at 64ºN latitude in Alaska and as far as 45ºN in Europe. Future collectingexpeditions are planned to target even higher latitudes in Europe in the hope oflearning more about the distribution of Neurospora species.

The absence of N. intermedia in our European collection was unexpectedgiven reports in the literature (Ramsbottom and Stephens 1935, Yassin andWheals 1992). For example Yassin and Wheals (1992) reported nine of 345isolates (<3%) as N. intermedia, eight of which were from nonbakery sources,including imported Indonesian ontjom. Moreover N. intermedia is by far the mostcommon species collected world wide, particularly at latitudes >30ºN and S(China, Japan, Australia and New Zealand) (Turner et al 2001). Given that thereare likely to be sources of N. intermedia in Europe, the lack of N. intermedia inour collections from temperate northern latitudes in Europe and western NorthAmerica is intriguing but its significance cannot be assessed currently.

The physical appearance in nature of N. crassa and N. discreta from Europeand the southeastern US is remarkably similar and unlike that of N. discreta fromwestern North America. However phylogenetic analysis of DNA sequences indi-cated that the European isolates of the two species were highly similar to thosefrom temperate western North America and dissimilar to those found in thesoutheastern United States.

N. crassa clades NcA, NcB and NcC are genetically distinct from one anotherbut do not meet the strict criteria that would make them separate phylogeneticspecies (Dettman et al 2003a). These clades have distinct geographical distribu-tions. NcA was widespread across the Caribbean basin and Africa. NcC waslimited to the state of Tamil Nadu in India, and the rare isolates of NcB werelimited to equatorial Africa and southern India. The addition of all the Europeanand western North American isolates of N. crassa to clade NcB significantlychanges the biogeography of the species. Now NcB also appears geographicallywidespread, similar to NcA, although its prevalence outside of Europe remains inquestion.

Distributions of NcA and NcB in the western hemisphere and Europe arenonoverlapping, but the clades do coexist in equatorial Africa. NcA and NcC,whose ranges overlap in southern India, have developed reproductive isolationphenotypes, which correlate with the genetic distance (Dettman et al 2003b; E.Turner, University of California at Berkeley, unpublished). No attempt was madein this or previous studies to characterize the reproductive relationships betweenmembers of the NcA and NcB clades; biological species recognition was limited tocrossing European isolates to the species tester strains. We therefore do not knowwhether NcA and NcB show reproductive isolation anywhere in their range.

NEW FINDINGS

39

The N. discreta complex from Europe and North America also shows a combi-nation of widespread and more narrowly distributed species. European N. discretaisolates, which represent only 7% of all collected European isolates, are placedphylogenetically within the predominant species in western North America, thewidespread PS 4B. PS 4B however is phylogenetically distant from the two otherN. discreta species in North America (i.e., N. discreta sensu stricto [Texas] and PS 7[Florida, Mexico and Guatemala]). Striking differences were seen in the growthhabit and the morphology of colonies of Neurospora on natural substratesbetween the two temperate continents. As mentioned, Neurospora in Europe wascommonly seen apparently growing on the surface of charred bark (Fig. 2.1A). Incontrast extensive colonies of Neurospora were seen in western North America butalways under the bark of woody plants. Only rarely, and after prolonged incuba-tion periods, did the fungus erupt through the bark (Jacobson et al 2004, Fig. 1).The extensive amount of sporulation on the surface of burned bark, as seen inEurope, was never observed in western North America. Moreover recognizablecolonies of Neurospora were not observed under the bark in Europe.

Neurospora in Europe grew on both charred woody and herbaceous plants,such as the grass in Spain (Fig. 2.1B). In contrast Neurospora was never observedin western North America on herbaceous plants. Although the latitude, climate,geography and vegetation are similar between Europe and temperate westernNorth America, the growth habit and substrate of Neurospora in Europe aresimilar to those in tropical and subtropical areas, including Florida and Texas inthe southern United States (Powell et al 2003,Turner and Perkins 1988, Turner etal 2001).

The proportion of N. sitophila Spore killer strains reported here for Europe ismuch higher (45 of 83, 54%) than worldwide (77 of 469, 16%) (Fig. 2.5) andmight provide an opportunity to study the dynamics of Spore killers in natureand the effect of meiotic drive on populations. Existing data have been insuffi-cient to determine the potential of Spore killer to become fixed in any population(Turner 2001), which makes the spatial and temporal dynamics of killer andsensitive haplotypes in European populations of great interest (Burt and Trivers2006). Whether a stable equilibrium is maintained can be tested only wherekiller and sensitive coexist in the same populations, but Sk-1 killer and sensitivehaplotypes coexist from only 10 out of 92 (11%) sites where Spore killer hasbeen found outside of Europe: one in Hawaii, one in Vanuatu and eight in Tahiti.In Europe, as mentioned, killer and sensitive coexist in three of six sites with Sk-1frequency of 13–83%. European Spore killer isolates also were reported by Yassinand Wheals (1992) who found that all nine of their bakery N. sitophila isolateswere Sk-1. Re-sampling of European N. sitophila populations over time mightprovide the data needed to understand Spore killer and meiotic drive in nature.

CHAPTER 2

40

Together with the recent discovery of Neurospora in western North America(Jacobson et al 2004), documentation presented here of its occurrence in Europefirmly establishes it as a common inhabitant of temperate climates, perhapsworldwide. The broad distribution of N. discreta, particularly its longitudinalcomponent, place it, along with N. crassa, among the handful of species thathave the attributes to serve as evolutionary and ecological model organisms.There will be no lack of ecological questions because of the large gaps in ourknowledge of the basic ecology of Neurospora and fire adapted fungi in general.The differences in regional diversity, reported here, can form the basis of testablehypotheses. Questions of phylogeography and adaptations specifically can beaddressed with Neurospora: Where did these species of Neurospora originate andhow did they arrive at their modern distributions? What role have human activi-ties played in the current distribution of Neurospora lineages? Have populationsof Neurospora changed genetically to adapt to local conditions, such as ambienttemperature or photoperiod (Tan et al 2004)? How organisms spread and adaptto new environments are fundamental ecosystem properties, yet they are littleunderstood. We hope that understanding of these fundamental features willcome from studies of N. discreta that blend ecology and evolutionary biologywith genetics and genomics.

References

NEW FINDINGS

41

Acknowledgements and contributions of autorsD.J. conceived and coordinated the project with the assistance of J.T., M.M., T.R., A.V., andL.M.C. D.J., C.B. and S.S. collected and cultured most strains. M.D., I.M., A.U., P.C. and A.V.collected and cultured Portugal strains, identifying some. L.M.C., L.N.S. and M.O. collectedin Seville and provided needed logistical help in locating other collection sites in Spain. D.J.identified or confirmed identification of all strains. J.D., R.A. and D.J. sequenced andperformed phylogenetic analyses. D.J. analyzed the data and with J.T. wrote the paper. Thework was supported by a grant from the US National Science Foundation to J.T. (DEB-0316710); by grants from the Deutsche Forschungsgemeinschaft and the Meyer-StruckmanStiftung to T.R.; by grants from Fundac¸a˜o para a Ciencia e a Tecnologia to A.V.; and bygrants from the Ministerio de Educacio´n y Ciencia, Spain, (INIA RM2004-00007) andJunta de Andalucı´a (CVI 0119) to L.M.C. D.J. also is supported in part by US NationalScience Foundation grant MCB-0417282, awarded to David D. Perkins, Stanford University.We thank David Perkins for allowing part of this work to be completed in his laboratory andN.B. Raju for the photomicrographs used in Fig. 2.3.

Burt A, Trivers R. 2006. Genes in conflict: the biology of selfish geneticelements.Cambridge, Massachusetts: Belknap Press of Harvard University Press. 602 p.

Cai L, Heewon R, Hyde KD. 2006. Phylogenetic investigations of Sordariaceae based onmultiple gene sequences and morphology. Mycol Res 110:137–150.

Davis RH. 2000. Neurospora: contributions of a model organism. New York: OxfordUniversity Press. 333 p.

CHAPTER 2

42

Dettman JR, Jacobson DJ, Taylor JW. 2003a. A multilocus genealogical approach to phylo-genetic species recognition in the model eukaryote Neurospora. Evolution 57:2703–2720.

———, ———, ———. 2006. Multilocus sequence data reveal extensive phylogeneticspecies diversity within the Neurospora discreta complex. Mycologia 98:437–447.

———, ———, Turner E, Pringle A, Taylor JW. 2003b. Reproductive isolation and phyloge-netic divergence in Neurospora: comparing methods of species recognition in a modeleukaryote. Evolution 57:2721–2741.

Garcı´a D, Stchigel AM, Cano J, Guarro J, Hawkworth DL. 2004. A synopsis and recircum-scription of Neurospora (syn. Gelasinospora) based on ultrastructural and 28S rDNAsequence data. Mycol Res 108:1119–1142.

Jacobson DJ, Powell AJ, Dettman JR, Saenz GS, Barton MM, Hiltz MD, Dvorachek WH,Glass NL, Taylor JW, Natvig DO. 2004. Neurospora in temperate forests of western NorthAmerica. Mycologia 96:66–74.

Legan JD. 1993. Mould spoilage of bread: the problem and some solutions. Int BiodeteriorBiodegrad 32:33–53.

Perkins DD, Turner BC. 1988. Neurospora from natural populations: toward the populationbiology of a haploid eukaryote. Exp Mycol 12:91–131.

———. 1991. The first published scientific study of Neurospora, including a description ofphotoinduction of carotenoids. Fung Genet Newsl 38:64–65.

Powell AJ, Jacobson DJ, Natvig DO. 2003. Variation among natural isolates of Neurospora onsmall spatial scales. Mycologia 95:809–819.

Raju NB. 2002. Spore killers: meiotic drive elements that distort genetic ratios. In: OsiewaczHD, ed. Molecular biology of fungal development. New York: Marcel Decker Inc. p275–296.

Ramsbottom J, Stephens FL. 1935. Neurospora in Britain. Trans Brit Mycol Soc 19:215–220.Shear CL, Dodge BO. 1927. Life histories of and heterothallism of the red bread-mold fungi

of the Monilia sitophila group. J Ag Res 34:1019–1042.Swofford DL. 2003. PAUP*: phylogenetic analysis using parsimony (*and other methods).

Version 4.0.b10. Sunderland, Massachusetts: Sinauer Associates.Tan Y, Merrow M, Roenneberg T. 2004. Photoperiodism in Neurospora crassa. J Biol Rhythm

19:135–43.Turner BC. 2001. Geographic distribution of Neurospora Spore killer strains and strains

resistant to killing. Fung Genet Biol 32:93–104.———, Perkins DD, Fairfield A. 2001. Neurospora from natural populations: a global study.

Fung Genet Biol 32:67–92.Yassin S, Wheals A. 1992. Neurospora species in bakeries. J Appl Bacteriol 72:377–380.

NEW FINDINGS

43

A population study: Latitudinal clinesand chronotype of Neurospora wild type strains

C. Madeti, T. Radiç, D.J. Jacobson, T. Roenneberg, M. Merrow

Manuscript

3CHAPTER

ABSTRACT

The length of day and night (photoperiod) and changes in temperature are signals used byorganisms to adjust their biology to the environment. Predicting seasonal changes –just likepredicting daily changes- is thought to be mediated by the circadian system. This principlehas been shown for example in the filamentous fungus Neurospora crassa, where one of thecentral players of the circadian clock, the gene frequency (frq) is indispensable for photoperi-odism. If the clock is essential for seasonal behaviour, then it follows that clock characteris-tics would vary aaccording to the latitude of origin. This principle has been shown inDrosophila. Here we report a phenotypic analysis of wild type strains of Neurospora crassafrom different latitudes assessed for circadian traits in light and temperature cycles.Furthermore, we examine the correlation with the latitude of the collection. We found ahighly significant correlation of phase (in a light-dark-cycle) and latitude, with strains fromhigher latitudes entraining earlier compared to strains from the lower latitudes. Theseresults were confirmed and extended by the correlation of latitude and length of twilight,with strains from the North entraining earlier in longer twilights.

Introduction

The ascomycete Neurospora crassa is one of the leading model organisms to studythe circadian clock. As for many model genetic organisms, the ecology ofNeurospora crassa in nature and its life cycle are not completely described. Sexualand asexual reproduction, as well as dormancy and growth are probably tempo-rally distributed –clearly separated from each other- up to several months (Pandit& Maheshwari, 1994). The asexual spores (conidia) are short-lived and can growalmost anywhere, if sufficient nutrients are available. The sexual ascospores areequipped with a tough wall which preserves them over many years. To activatean ascopore it usually takes temperatures of about 50 degrees Celsius. Hightemperatures and readily available carbon sources might make burnt wood aperfect growth medium. So, in nature, Neurospora is mostly found after fires onor below the bark of burnt trees, where the orange conidia are conspicuous, inpart due to their carotenoids. Carotenoids in conidia might serve severalpurposes: firstly, it has been shown that carotenoids help diminish mutations viaUV-irradiation (Sies & Stahl, 2004) and secondly, the orange ‚blossom’ that canbe seen from afar attracts insects that can spread the conidia further (Shaw,1990; Shaw, 1998). It has been speculated that there might be an alternateoxygen- independent life cycle in Neurospora which allows for growth under-neath the bark of trees until a fire propagates the aerobic form that is commonlyknown (M. Merrow, D.J. Jacobson, personal communication).

Due to its obvious banding pattern (bouts of asexual spore formation)Neurospora crassa has been a valued model organism for the study of circadianrhythms since the late 1950’s (Pittendrigh et al, 1959). Researchers doingchronobiological experiments on Neurospora strains collected from the wild (asopposed to strains created in the laboratory = lab strains) often face a problem:wild type strains grow much faster than the lab „wild-type“strain band (bd).Furthermore, race tube experiments are complicated by obscured rhythmicity inthe ‚real’ wild type isolates. Several attempts have been made to overcome thesedifficulties, mostly by adjusting media. One method involves addition of Rb-chlo-ride (Gall & Lysek, 1981) and another uses acetate/casamino acids together withfructose as an alternative carbon source (Morgan & Feldman, 1998). Anotherapproach calls for ‚inverted race tubes’ to minimize the suppression of conidialbanding by CO2-accumulation in race tubes over the course of an experiment(Sargent & Kaltenborn, 1972; Park & Lee, 2004). In a recent publication on themutation underlying bd, it was suggested that reactive oxygen species (ROS) thataccumulate during an experiment inside race tubes are the cause of band forma-tion. bd is a strain carrying a single point mutation in the ras1-gene, which leadsto an imbalance in ROS levels. Hyperoxidant states are hypothesized to trigger

CHAPTER 3

46

cell differentiation and might therefore also confer the banding phenotype(Belden et al, 2007). Vitamin K (menadione) or other pro-oxidant ROS-gener-ating agents also increase ROS-levels and therefore can be used to convert wildtype strains into banding strains (Belden et al, 2007).

Our interest in world-wide Neurospora populations coincides with intenseinterest in understanding how we (as humans) adjust our physiology to dailychanges in our environment with help of our circadian clock. On the experi-mental front, researchers are applying quasi-natural conditions (e.g. twilight asopposed to square-wave light pulses (Boulos et al, 1996) or are studying organ-isms collected from the wild or directly in their natural environment (Daan &Aschoff, 1975; Everts et al, 2004; van Oort et al, 2005). Recent work focusedalso on the variabilty of circadian qualities depending on the geographic origin(so-called latitudinal clines). Several studies show their existence, for example inwild Drosophila strains. While in one study (Joshi, 1999) locomotor activity wasdemonstrated to be variable and latitude-dependent, Pittendrigh and Takamuraanalysed the circadian rhythm of eclosion activity and the endogenous rhyth-micity (i.e. the pacemaker) in a series of Drosophila auraria races coming fromdifferent latitudes of 34.2° to 42.9° in Japan (Pittendrigh & Takamura, 1989a).They found that the phase relationship of the rhythm to the daily photoperiodchanges as day length increases, and that the amplitude of the rhythm changeswith latitude. The amplitudes of Drosophila phase response curves (PRCs) wereobserved to be lower in the North and free running periods (FRPs) were longerin northern than in southern races. In rodents, latitudinal clines have beenshown, where the ratio of morning over evening PRC slopes increases as theperiod lengthens (Pittendrigh & Takamura, 1989a). North-south differences inthe phase relation of both the pacemaker and the rhythm in relation to the lightcycle were explained by latitudinal clines in pacemaker properties and depend-ency on day- (with a FRP longer than 24 h) or night-activity (with a FRP shorterthan 24 h).

Measuring day or night-length might be one way in which plants or animalscan track these changes. That seasonal adjustment uses the circadian system hasbeen shown e.g. in the filamentous fungus Neurospora crassa, where the centralclock components FREQUENCY (FRQ) and WHITE-COLLAR-1 (WC-1) are indis-pensable for photoperiodic responses (Tan et al, 2004). Several lines of evidencesuggest that non-reproductive events (e.g. carotenogenesis) as well as propaga-tion (asexual spore formation, i.e. conidiation) and reproduction (sexual sporeproduction) show systematic responses to different photoperiods (Tan et al,2004; Rémi, 2007). In Neurospora as in many other organisms, like Drosophila(Pittendrigh & Takamura, 1987), measuring night-length is the way by which thecircadian system detects seasonal changes (Roenneberg & Merrow, 2001b).

LATITUDINAL CLINES

47

Experimentally, this has been demonstrated mostly with light (photoperiods), butthere is some indication that temperature plays a role in this process, also (seechapter 6, this thesis).

The circadian system with a free running period (FRP) close to 24 hoursenhances fitness (Hotta et al, 2007; Young & Kay, 2001). However, a certain vari-ability of clock properties over a broad range of latitudes and longitudes is surelyindispensable for the survival of an organism (for example in migratory birds, infungi that are spread via spores, to name just a few).

Until recently, the geographical range of Neurospora crassa was thought to belimited to the tropics and subtropics, where it would theroretically be subjectedto only minor changes in photoperiod. Recent findings have Neurospora as farNorth as Alaska (Jacobson et al, 2004). Neurospora species are endemic inEurope, from Portugal to Scotland. The sum of the collections of older strainscollected in the tropics and subtropics together with the newly collected wildNeurospora strains essentially make a world-wide population. Here, we haveused this collection to study natural variability of circadian behavior and itscorrelation with latitude.

Material and methods

StrainsThe Neurospora crassa strains utilized are wild type strains collected from lati-tudes between 4º5’S (e.g. strain FGSC-No. 8834 from Mandingo, Congo) and55º57’N (e.g. strain FGSC-No. 1672 from Edinburgh, Scotland). Unless otherwiseindicated, three replicates per strain have been assessed. For the twilight experi-ment, 6 replicates were evaluated. In our laboratory several approaches have been used to assess circadian pheno-types of Neurospora strains from the wild. Our method uses the standard glucose-arginine medium described by Sargent and Kaltenborn (Sargent & Kaltenborn,1972) containing 1X Vogel’s solution (Vogel, 1956), 0.5% Arginine, 10µl/100mlBiotin, 2% Agar and no glucose (Munich Minimal Medium).

Light and temperature cyclesAll strains were assayed in both light-dark- and temperature cycles, as well asconstant darkness (DD at 25ºC). Phase of entrainment and free running period(FRP) were assessed. Furthermore we assessed whether phases and periodscomply with the phase-period rule, i.e. whether phase and period are correlatedin Neurospora wild type strains.

CHAPTER 3

48

All light-dark cycles were performed at a constant temperature of 25°C andwere carried out in light-tight boxes (Rémi, 2007). An air-circulating fan wasused to minimize temperature effects during illumination phases with a whitefluorescent tube light (OSRAM, ca. 4µE). A layer of diffuser was insertedbetween the light source and the race tubes to improve light distribution. Lightcycles applied were 12h light/12h darkness (12/12 LD) and 2h light/10h dark-ness (2/10 LD, skeleton photoperiods). All temperature cycles were performed inwater baths which allow gradual temperature steps up and down (e.g. 22/27 ºC,see details in chapter 6). The cycle length was 24h with alternating cool (22 ºC)and warm (27 ºC) periods of 12 hours each.

The temperature cycles applied here display rather slow transitions (ca. 1.5hours) from warm to cold, as opposed to light cycles, where light is given aspulses in steps up or down.

Analysis of period lengths and phase was done using Chrono (Versions 6.4mto 6.7.1m (Roenneberg & Taylor, 2000)). As phase reference point we used onsetof conidiation relative to lights-off (shown in the figures as Φon, ‘phi on’).

LATITUDINAL CLINES

49

120°W 90°W 60°W 30°W 0° 30°E 60°E 90°E 120°E

30°S

0°

30°N

60°N

60°S

Figure 3.1 Collection sites of Neurospora crassa strains used in this study. The latitudes thestrains are collected from (shown as dots) range from 4º 5’S (e.g. strain FGSC-No. 8834from Mandingo, Congo) to 55º57’N (e.g. strain FGSC-No. 1672 from Edinburgh, Scotland).

Results

To measure free running period (FRP) we subjected all strains to complete dark-ness at a constant temperature of 25 ºC. These conditions are completely stan-dard except for the use of media without glucose, a concession to the bandingphenotype of the wild type strains. Under these conditions, the circadian controlstrain, bd, extends ist FRP from 22h to ca. 24 hours. The distribution of FRP’s inwild type strains can be seen in Fig. 3.2, together with the distributions of strainscollected from Europe and the strains from the rest of the world. The separaterepresentation was chosen due to differences in the (phylo) genetics and prove-nience of the strains from Europe (Jacobson et al, 2006). The average FRP’s wereslightly shorter (23.4 h) in the European compared to the non-European strains(23.8 h), with most (55%) strains having FRP’s between 23 and 25 hours (Fig.3.3 A to B).

The distribution of periods is ranging from 18 to 30 hours, whereas on theextreme ends only few strains have been observed. Most strains show FRPs in therange of 21 to 26 hours. We found no correlation between period and latitude(Fig. 3.4), even when European and Non-European strains were analysed sepa-rately (data not shown).

Given the broad distribution in the wild type population strains concerningtheir free running period, we expected to see a broad distribution in entrainedphase, as well. This is a simple and logical extension of the ‘phase-period rule’.

CHAPTER 3

50

0

10

20

5

25

15

num

ber

of s

train

s

18 20 22 24 26 28 30

FRP (h)

allEuropean

non-European

Figure 3.2 The distribution of free running periods (FRP’s, in h) in Neurospora crassa wildtype strains. The y-axis indicates the number of strains. Black bars represent all strains fromour collection (n = 97, average FRP = 23.5 h), dark grey bars strains from Europe (n = 66,averagef FRP = 23.4 h) and light grey bars strains from the rest of the world (n = 31,average FRP = 23.8 h).

Phase of entrainment (onset of conidiation relative to lights-off) was calculatedin all strains in 2/10 LD, 25 ºC. Surprisingly, we found no correlation, neitherwhen all strains were compared nor when only strains from outside of Europe orEurope were considered (Fig. 3.5).

For example in Drosophila ananassae, the activity rhythm varies according tolatitude (with early phase coninciding with lower latitude (Joshi, 1999)). Since

LATITUDINAL CLINES

51

0

10

20

5

25

15

perc

enta

ge

18 20 22 24 26 28 30

all strains

B

18 20 22 24 26 28 30

FRP (h)

non-European strains

A

18 20 22 24 26 28 30

European strains

C

Figure 3.3 The distribution of free running periods (FRPs, in h) in Neurospora crassa wildtype strains in %. In total 97 strains were assessed (panel A), of these 31 were collectedoutside of Europe (panel B) and 67 in Europe (panel C).

0

18

20

22

24

26

28

30

FRP

(h)

10 20 30 40 50 60latitude (°)

Figure 3.4 Free running period (in h) in DD/25 ºC versus latitude (in º) in Neurosporacrassa wild type strains. The data shows no correlation (p-value = 0.0732).

the entrained state reflects pacemaker characteristics of an organism, differencesin phase might be latitude-dependent, as well. Thus, in a next step we correlatedΦon to latitude and found a highly significant correlation (p-value = 0.0011)(Fig. 3.6).

CHAPTER 3

52

280

60

120

180

240

300

360Φ

on (

°)

18 20 22 24 26FRP (h)

30

all strainsnon-European strains

Figure 3.5 Φon (‘phi on’, onset of conidiation relative to lights-off, in º) in 2/10 LD vs FRP(in h) in DD/25 ºC in Neurospora crassa wild type strains (n = 49). The values are not corre-lated, neither in all strains (grey line, p-value = 0.68) nor in the strains from outside ofEurope (black line, p-value = 0.11)

00

60

120

180

240

300

360

Φon

(°)

10 20 30 40 50latitude (°)

Figure 3.6 Φon (onset of conidiation relative to lights-off, in º) of Neurospora crassa wildtype strains (n = 78) in 2/10 LD vs latitude (in º). The values are significantly correlated (p-value = 0.0011).

In previous experiments (see also chapter 4) we had seen that wild typestrains and their progeny are sometimes more robustly synchronised (judged byeye) when temperature is used as an entraining stimulus. Therefore, in anotherset of experiments, we used a temperature cycle (12/12 h 27/22 ºC) to deter-mine phase of entrainment.

When the distributions of wild type strains assessed in a light cycle and atemperature cycle were compared, different curve shapes can be observed(shown in Fig. 3.7). The curve yielded for Φon in a light cycle is much broadercompared to the curve yielded by the temperature cycle. The phases are signifi-cantly different in both cycles and have a different mean Φon with 112.7º (2/10LD) and 87.14º (12/12 h 27/22ºC). Thus, the average Φon in the light cycle is 1h 42 min later compared to the temperature cycle.

Then we plotted the phase in the temperature cycle vs free running period totest whether the phase-period-rule applies. We found a highly significant correla-tion of phase and period (p-value = 0.0026, Fig. 3.8).

Interestingly, when phase in the temperature cycle was graphed versus lati-tude-of-origin we did not find a correlation (Fig. 3.9). This was the case for allstrains, and also for non-European and European strains examined separately.

LATITUDINAL CLINES

53

28

5

20

10

15

0

perc

enta

ge

16 20 24Φon (°) 32

12840

phase in 2/10 LDphase in 12/12h 27/22°C

Figure 3.7 (Distribution of phases (Fon, onset of conidiation relation to lights/warm off in º)in a light (2/10 LD) and a temperature cycle (12/12h 27/22ºC). The phases have beencalculated referring to the middle of the dark/cold phase for comparison.

To study the effects of twilight on entrainment in Neurospora wild type strainswe selected 8 strains from latitudes of 1.78º to 55.95ºC and subjected them to12/12 h light cycles with twilights of different lengths (ranging from 0 to 0.5, 1,1.5, 2 and 2.5 hours).

CHAPTER 3

54

28

-60

60

120

180

240

0

Φon

(°)

16 20 24FRP (h)

32

Figure 3.8 Φon (onset of conidiation referring to warm-off, in º) in a temperature cycle of12/12h at 27/22 ºC vs FRP in DD/25 ºC (in h) in Neurospora crassa wild type strains (n =66). The values are highly correlated (p-value = 0.0026).

0 10 20 30 40 50 60latitude (°)

0

60

120

180

240

300

Φon

(°)

non-European strainsEuropean strains

Figure 3.9 Φon (onset of conidiation referring to warm-off, in º) in a temperature cycle of12/12 h at 27/22 ºC vs latitude of origin in Neurospora crassa wild type strains. The valuesare not correlated, also when non-European or European were looked at separately.

Averaged phases measured in these 8 wild type strains showed a significantphase advance (related to the midpoint of twilight) with longer twilight. Whengraphed separately, however, we found the strains from higher latitudes toadvance their phase with increasing twilight, whereas the strains from lower lati-tudes where almost not affected and showed a stable phase relationship to themiddle of the dark phase (Fig. 10).

LATITUDINAL CLINES

55

1.5-4

4

1216

24

0

0.0 0.5 1.0Φon (h)

2.0

20

8

2.5

14D1 44.43°

1.50.0 0.5 1.0Φon (h)

2.0 2.5

1672 55.95°

-4

4

1216

24

0

20

8

8873 29.80° 1545 37.40°

-4

4

1216

24

0

20

8

8859 12.67° 1451 18.48°

-4

4

1216

24

0

20

8

1455 1.78° 883D 7.43°

Figure 3.10 Φon (onset of conidiation referring to the midpoint of twilight, in h) in12/12LD-cycles with different twilight lengths ranging from 0 to 2.5 h (x-axis) in 8Neurospora wild type strains from different latitudes of 1.78º to 55.95º (latitude-of-origingiven next to each panel). Light is shown in white, darkness in dark grey, twilight in lightgrey; midnight is indicated as dashed line. The FGSC-numbers of the strains used are givenin the upper left corner of each panel.

Discussion

Correlations of free-running period and the latitude-of-origin of an organismhave been established. We therefore set up a series of experiments to elucidatelatitudinal clines in Neurospora crassa. To this end, we assessed the followingcircadian phenotypes in Neurospora crassa isolates from the wild:● free running period in DD/25 ºC● phase of entrainment in a light cycle (2/10 LD, 25 ºC)● phase of entrainment in a temperature cycle (12/12 h 27/22 ºC)● phase of entrainment in twilight (twilight of 0 - 2.5 hours in 12/12 LD, 25 ºC,

8 strains only)Subsequently, these have been correlated to latitude, and additionally phase ofentrainment in both light and temperature cycles related to FRP.

Our correlations deviate from results obtained by Pittendrigh and Takamura(Pittendrigh & Takamura, 1989), in that we did not find a correlation of FRP andlatitude. Since, in nature, Neurospora is probably never exposed to constant dark-ness, the ‚period’ trait is likely not selected for. Hence, if there is a relationbetween the circadian system and the latitude at which it evolved, period is onlyas good as its relationship to phase. Hence, phase of entrainment in light cycleswas determined and correlated to period. Contrary to circadian theory – andwhat has been shown many times e.g. in humans (Daan & Aschoff, 1975; Duffyet al, 1999; Duffy et al, 2001; Jones et al, 1999b) - we did not find a correlationof phase derived from light dark cycles and period. A similar finding wasobserved in Arabidopsis (Michael et al, 2003), where different quantitative traitloci (QTL) could be found to affect either period or phase of entrainement. Thisfinding suggests that the traits phase and period are likely not under control ofthe same genes. Given the complexity of the circadian system - from input path-ways to oscillator with likely several interconnected feedback loops and outputpathways-, it can be argued that polymorphisms in output pathway genesaffecting phase do not neccessarily also need to affect period.

Interestingly, we found a highly significant correlation of phase (in LD) andlatitude-of-origin. This is consistent with results from Arabidopsis (Michael et al,2003) and Drosophila studies (Joshi, 1999). As pointed out before, phase ofentrainment is a trait under evolutionary selection and could as such representthe entrainment properties of strains depending on their latitude-of-origin.

As a next set-up for the characterization of Neurospora wild type strains wehave chosen a temperature cycle. In previous work, the progeny of two wild typestrains was studied in temperature cycles and gave a completely surprising result:in DD/25 ºC and in several light cycles, relatively few strains could be assessedfor period or phase due to poor band formation. When put into temperature

CHAPTER 3

56

cycles, however, the same strains that did not ‚band’ well showed clear banding.When we correlated the phase yielded in temperature cycles to FRP we found ahighly significant correlation. Why we find this correlation in temperature cyclesbut not when light cycles are applied can only be speculated about. Oneexplanantion could be that Neurospora crassa is under a bigger selective pressureby temperature compared to light. At least during certain parts of its life,Neurospora could be exposed to light levels too low to function as a zeitgeber(like below the light-shielding bark of a tree), which would mean that theendogenous clock is rather entrained by thermoperiods. In that case, the endoge-nous clock would be reflected rather by the temperature-entrained state andphase – entrained by a temperature cycle - could be found to correlate withperiod. Contrary to what we found with phase in LD 2/10, the entrained phaseusing a temperature cycle is not correlated with latitude. This finding can bepartly explained by the fact that –for light cycles- a latitude-dependence makesmuch more sense than for temperature cycles, as photoperiods in nature aremuch more predictable than thermoperiods which get less predictable, the higherthe latitude. But to clear up all these – partly seasonal – questions it needsfurther research.

When, for example, birds are studied in quasi-natural experiments, the timingof activity correlates with variations in photoperiod and twilight duration (Daan& Aschoff, 1975). In birds and mammals at the arctic circle an increased vari-ablity of phase (measured as ψmidpoint) with longer twilights (Daan & Aschoff,1975) was observed. So, taking the experiments with Neurospora crassa a stepfurther we determined the phases of a subsample of eight strains from differentlatitudes ranging from 1.78º to 55.95º in light cycles (12/12 LD, 25ºC) withdifferent twilight lengths ranging from 0 up to 2.5 hours. The strains from lowerlatitudes showed almost the same phase in all different protocols, independent oftwilight length, whereas the strains from latitudes greater than 40º showed aphase advance (twards twilight-on) when twilight was longer than 1.5 hours.This finding is not surprising, since at the low latitudes shorter or almost notwilight is observed in nature. Hence, there is no selective advantage for strainsthere to detect twilight. In the strains from higher latitudes, however, apparentlyeven very dim light at the beginning of the twilight can be detected and taken asphase relation point for entrainment. We did not find an increased variability(measured as standard deviation of onset of conidiation in 6 replicates per strain)with twilight. However, a subsample of 8 strains might be too few, and futureefforts are recommended to concentrate on wild type strains in more naturallight or temperature environments or, even better, in a combination of both.

For all experiments presented here, it has to be taken into account thatNeurospora crassa from different geographic origin will be also genetically

LATITUDINAL CLINES

57

diverse. NcA, NcB and NcC are the three major phylogenetic clades in which thegenus Neurospora crassa can be subclassified. All European strains isolated so farfall into the clade NcB (Jacobson et al, 2006), together with strains from westernNorth America and equatorial Africa. NcC exists in southern India exclusively,together with NcA, that is widespread across the Caribbean Basin and Africa.Furthermore Neurospora from specific regions might have a distinct set ofmarkers (e.g. the strains from Seville, Spain, have a particular single nucleotidepolymorphism). Therefore, it is seemingly not justified to compare betweenclades but only within clades and to not include genetically distinct isolates instudies comparing world wide collections of strains. This was partly considered,when European and non-European strains were compared. But at the currentpoint of time our knowledge of all these factors is limited. So, future studiesconcentrating on chronoecology should shed more light on genetic factors influ-encing diversity in Neurospora crassa.

Conclusions and outlook

The phenotyping of Neurospora crassa isolates from different latitudes showedhighly significant correlations for phase (in a light-dark-cycle) and latitude, andphase (in a temperature cycle) and period. Further, entrainment according totwilight length and latitude-of-origin may show a systematic relationship. Toconfirm and extend these correlations with a different clock readout, like e.g.clock gene expression, will be the task of future studies. Selected strains could beassayed in light- and temperature cycles delivered simultaneously. Informationon entraining in twilight conditions versus square-wave light cycles is also a highpriority. Finally, to advance our knowledge about Neurospora crassa, its circadianclock and its zeitgebers, it would be informative to elaborate genetic markersusing linkage analysis.

References

CHAPTER 3

58

Belden WJ, Larrondo LF, Froehlich AC, Shi M, Chen CH, Loros JJ, Dunlap JC (2007) Theband mutation in Neurospora crassa is a dominant allele of ras-1 implicating RASsignaling in circadian output. Genes Dev 21(12): 1494-1505

Boulos Z, Macchi M, Terman M (1996) Effects of twilights on circadian entrainment patternsand reentrainment rates in squirrel monkeys. J Comp Physiol [A] 179(5): 687-694

Daan S, Aschoff J (1975) Circadian Rhythms of locomotor activity in captive birds andmammals: their variations with season and latitude. Oecologia 18: 269-316

Duffy JF, Dijk DJ, Hall EF, Czeisler CA (1999) Relationship of endogenous circadian mela-tonin and temperature rhythms to self-reported preference for morning or eveningactivity in young and older people. J Investig Med 47(3): 141-150

LATITUDINAL CLINES

59

Duffy JF, Rimmer DW, Czeisler CA (2001) Association of intrinsic circadian period withmorningness-eveningness, usual wake time, and circadian phase. Behav Neurosci115(4): 895-899

Everts LG, Strijkstra AM, Hut RA, Hoffmann IE, Millesi E (2004) Seasonal variation in dailyactivity patterns of free-ranging European ground squirrels (Spermophilus citellus).Chronobiol Int 21(1): 57-71

Hotta CT, Gardner MJ, Hubbard KE, Baek SJ, Dalchau N, Suhita D, Dodd AN, Webb AA(2007) Modulation of environmental responses of plants by circadian clocks. Plant CellEnviron 30(3): 333-349

Jacobson DJ, Dettman JR, Adams RI, Boesl C, Sultana S, Roenneberg T, Merrow M, DuarteM, Marques I, Ushakova A, Carneiro P, Videira A, Navarro-Sampedro L, Olmedo M,Corrochano LM, Taylor JW (2006) New findings of Neurospora in Europe and compar-isons of diversity in temperate climates on continental scales. Mycologia 98(4): 550-559

Jacobson DJ, Powell AJ, Dettman JR, Saenz GS, Barton MM, Hiltz MD, Dvorachek WH,Glass NL, Taylor JW, Natvig DO (2004) Neurospora in temperate forests of westernNorth America. Mycologia 96: 66-74

Jones CR, Campbell SS, Zone SE, Cooper F, DeSano A, Murphy PJ, Jones B, Czajkowski L,Ptacek LJ (1999) Familial advanced sleep-phase syndrome: A short-period circadianrhythm variant in humans. Nat Med 5(9): 1062-1065

Michael TP, Salome PA, Yu HJ, Spencer TR, Sharp EL, McPeek MA, Alonso JM, Ecker JR,McClung CR (2003) Enhanced fitness conferred by naturally occurring variation in thecircadian clock. Science 302(5647): 1049-1053

Morgan W, Feldman JF (1998) Preliminary investigation of the circadian rhythms of wild-collected Neurospora strains. FGN 45: 30-31

Pandit A, Maheshwari R (1994) Sexual reproduction by Neurospora in nature. FGN 41: 67-68

Pittendrigh CS, Bruce VG, Rosensweig NS, Rubin ML (1959) Growth patterns in Neurosporacrassa. Nature 184: 169-170

Pittendrigh CS, Takamura T (1987) Temperature dependence and evolutionary adjustmentof critical night length in insect photoperiodism. Proc Natl Acad Sci USA 84(20): 7169-7173

Pittendrigh CS, Takamura T (1989) Latitudinal clines in the properties of a circadian pace-maker. J Biol Rhythms 4: 217-235

Rémi J (2007) The circadian surface of Neurospora crassa- From physiology to molecularmechanisms Doctoral Thesis University of Munich, LMU

Roenneberg T, Merrow M (2001) Seasonality and photoperiodism in fungi. J Biol Rhythms16(4): 403-414

Roenneberg T, Taylor W (2000) Automated recordings of bioluminescence with special refer-ence to the analysis of circadian rhythms. Methods Enzymol 305: 104-119

Sargent ML, Kaltenborn SH (1972) Effects of medium composition and corbon dioxide oncircadian conidiation in Neurospora. Plant Physiol 50: 171-175

Sies H, Stahl W (2004) Carotenoids and UV protection. Photochem Photobiol Sci 3(8): 749-752

Tan Y, Merrow M, Roenneberg T (2004) Photoperiodism in Neurospora crassa. J BiolRhythms 19(2): 135-143

van Oort BE, Tyler NJ, Gerkema MP, Folkow L, Blix AS, Stokkan KA (2005) Circadian organi-zation in reindeer. Nature 438(7071): 1095-1096

Vogel HJ (1956) A convenient growth medium for Neurospora (Medium N). Microbiol GenetBull 13: 42-43

Young MW, Kay SA (2001) Time zones: a comparative genetics of circadian clocks. NatureReviews Genetics 2: 702-715

Using entrainment to discover clock genes: A QTL-Analysis in Neurospora crassa

C. Madeti, A. Marchetti, E.Turner, J. Taylor,T. Roenneberg and M. Merrow

Manuscript

4CHAPTER

ABSTRACT

One of the central aims of the field of chronobiology is to understand how organisms predictdaily changes in the environment and adjust to them. How these environmental rhythms areperceived by the organism and how biological rhythms are generated on a cellular level hasbeen studied extensively during the last decades. In all phyla investigated so far, distinctsets of clock genes have been found to generate circadian oscillations via a complex networkof positive and negative feedback loops. Furthermore, the inheritance of circadian pheno-types does not simply follow the rules of Mendelian genetics, but is under control ofmultiple genes plus environment and is therefore an example of a complex genetic trait.Mapping quantitative trait loci in the context of circadian rhythmicity has been used inArabidopsis, mice and other organisms and indicates that circadian behavior is influenced bymany more clock genes than have been identified, yet. A QT (quantitative trait) is a pheno-type, that can be quantified (e.g. flowering time, growth rate, fertility, spore size, hyphalgrowth, tolerance to toxins etc.). Given the complexity of the circadian clock and the factthat in a previous study multiple genes could be attributed to circadian phenotypes likephase of entrainment and period, these traits seem to be optimal candidates for a QTL-analysis in Neurospora crassa.In addition to an obvious circadian rhythm (conidial band formation), N. crassa’s advan-tages include a relatively small, sequenced genome, high numbers of offspring from a singlecross, and haploidy.To date, there has been only one published QTL-analysis in N. crassa.The aim of the underlying study is to find novel clock genes by associating circadian pheno-types with a linkage map for Neurospora crassa. Contrary to circadian theory and experi-mental evidence, but similar to findings in Arabidopsis, we found no straightforward correla-tions between phase and free running period (FRP). One QTL that we identified for theperiod trait was linked to the known clock gene frequency, demonstrating proof of principle..

Introduction

More than 2000 years ago Androsthenes, scribe of Alexander the Great,described how leaves of the tamarind tree were horizontally oriented during theday and folded at night (Satter et al, 1974). Daily or circadian (‚about a day’)rhythms in behaviour of virtually all eukaryotes - even unicellular algae – andalso prokaryotes have since been demonstrated (Roenneberg & Merrow, 2005).The system maintaining, predicting and adjusting to daily rhythms, termed ‚circa-dian clock’, is heritable, as was shown already by Erwin Bünning in 1932(Bünning, 1932). The first clock genes, the period (per) gene in Drosophila(Konopka & Benzer, 1971) and the Neurospora crassa gene frequency (frq)(Feldman & Hoyle, 1973), were discovered in the early 1970s. As a model forhow circadian rhythmicity is generated at the molecular level the model of a‚transcriptional-translational autoregulatory negative feedback loop’ has beenproposed (Hardin et al, 1990): the transcription of a clock gene into RNA leadsto its translation into protein. Upon reentering the nucleus the protein blocks itsown transcription until - after degradation of protein and RNA - this inhibition isreleased to restart the cycle. This model is to date still in use with new compo-nents (other clock genes or homologs of known clock genes) simply added, suchas positive feedback loops, additional coupled loops (e.g. the Frq-Less-OscillatorFLO in Neurospora, see below), and cellular regulators (like phosphatases andkinases) adding more complexity to the system (Roenneberg & Merrow, 2005).

In several studies this complexity has been shown also on the genetic level,for example in mice, where the progeny of a cross displayed a wide variety ofphases and periods despite the parents having similar phenotypes (Shimomura etal, 2001). This speaks for a non-Mendelian inheritance and the involvement ofmultiple genes. The circadian system is a complex trait.

Mutant screens in the filamentous fungus Neurospora crassa so far have usedmostly constant conditions and focused therefore on alterations in period lengthor on arrhythmic mutants. Through these screens also the first Neurospora clockgene, the above mentioned gene frequency (frq), as well as mutant allelesaccounting for abnormal period lengths or loss of rhythmicity have been discov-ered (Dunlap & Loros, 2004; Feldman & Hoyle, 1973; Loros & Dunlap, 2001b).Though the molecular properties of frq and other clock genes have been exten-sively characterized, we are still far away from a full understanding of the molec-ular clock machinery in Neurospora crassa. Furthermore, there is ample evidencethat there are additional components that contribute to oscillatory machinery.Key evidence concerning this hypothesis comes from the demonstration of circa-dian entrainment in a frq-less strain (Merrow et al, 1999a; Roenneberg et al,2005) and the demonstration of 3 oscillating transcripts in frq-less strains

CHAPTER 4

62

(Vitalini et al, 2004). This has led to the prediction of a FRQ-Less-Oscillator(FLO) in Neurospora. Clearly, in a wild type background, the FLO is integratedinto the system. Without critical clock genes, the FLO is weak and difficult tocharacterize. To detect these subtle characteristics and their molecular basis weare applying a QTL-analysis on a progeny of Neurospora crassa strains subjectednot only to constant conditions but also to light-dark-cycles.

Quantitative traits (QTs) are per definitionem phenotypes influenced bymultiple loci contributing to the natural variation of these traits. The complexityof the molecular mechanisms underlying circadian rhythmicity makes the circa-dian clock traits period and phase of entrainment ideal targets for a QTL-study.Quantitative Trait (QT) studies have revealed interacting alleles in the context ofcircadian rhythmicity and have furthermore helped to discover novel clockcomponents e.g. in Arabidopsis (Darrah et al, 2006; Edwards et al, 2006;Edwards et al, 2005; Michael et al, 2003; Swarup et al, 1999), the pitcher-plantmosquito Wyeomyia smithii (Mathias et al, 2007) and in mice (Shimomura et al,2001; Suzuki et al, 2001). A recent QTL-study on Neurospora crassa (Kim et al,2007) has found three potentially novel QTLs affecting period and eight QTLsaffecting phase of the oscillation in a free run.

Neurospora crassa is an ideal candidate for a QTL-analysis: its relatively smalland fully sequenced genome (Galagan et al, 2003) provides the basis fordesigning markers and finding clock QT loci. The haploid genome makes reversegenetics easier, as no backcrosses are needed, genetic dominance does not influ-ence the phenotypic variation. Furthermore, Neurospora crassa has a relativelyshort generation time (a cross is available within weeks) and a large progeny.The additional advantage for circadian researchers is the easily monitored circa-dian rhythm in Neurospora crassa: its conidiation rhythm can be assessed in glasstubes filled with solid growth medium (so-called race tubes), where stretches ofspore formation (visible as ‚bands’) alternate with less dense stretches of hyphalgrowth (‚interbands’). These can be visualized as areas of high pixel density andinterbands as areas of low density.

We measured free-running period (FRP) in constant darkness and the phaseof entrainment (Φon, ‚phi on’) in light-dark-cycles for each member of a QTLmapping population comprising up to 500 offspring of a cross between two wildtype strains of N. crassa. Additionally we did both experiments at two distincttemperatures to assess temperature compensation. Traits conferring fitness inNeurospora crassa (e.g. spore size, hyphal growth, tolerance to toxins etc.) werestudied using a QTL-analysis in order to unravel novel loci affecting these traits.Since the natural variation in clock parameters like period, phase, temperaturecompensation etc. is neccessary for the fitness of an organism, these traits havebeen assessed and analysed in the same QTL-study. The correlation of pheno-

QTL

63

types with the pattern of 103 AFLP-markers allows mapping of the underlyinggenes through their cosegregation with these markers.

The results of this QTL-analysis offer proof of principle: cosegregating traitsand likely genes were FRP and frq, temperature compensation (see details below)and prd-4 and os-1.

Material and methods

Strains and MediaThe strains utilized are the offspring of a cross between two wildtype strains,1202 (FGSC-No. 8866, from Tamil Nadu, India) and 74OR23-IVA (FGSC-No.2489, from the Caribbean Basin, St. Lawrence). Up to 500 of these have beenanalyzed in race tube experiments and were mapped (Turner and Taylor, unpub-lished data) by using 103 AFLP-markers (Vos et al, 1995) that are polymorphicbetween the two parental strains.

Each of the strains was assessed in duplicates. Conditions included:● constant darkness (DD) at 25°C ● constant darkness (DD) at 22°C● a cycle with 12 hours of light and 12 hours of darkness at 25°C● a cycle with 12 hours of light and 12 hours of darkness at 20°C

Vogel’s medium N (Vogel HJ, 1956) (1X Vogel’s solution, 0.5% Arginine, 10µl/100ml Biotin, 2% Agar) without glucose (Munich Minimal Medium) wasused. All race tubes were marked daily.

After completing the experiment all race tubes were scanned and the fileswere converted to pict- files, which were modified to increase contrast. All datafrom these pictures was analyzed to obtain free running periods or phase ofentrainment by using Chrono 6.4m (Roenneberg & Taylor, 2000).

Defining ‚rhythmicity’In the course of the experiments, it became obvious that many of the QTL-strainscould not be assessed because they either did not show clear conidiation rhythmsor were not stably entrained. We therefore developed criteria to categorize thestrains according to the quality of their rhythmic/non-rhythmic behaviour. Fig.4.1 shows two examples of line plots from a period analysis of two QTL-strains.The strain on top, No. 142, is an example of one of the least rhythmic strains,having a low amplitude rhythm with a low r-value. The r-value is a measurementfor how well a sine-curve can be fitted to the experimental rhythm. r = 1 if thefit is perfect and is <1 as the fit deteriorates. Strain No. 23, shown below, is an

CHAPTER 4

64

example of one of the most rhythmic strains, with an r-value close to 1 and ahigh amplitude of the oscillation.

A calculation was added for ‚robustness’ in the trait period. We hypothesizedthat the most robust rhythms would (analytically) change the least when thetrend correction was altered. Thus, periodicity was assayed with 13 and 23 htrend correction, in addition to a default 24h. A Dtau (= |tau tc23- tau tc13|) of>1 was taken as cut-off for non-robust rhythms in QTL-strains. In the figuresshown below, only „robust“ strains according to our classification with a Dtau <1were included. This means that their rhythm changes its period for less than 1hour even if the period is trend corrected by 13 hours.

Another observation was that in some cases the values obtained for periodfrom two replicates of one strain were different. To quantify this difference andfind whether this variability within one strain is associated with a QT, we intro-duced the traits ‚rhythmicity’ and ‚Race Tube Standard Deviation’ (RTSD).

In the trait „rhythmicity“ we used categorical values. In some cases the bothreplicates of one strain were found to be arrhythmic, in others not. We scoredstrains using the values 1 (in case both replicates were arrhythmic), 0.5 (if one of

QTL

65

79: 142 1–20.27

22.94

amplitude = 22.55

90 180 270 360

r = 0.79

04: 23 2–150.5

255.5

amplitude = 372.56 r = 0.99

Figure 4.1 Line plots showing averaged periods of two QTL-strains, No. 142 (on top) andNo. 23 (below) over two days. Amplitudes are 22.55 (with an r-value of 0.71) for 142 and372.56 (with an r –value of 0.95). Graphs generated with Chrono 6.4 (Roenneberg & Taylor,2000).

two replicates was arrhythmic) and 0 (when both replicates were rhythmic).For the trait ‚RTSD’ a standard deviation of the values for free running period ofthe two replicates of one strain was calculated.

We also developed a scoring system for phase of entrainment (‚entrainmentscore’). For each strain, minimum, onset, maximum and offset of conidiation andr-value (see above) were calculated based on two-harmonic cosine curveanalysis. This was determined for both tubes for each strain. Then we measuredthe standard deviation (StDev in Table 4.1) of the two replicates for each phasemarker. These 4 values were averaged and divided by the averaged r-value of therespective race tube pair. In this way, the robustness of the curve (represented bythe r-value) incorporates intra-strain-variability. An ‚entrainment score’ of morethan 15 (equal to degrees, roughly meaning an intra-strain-difference from 1 racetube to the second of 1 hour) was the threshold for strains to be discounted (seeTable 4.1).

Taken together, the phenotypes assessed in this study were: ● ττ (period) in constant darkness (DD) at 25°C (measured in hours, see Fig.

4.2). ● ττ in constant darkness (DD) at 22°C (measured in hours, see Fig. 4.2).● “rhythmicity”: we scored strains as categorical data with 1 (both race tubes

arrhythmic), 0.5 (one of two race tubes arrhythmic) and 0 (both race tubesrhythmic); thus rhythmicity is quantified.

CHAPTER 4

66

Φmin Φon Φmax Φoff r

Race tube 1 307.50 116.25 187.50 253.75 0.98Race tube 2 313.50 114.25 187.00 256.75 0.97Average 310.50 115.25 187.25 255.25 0.97StDev 4.24 1.41 0.35 2.12Average StDev 2.03Entrainment score 2.09(=average StDev/average r)

Table 4.1 How to measure the ‚entrainment score’: The lines Race tube 1 and Race tube 2are examples of the output from the Chrono program for two replicates of an individualstrain, Φmin, Φon, Φmax, Φoff and r-value (see description in text). For the two values perstrain and all 4 phase reference points an average was calculated. In a next step the stan-dard deviation (average StDev in Table 4.1) of each of these 4 values was determined, aver-aged and devided by the average r-value from both race tubes. An ‚entrainment score’ >15(which roughly equals 1 hour) was used as cut-off for strains that do not stably entrain.

● “RTSD” (race tube standard deviation): The trait RTSD is calculated as astandard deviation of the periods measured in two replicates per strain

● phase of entrainment (hereafter represented as Φon, onset of conidiationcompared to lights off) in a cycle with 12 hours of light and 12 hours of dark-ness at 25°C, measured in º (see Fig. 4.3)

● phase of entrainment (phase, here represented as Φon, onset of conidiationcompared to lights off) in a cycle with 12 hours of light and 12 hours of dark-ness at 20°C, measured in º (see Fig. 4.3)

● „temperature compensation“: to assess the trait temperature compensationthe absolute difference of FRPs in DD/22ºC and DD/25ºC was analysed.

Since the robustness of strains also shows a large variability and might conferfitness, we also subjected the traits ‚RTSD’ and the categorial values in the trait‚rhythmicity’ (see below for description of both), to the QTL-analysis.

QTL-AnalysisThe QTL-analysis was done via interval mapping using ‚QTL Cartographer’. Onlytest intervals (QTLs) with likelihood ratio statistics greater than 12 were includedin the output files.

Results

RESULTS OF THE PHENOTYPING IN RACE TUBE EXPERIMENTS

To phenotype the up to 500 QTL-strains, we measured free-running period (FRP)in constant darkness (DD) and the phase of entrainment (Φon) in light-dark-cycles with alternating 12 hours of light and 12 hours of darkness (LD 12/12).Both experiments were additionally done at two temperatures to assess tempera-ture compensation.

Free-running period (FRP) in DD/25ºC and DD/22ºCThe distributions of free-running periods in 25 and 22°C (Fig. 4.2) have curveswith a similar shape and the mean FRP is almost equal with 23.1 hours at 22°Cand 23 hours at 25°C. The period values spanned reach from 16 hours to 29.7hours in DD/22ºC and 20.9 hours to 28.8 hours in DD/25ºC. The parental strainsare arrythmic (both by eye and by aforementioned FRP and ‚robustness’- analysiswith Chrono 6.4 (Roenneberg & Taylor, 2000)). Only 228 of the ca. 500 strains(equal to 46%) were rhythmic at DD/22ºC and 119 of the ca. 200 strains (equalto 60%) at DD/25ºC.

QTL

67

Phase of entrainment (phase) in LD 12/12 at 25ºC and 20ºCSimilar distributions are observed for phase of entrainment (onset of conidiationcompared to lights-off) in 12/12 LD cycles (Fig. 4.3). The mean value at 25°C islater with 109° (ca. 1 hour after midnight) compared to 94° at 20°C, whichequals shortly (16 minutes) after midnight (90º after lights-off). So, the averagephases are about 1 h (15º) earlier in the colder temperature. The distribution ofphases is wider at 20ºC, running from 16º up to 140.5 º compared to a range of41º to 145.8º at 25ºC. At some level, this is thus consistent to the situation inDD, where the range of periods is wider in colder temperatures. At another level,it is not. Earlier entrainment would be expected to correlate with shorter FRP, buton average, there is no population difference.

QTL- strains kept in LD-cycles can also weak rhythmicity, as in DD: onlyabout 70% (in the LD-cycle with 25ºC) or 61% (in the LD-cycle with 20ºC) ofthe wild type progeny show entrainment, as assessed by eye and Chrono-analysis, i.e. almost 10% fewer strains entrained when subjected to coldertemperatures. In Fig. 4.3 only the strains that have an ‚entrainment score’ (seedescription in Materials and Methods section) of less/equal to 15 are included.

Correlation of period and phase in QTL-strainsIn oscillator theory, phase and period are correlated in a direct relationship,where long period oscillations entrain to a later phase and short period oscilla-tions entrain to an earlier phase. However, the relationship can theoretically beinfluenced by other factors, as well. In a QTL-study utilizing Arabidopsis (Michael

CHAPTER 4

68

0

20

40

10

50

30

num

ber

18 20 22 24 26 28 30

FRP, τ (h)

FRP in DD/22°C, n=228FRP in DD/25°C, n=119

16

60

Figure 4.2 Distribution of FRPs (Free-running periods) in hours of QTL-strains in DD/25ºCand DD/22ºC. Only strains that fulfill the robustness criterium of ∆tau <1 are included (seedescription in text). Data analysed using Chrono 6.4 (Roenneberg & Taylor, 2000)

et al, 2003) a negative correlation for phase and period was found in accessions,suggesting that in this case the QTL accounting for lengthening of the periodsegregated together with the QTL responsible for an advanced phase. A positivecorrelation was found in the same study when different clock mutants wereanalysed (although within a genotype there is again negative correlation of phase

QTL

69

0

20

40

10

50

30

num

ber

100-20 40

Φon , onset of conidiation (°)

Φon LD/25°C, n=139Φon LD/22°C, n=120

160

Figure 4.3 Distribution of phases (Φon in relation to lights-off) of QTL-strains in 12/12 LD(25ºC, n=139) and 12/12 LD (22ºC, n=120). The grey area indicates the 12 h of darkphase. Only robust strains (entrainment scores < 15, see text) are included. Data analysedusing Chrono 6.4 (Roenneberg & Taylor, 2000). Fluorescent white light with an intensity ofapproximately 4µE was used.

28

60

120

150

90

Φon

(°)

20 22 24 26FRP, τ (h)

30

Figure 4.4 Φon (onset of conidiation after lights-off, in º, y-axis) in LD 12/12, 25ºC vs. τ inDD/25ºC. No significant correlation of phase vs. period could be detected in QTL-strains (allstrains shown as squares), even if the 15. Percentile -15% extreme early and late entrainedstrains- had been omitted (shown as black dots). All strains shown had robust periods andwere well entrained according to our categories (see text for details) .

and period). In previous work with Neurospora, clock gene mutant strains alsoshowed a clear phase/period relationship. However, the data presented here failsto identify a correlation between these measurements.

RESULTS OF THE QTL-ANALYSIS(DONE BY ELIZABETH TURNER, BERKELEY)

The results of our phenotyping, periods, phases, ‘rhythmicity’, DRT and tempera-ture compensation (see Materials and Methods section for explanation) wereanalysed in a QTL-Analysis using 103 AFLP-markers. The results for the QTL-analysis on the traits period, ‘rhythmicity’ and RTSD (free running period inDD/25ºC and DD/22ºC) are shown in Table 4.2.

CHAPTER 4

70

Table 4.2 Results from the QTL-analysis for the traits τ in DD/25ºC and DD/22ºC and“rhythmicity” (see explanation in text) and RTSD (“Race Tube Standard Deviation”, seeexplanation in text). R2*= percentage of variance explained by QTL conditioned on thebackground markers.

DD/25°C DD/22°C

Gene Chromo- R2* Gene Chromo- R2*some of QTL some of QTL

ττ (all strains) frq 7 12 DA25 4 11DA25 4 14DA25 4 16DA25 4 14

ττ (rhythmic strains only) frq 7 16

“Rhythmicity” ccg-1, chs-2 8 27 Tel-IVL, trp-4, fsr-62 3 12ccg-1, chs-3 8 21hsp80, inl 8 22hsp80, inl, 8 16fsr029, ro-4

RTSD (∆RTSD<2) ccg-1, chs-2 8 30 DA25 4 88ccg-1, chs-2 8 30hsp80, inl 8 22hsp80, inl, 8 16fsr029, ro-4

RTSD (all) ccg-1, chs-2 8 24 Tel-IVL, trp-4, fsr-62 3 9ccg-1, chs-2 8 19 DA25 4 20hsp80, inl 8 18 ccg-1, chs-2 8 11hsp80, inl, 8 12 ccg-1, chs-2 8 9fsr029, ro-4

hsp80, inl 8 7

The QTL found for the trait FRP in DD/25ºC is associated with a markerlikely linked to the frq-gene (highlighted in rey n Table 4.2). In DD/22ºCdifferent genes segregated with the trait ‘FRP’. This region, ‘DA-25’, is a ca.1000000 bp stretch on Chromosome 4 comprising several genes, none of whichare to date known as clock or clock-related genes.

The categorical trait ‘rhythmicity’ yielded hits on Chromosomes 8 (DD/25ºC)and 3 (DD/22ºC), linked to-among others- the genes ccg-1, hsp-80 (DD/25ºC)and trp-4 (DD/22ºC). The same genes are likely to be a part of the QTLs segre-gating with the trait DRT in DD/25ºC and DRT in DD/22ºC (only when allstrains are included).

In the analysis of the trait phase the genes/stretches of DNA likely to be partsof a QTL are DA25 and hsp80 in LD 12/12 at 20ºC. In LD/25ºC, DA25 again isco-segregates with the trait (Table 4.3).

Temperature compensation (measured as absolute difference Tau DD/25ºC–Tau DD/22ºC) was also analyzed in the underlying QTL-analysis. The genesfound close to the markers showing a significant hit were prd-4, os-1 and againDA-25 on chromosome 4.

Discussion

The parental strains used in the QTL-experiments were both arhythmic and theyhave a different phase. However, in our analysis we found a wide distributionamongst the progeny for period in DD/25 ºC and DD/22 ºC and for phase in LD12/12 at 20ºC and 25ºC, consistent with the involvement of multiple genes inthese traits. This is reminiscent of what has been found in mice (Shimomura etal, 2001) and Arabidopsis (Michael et al, 2003), where the FRPs and phases ofthe offspring show a wide distribution compared to the rather similar values ofthe parental strains. These data collectively confirm that FRP and phase arecomplex phenotypic traits.

QTL

71

Table 4.3 Results from the QTL-analysis for the trait phase (onset of conidiation in relationto lights-off). R2*= percentage of variance explained by QTL conditioned on the back-ground markers.

Gene Chromosome of QTL R2*

Φon in LD 12/12 25°C DA25 4 11DA25 4 11

Φon in LD 12/12 2o°C hsp80, ..? 8 11

Concerning the distribution for phase of entrainment (which roughly equals“chronotype” in humans) in 12/12 LD cycles, similar curves are observed for dataderived at both temperatures. Similar curve shapes are also observed for periodat 25ºC and 22ºC. This demonstrates temperature compensation of both traits,FRP and phase of entrainment. However, the curves of the colder experimentalset-ups, for FRP (at 22ºC) and phase (at 20ºC), are broader, spanning a widerarray of FRPs and phases.

According to circadian theory, phase of entrainment and period are correlatedin a direct relationship, where a long period corresponds to a later phase and ashort period results in an earlier phase. This has been shown e.g. for the sleep-wake-rhythm in humans by Wright et al. (Wright et al, 2005), and is examplifiedby the Familial advanced sleep-phase syndrom (FASPS), where a short-periodrhythm in humans leads to advanced sleep. Affected individuals wake up at 4o’clock or earlier (Jones et al, 1999a). The phase of entrainment is not only influ-enced by the free running period but also by the strength of the zeitgeber (forexample the amplitude of light-intensity differences in the day vs. in the night)and also the strength of the oscillator. The effect however, depends on the indi-vidual t, where a clock with a τ < 24 hours, like in Neurospora, moves forwardand a clock with a τ >24 hours, like in humans, moves to a later phase. This rela-tion also accounts for distributions of human chronotypes becoming broader withweaker zeitgeber strength, e.g. in the low-light environment of urban white-collar workers compared to farmers in rural areas (Roenneberg et al, 2003). Inour results presented here a wider distribution of periods (Fig. 4.2) and alsophases (Fig. 4.3) can be observed in the colder experimental setup (at 22 or

CHAPTER 4

72

met-8

III

cum

r(Sk-2

)

T(NM18

3)

acr-7

T(OY33

9)

T(AR17

7)

sc scr-1

pro-1

ff-s com

col-1

6ad

-4his

-7ch

i-2ad

-2trp

-1os

-8ac

r-2ac

e-2lcu

-1

Sk-2, Sk-3

( ) ( ) ( )

dow

ro-2

T(D30

5)L

vel

tyr-1

T(UK8-1

8)R

hct-7

T(AR17

)L

ro-11

nn-17

erg-3

Tcl-IIIR

phe-2

acu-7

nic-7

T(DBL9)

( )( ) ( )

T(OY320)T(OY329)

Figure 4.5 The location of DA25 on linkage group III (taken from (Perkins, 2000)). DA25 isa ca. 1000000 bp stretch of DNA located between the markers acr-2 and os-8 (see arrows).

20ºC) compared to the distributions at 25ºC. Since there is no zeitgeber in DDand since the zeitgeber in the LD-cycles is the same, this has to be a meretemperature effect. How an elevated temperature narrows down a distribution ofphenotypes is a very interesting phenomenon and could be the aim of furtherinvestigation. Interestingly, for some strains light apparently does not overcomethe effects of a 5ºC colder environment, as around 14% less strains entrained in12/12 LD at 20ºC compared to 12/12 LD at 25ºC.

In the following the results of the QTL-analysis wil be discussed: A gene likely to be part of the period trait in DD/25ºC is frq (frequency). frq isone of the so-called canonical clock genes and its mutant alleles affect periodlength. frq has also been found as an QTL-gene in the study of Kim et al. (Kim etal., 2007). However, frq was not associated with FRP at the colder temperature.Rather, the stretch of DNA associated with FRP in DD/22ºC, DA25, does notcomprise any gene that has been associated with circadian rhythmicity, so far(see Fig. 4.5). It is strongly associated with the trait RTSD in DD/22ºC (here itaccounts for 88% of the variance!) and also in phase in LD/25ºC, making it aninteresting candidate for further genetic studies.

The involvement of clock-controlled gene 1 (ccg-1), which codes for an early-morning specific transcript (Loros et al, 1989), in the traits ‚rhythmicity’ andRTSD could be a proof of principle for the genetic basis of robust rhythmicity, buthas to be assessed further. Ccg-1 is also known as glucose-repressed gene 1 (grg-1)for its preferential expression under low glucose conditions, one of the stringentconditions for a gene to be designated as a ‚clock-controlled gene’ (ccg). With ournovel race tube conditions (0% glucose in our race tube medium) we may beuncovering a novel feedback on the clock system. hsp-80 (coding for an 80kDheat-shock protein) cosegregates with the traits phase in LD/20ºC and RTSD. Itcould be another example of a ccg –like other ccgs, that have additional func-tions in the cell metabolism e.g. in stress responses, cell metabolism, and others(Correa et al, 2003). Interestingly, heat-shock factor 1 (hsf-1) has recently beenimplicated in the mammalian circadian clock (Reinke et al, 2008). This is aninteresting connection between the clock and environmental signals that areespecially relevant for Neurospora.

Apart from DA25, that is found associated with FRP in DD/22ºC and phase,different QTLs are associated with the traits period and phase. This might suggest-as in Arabidopsis, where the alleles studied in a QTL-analysis affect either periodor phase- that period and phase are not under the same genetic control and thusbegins to explain why these traits are likely not correlated (Michael et al., 2003).Another example of the same phenomenon can be seen in chapter 3 of thisthesis, where phase and period don’t correlate in some cases and do in others,depending on the zeitgeber cycles applied.

QTL

73

It is worthwhile noting, that in most cases, where only the rhythmic strainshave been analyzed or where the ones with a high SD have been omitted, the R2

increases in comparison to the case where all strains are included (see Tab. 4.1and 2)! By increasing the ‘quality’ of the phenotyping by omitting non-robuststrains the outcome of the QTL-analysis can apparently be improved drastically.The improvement in phenotype more than offsets the effect of a diminishing ‘N’.

The genes linked to the trait temperature compensation, yielded interestingcandidates in the genes prd-4, os-1 and DA-25. os-1 (osmotic-1) encodes a histi-dine kinase osmosensor, and regulates sensitivity to high osmotic pressure(Schumacher et al. 1997). The involvement of os-1 in circadian rhythmicity andthe clock controlled OS-pathway –perhaps preparing a cell for daily challenges ofosmotic stress - has been shown recently (Vitalini et al, 2007). prd-4 shows ashort period in k.o.-mutants and is required for temperature compensation, there-fore offering proof of principle, that this QTL-analysis yields reasonable resultsand should be pursued further with more markers and a higher offspringnumber. Interestingly, prd-4 was also identified as a clock QTL in the analysis byKim et al. (Kim et al., 2007). The future aim should be to narrow down the QTLregions found, especially DA25, to be able to find new clock genes. Furthermore,additional strains should be analyzed to increase the statistical significance andtherefore the power of the QTL-analysis.

References

CHAPTER 4

74

Bünning E (1932) Über die Erblichkeit der Tagesperiodizität bei den Phaseolus Blättern. Jbwiss Bot 81: 411-418

Correa A, Lewis ZA, Greene AV, March IJ, Gomer RH, Bell-Pedersen D (2003) Multiple oscil-lators regulate circadian gene expression in Neurospora. Proc Natl Acad Sci USA100(23): 13597-13602

Darrah C, Taylor BL, Edwards KD, Brown PE, Hall A, McWatters HG (2006) Analysis ofphase of LUCIFERASE expression reveals novel circadian quantitative trait loci inArabidopsis. Plant Physiol 140(4): 1464-1474

Dunlap JC, Loros JJ (2004) The Neurospora circadian system. J Biol Rhythms 19(5): 414-424Edwards KD, Anderson PE, Hall A, Salathia NS, Locke JC, Lynn JR, Straume M, Smith JQ,

Millar AJ (2006) FLOWERING LOCUS C mediates natural variation in the high-tempera-ture response of the Arabidopsis circadian clock. Plant Cell 18(3): 639-650

Edwards KD, Lynn JR, Gyula P, Nagy F, Millar AJ (2005) Natural allelic variation in thetemperature-compensation mechanisms of the Arabidopsis thaliana circadian clock.Genetics 170(1): 387-400

Feldman JF, Hoyle MN (1973) Isolation of circadian clock mutants of Neurospora crassa.Genetics 75: 605-613

Galagan JE, Calvo SE, Borkovich KA, Selker EU, Read ND, Jaffe D, FitzHugh W, Ma LJ,Smirnov S, Purcell S, Rehman B, Elkins T, Engels R, Wang S, Nielsen CB, Butler J,Endrizzi M, Qui D, Ianakiev P, Bell-Pedersen D, Nelson MA, Werner-Washburne M,Selitrennikoff CP, Kinsey JA, Braun EL, Zelter A, Schulte U, Kothe GO, Jedd G, Mewes W,

QTL

75

Staben C, Marcotte E, Greenberg D, Roy A, Foley K, Naylor J, Stange-Thomann N,Barrett R, Gnerre S, Kamal M, Kamvysselis M, Mauceli E, Bielke C, Rudd S, Frishman D,Krystofova S, Rasmussen C, Metzenberg RL, Perkins DD, Kroken S, Cogoni C, Macino G,Catcheside D, Li W, Pratt RJ, Osmani SA, DeSouza CP, Glass L, Orbach MJ, Berglund JA,Voelker R, Yarden O, Plamann M, Seiler S, Dunlap J, Radford A, Aramayo R, Natvig DO,Alex LA, Mannhaupt G, Ebbole DJ, Freitag M, Paulsen I, Sachs MS, Lander ES, NusbaumC, Birren B (2003) The genome sequence of the filamentous fungus Neurospora crassa.Nature 422(6934): 859-868

Hardin PE, Hall JC, Rosbash M (1990) Feedback of the Drosophila period gene product oncircadian cycling of its messenger RNA levels. Nature 343: 536-540

Jones CR, Campbell SS, Zone SE, Cooper F, DeSano A, Murphy PJ, Jones B, Czajkowski L,Ptacek LJ (1999) Familial advanced sleep-phase syndrome: A short-period circadianrhythm variant in humans. Nat-Med 5(9): 1062-1065

Kim TS, Logsdon BA, Park S, Mezey JG, Lee K (2007) Quantitative trait loci for the circa-dian clock in Neurospora crassa. Genetics 177(4): 2335-2347

Konopka R, Benzer S (1971) Clock mutants of Drosophila melanogaster. Proc Natl Acad SciUSA 68: 2112-2116

Loros JJ, Denome SA, Dunlap JC (1989) Molecular cloning of genes under control of thecircadian clock in Neurospora. Science 243(4889): 385-388

Loros JJ, Dunlap JC (2001) Genetic and molecular analysis of circadian rhythms inNeurospora. Annu Rev Physiol 63: 757-794

Mathias D, Jacky L, Bradshaw WE, Holzapfel CM (2007) Quantitative trait loci associatedwith photoperiodic response and stage of diapause in the pitcher-plant mosquito,Wyeomyia smithii. Genetics 176(1): 391-402

Merrow M, Brunner M, Roenneberg T (1999) Assignment of circadian function for theNeurospora clock gene frequency. Nature 399(6736): 584-586

Michael TP, Salome PA, Yu HJ, Spencer TR, Sharp EL, McPeek MA, Alonso JM, Ecker JR,McClung CR (2003) Enhanced fitness conferred by naturally occurring variation in thecircadian clock. Science 302(5647): 1049-1053

Perkins DD (2000) Neurospora crassa genetic maps and mapped loci. FGN 47: 40-58Reinke H, Saini C, Fleury-Olela F, Dibner C, Benjamin IJ, Schibler U (2008) Differential

display of DNA-binding proteins reveals heat-shock factor 1 as a circadian transcriptionfactor. Genes Dev 22(3): 331-345

Roenneberg T, Daan S, Merrow M (2003) The art of entrainment. J Biol Rhythms 18(3):183-194

Roenneberg T, Dragovic Z, Merrow M (2005) Demasking biological oscillators: propertiesand principles of entrainment exemplified by the Neurospora circadian clock. Proc NatlAcad Sci USA 102(21): 7742-7747

Roenneberg T, Merrow M (2005) Circadian clocks - the fall and rise of physiology. Nat RevMol Cell Biol 6(12): 965-971

Roenneberg T, Taylor W (2000) Automated recordings of bioluminescence with special refer-ence to the analysis of circadian rhythms. Methods Enzymol 305: 104-119

Satter RL, Geballe GT, Applewhite PB, Galston AW (1974) Potassium flux and leaf move-ment in Samanea saman. I. Rhythmic movement. J Gen Physiol 64(4): 413-430

Shimomura K, Low-Zeddies SS, King DP, Steeves TD, Whiteley A, Kushla J, Zemenides PD,Lin A, Vitaterna MH, Churchill GA, Takahashi JS (2001) Genome-wide epistatic interac-tion analysis reveals complex genetic determinants of circadian behavior in mice.Genome Res 11(6): 959-980

Suzuki T, Ishikawa A, Yoshimura T, Namikawa T, Abe H, Honma S, Honma K, Ebihara S(2001) Quantitative trait locus analysis of abnormal circadian period in CS mice. MammGenome 12(4): 272-277

CHAPTER 4

76

Swarup K, Alonso-Blanco C, Lynn JR, Michaels SD, Amasino RM, Koornneef M, Millar AJ(1999) Natural allelic variation identifies new genes in the Arabidopsis circadian system.Plant J 20(1): 67-77

Vitalini MW, de Paula RM, Goldsmith CS, Jones CA, Borkovich KA, Bell-Pedersen D (2007)Circadian rhythmicity mediated by temporal regulation of the activity of p38 MAPK.Proc Natl Acad Sci USA 104(46): 18223-18228

Vitalini MW, Morgan LW, March IJ, Bell-Pedersen D (2004) A genetic selection for circadianoutput pathway mutations in Neurospora crassa. Genetics 167(1): 119-129

Vogel HJ (1956) A convenient growth medium for Neurospora (Medium N). Microbiol GenetBull 13: 42-43

Vos P, Hogers R, Bleeker M, Reijans M, van de Lee T, Hornes M, Frijters A, Pot J, Peleman J,Kuiper M, et al. (1995) AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res23(21): 4407-4414

Wright KP, Jr., Gronfier C, Duffy JF, Czeisler CA (2005) Intrinsic period and light intensitydetermine the phase relationship between melatonin and sleep in humans. J BiolRhythms 20(2): 168-177

Appendix

At the end of the course of phenotyping QTL-strains we discovered that entrain-ment could –at the least visually- be improved by applying thermo periodsinstead of light-cycles or constant darkness (see examples for race tubes in Fig.4.6). Eventually the number of strains that can go into the QTL-analysis –simplybecause their rhythm is apparently more robust/visuable- could be increased byapplying temperature cycles, thereby resulting in a higher statistical exactness.From our experience with wild type strains it seems, furthermore, that tempera-ture cycles might help to overcome the robustness problems (see Defining ‘rhyth-micity’ section).

QTL

77

DD25°C

Wt 8866

bdA

WT 8859

2/10LD25°C

Wt 8866

bdA

WT 8859

12/12h27/22°C

Wt 8866

bdA

WT 8859

Figure 4.6 Examples for effects of the experiment on the growth of strains on racetubes.Three conditions are shown from top to bottom, DD/25ºC, a light-dark-cycle and a tempera-ture cycle with alternating 12 hours of cold (22ºC) and warm (27ºC) temperatures incomplete darkness. The strains shown (two race tubes, each) are the wildtype 1202 (FGSC-No. 8866) (one of the parental strains for the QTL-study), the lab standard strain bdA andthe wild type FGSC-No. 8859. The medium used is Vogel’s Medium N (Vogel, 1956) withoutglucose.

Entrainment reveals the photoreceptor genecryptochrome as a clock gene in Neurospora crassa

C. Madeti, T. Roenneberg and M. Merrow

Manuscript

5CHAPTER

ABSTRACT

Neurospora has been used as a model organism for studying blue light responses. Carotenoidproduction, asexual spore induction, phototropism and synchronization of the circadianclock are some important blue light-regulated physiologies of the filamentous fungusNeurospora crassa. The many candidates for photoreceptor molecules in Neurospora indicatecomplexity of light signalling in this simple organism. WHITE-COLLAR-1 (WC-1) a tran-scription factor and VIVID (VVD), a flavoprotein, are well-described blue-light sensingcomponents of Neurospora crassa. The sequence of the Neurospora genome containsphytochrome and cryptochrome homologs. The functional analyses of the former have beendescribed recently. The Neurospora CRYPTOCHROME (nCRY) has not yet been charac-terised, although its sequence-similarity to the animal cryptochromes suggests that it mightbe involved in the circadian photobiology of N. crassa. To test this hypothesis we made anncry-knock-out-mutant and subjected it to a series of circadian protocols. The results of ourphenotyping suggest that ncry is a clock gene. The mutant phenotype is accentuated inentrainment experiments using long photoperiods, therefore we propose non-standardentrainment protocols as valuable tools for the characterization of clock mutants inNeurospora crassa..

Introduction

Over 200 years ago Darwin demonstrated that plants were capable of sensingblue light (Cashmore et al, 1999). Obviously, this sensory function is mediated bya blue light photoreceptor. One such example of these molecules is known ascryptochrome. Across phyla, these proteins have pleiotropic functions rangingfrom regulation of blue-light-dependent development in plants and blue-light-mediated phase shifting of the circadian clock in insects to acting as a core circa-dian clock component in mammals (Van Gelder, 2002; Van Gelder et al, 2002;Van Gelder & Sancar, 2003).

Cryptochromes show great similarities to and probably evolved from DNAphotolyases (Ahmad & Cashmore, 1993), ca. 55-65 kDa enzymes that repair DNAdamage caused by exposure to UV-B light (Sancar, 2003). Cryptochromes,however, lack the capability of repairing DNA and apparently perform other func-tions (Cashmore et al, 1999). To date, we know most about the cryptochromes inDrosophila, mice and Arabidopsis. In Arabidopsis, CRY1 and CRY2 are predomi-nantly nuclear proteins, and they are involved in photoentrainment, regulation ofgene expression and also in the control of developmental changes (like floralinitiation, (Ahmad & Cashmore, 1993; Guo et al, 1998). In Drosophila, CRY islocated primarily in the nucleus, with small amounts found in the cytosol. dCRYinteracts with another protein, TIMELESS (dTIM), in a light-dependent mannerto promote degradation and suppress the formation of the dPERIOD (PER)-dTIM-protein complex. The inhibitory action (down-regulation of the activatorcomplex, dCLOCK and dCYCLE) of dPER-dTIM is thus suspended. It is proposedthat this results in resetting the phase of the circadian oscillation, leading toentrainment (Ceriani et al, 1999; Darlington et al, 1998). It has been shownfurther that dCRY has a non-photoreceptive role in peripheral oscillators ofDrosophila (e.g. olfactory responses in the antennae, (Krishnan et al, 2001)), aswell. In mammals, cryptochromes are also distributed between the nucleus andcytosol, and function there as components of the circadian clock. In mutantscreens, it has been shown that mice lacking either the CRY 1 or CRY 2 proteinsshowed abnormal free-running periods (a shortened FRP in the first case and alengthened FRP in the latter case (van der Horst et al, 1999). Mutants lackingboth CRY1 and CRY2 are arrhythmic (e.g. in behaviour and body temperature,(Nagashima et al, 2005), suggesting that both proteins are involved in main-taining circadian rhythmicity (Kavakli & Sancar, 2002; van der Horst et al,1999). When kept in light-dark-cycles, however, the cry1-/-,cry2-/- mutant isrhythmic, probably due to masking. It has been suggested that cryptochromesand classical opsins might have a function in signal transduction downstream ofa non-opsin photopigment, as well (Kavakli & Sancar, 2002). Sequence compar-

CHAPTER 5

80

ison of plant and animal cryptochrome families suggests a separate evolution ofthese, with animal cryptochromes being of rather recent origin, whereas thecryptochromes of plants have an ancient evolutionary background (Cashmore etal., 1999).

The Neurospora crassa genome sequence revealed a cry and photolyasehomolog (Galagan et al, 2003). Prior to the publication of the N. crassa genome,the Neurospora DNA photolyase was cloned and functionally described (Shimuraet al, 1999); this gene is distinct from the cry homolog. ncry belongs to the newlydescribed cry-DASH family (Daiyasu et al, 2004), although nCRY is in somerespect unique, because it has a C-terminal extension with no sequence similarityto any other described protein. Due to its close phylogenetic relationship to theanimal cryptochromes, it was suggested that ncry could be involved in circadianphotobiology (Daiyasu et al, 2004). However, to date, ncry has no known func-tion, and first experiments suggested that it has no function in circadian timingin Neurospora (J. Dunlap, personal communication).

In Neurospora two photoreceptors are known to be involved in the circadiansystem. WC-1 and VVD are blue light receptors with absorption maxima around465nm (Froehlich et al, 2002; Schwerdtfeger & Linden, 2003). WC-1 actstogether with WC-2 to activate the transcription of frequency (frq) and a host ofother genes in a light-dependent manner. WC-1 is regulated via rhythmic phos-phorylation by frq (Schafmeier et al, 2005) and other molecules includingPROTEIN KINASE C (nPKC) (Franchi et al, 2005). The cytoplasmic blue lightphotoreceptor and flavoprotein VIVID (VVD) (Schwerdtfeger & Linden, 2003)regulates entrainment via a mechanism termed photo-adaptation (Elvin et al,2005; Heintzen et al, 2001). Interestingly, the mutant has a normal free runningperiod in several conditions (Shrode et al, 2001), but abnormal phase of entrain-ment. Other photoreceptor candidates can be found via sequence similarity in theNeurospora genome: one of them is NOVEL OPSIN 1 (NOP-1), a homolog ofarchaeal opsins. However, ∆nop-1-strains displayed no apparent defects in lightsignaling (Bieszke et al, 1999a). Another photoreceptor candidate – as mentionedbefore - is nCRY. Its roles and involvement in circadian photoreception are to dateunknown. It has been speculated, however, that the Neurospora blue lightphotoreceptor protein nCRY is photoregulated by the WC-Complex (Froehlich etal, 2005). The genome sequence also reveals two genes that may be involved inred light photoreception, the phytochromes (phy1 and phy2). Knockouts of phy-1and phy-2 do not have an effect on any -so far- known photoresponses nor ongrowth rate, but phy1mRNA levels are under clock control (Froehlich et al,2005). According to sequence homology to plant phytochromes, the twophytochrome genes phy-1 and phy-2 are possibly encoding for Red/Far Redphotoreceptors. Furthermore, a homolog of the Aspergillus nidulans gene velvet

CRYPTOCHROME

81

can be found in the Neurospora genome sequence. In Aspergillus, this gene isinvolved in the signal transduction of both red and blue light (Borkovich et al,2004; Yager et al, 1998). The presence of a high number of photoreceptors in arather simple organism is surprising and indicates that photobiology as a wholeand circadian photoreception as a part of this might be more complex thanexpected in Neurospora.

This filamentous fungus has been a model organism in the field of circadianrhythms research since the late 1950’s and has features that make it remain anattractive eukaryotic species for this work. Its small and fully sequenced genome(Galagan et al, 2003) allows for a wide array of established molecular tech-niques. Neurospora is predominantly haploid and genetic studies are less compli-cated than in diploid organisms. When grown on a flat agar medium (mostly inhollow glass tubes called ‘race tubes’) Neurospora crassa alternately forms aconspicuous pattern of dense bands, where asexual spore (conidia) formationoccurs, interspersed with regions, where only vegetative hyphae are formed. Thefree running period of band formation is about 22h in constant darkness andexactly once per 24h in a 24h light/dark (LD) cycle.

We used a motif search in the Neurospora CRY sequence to find interestingfeatures in the context of the circadian clock, i.e. features or molecules that havebeen shown to be connected to circadian rhythmicity in other organisms. Toassay for involvement of the Neurospora crassa cry in circadian photoreceptionand entrainment, we generated a knock-out mutant to characterize phenotypiceffects of a lack of ncry on the circadian clock. The physiological output meas-ured as rhythmic formation of asexual spores (conidiation) in this mutant wasstudied in protocols applying various light/dark-cycles with white and blue light(465nm).

Material and methods

Motif search in the amino acid sequence of theNeurospora cryptochrome geneThe amino acid sequence of the Neurospora crassa CRY protein (NCU00582.3,available at the Broad Institute web page,http://www.broad.mit.edu/annotation/genome/ neurospora/Home.html)was scanned for motifs in the PROSITE data base(http://expasy.org/prosite/)and in ’Swiss-Prot’ via the primary accession number Q7SI68(http://www.expasy.org/uniprot/Q7SI68).

CHAPTER 5

82

Generation of a cryptochrome- k.o.- mutant in Neurospora crassaIn order to see effects of a lack of cry in Neurospora crassa, a mutant, crybdIm1A,was generated, where the cry-ORF has been removed by homologous recombina-tion. To select for the mutation we used the hygromycin- resistance cassette(HygR) from pCSN44. Briefly, three fragments were amplified by PCR: thehygromycin-cassette from the plasmid pCSN44, the 3´-UTR and the 5´-UTR ofthe cryptochrome gene (NCU00582.3) (see Fig. 5.1). The primers used (see Table5.1) generate specific restriction sites on each end of the fragments, so thathomodimers can be removed specifically after ligation and so that the fragmentsonly ligate in the following two combinations: (3´-UTR- cry) + (HygR) and(HygR) + (5´-UTR-cry).

These ligation products were used as templates for a PCR that uses primerswithin the HygR to generate an overlap within the hygR-cassette of about 300 bp(see Figure 5.1). These fragments were transformed into conidia of the strain bdA(FGSC# 1489), where they underwent homologous recombination.

Transformants were screened for hygromycin-resistance on plates containingbottom agar3 with 200µg/ml hygromycin. and verified by PCR, restriction digestsand sequencing. Homokaryons were produced by microconidiation.

Phenotyping of the cryptochrome-k.o.-mutant:The cry-k.o.-mutant (hereafter called cry-) was subjected to several protocols inrace tube assays compared to the lab standard strain bd (FGSC#1489, hereaftercalled cry+). The medium was comprised of 2% agar, 0.5% arginine, 1x Vogel’ssalts (Vogel, 1956) and 10 ng/ml Biotin. There was no additional carbon source.Race tubes were inoculated and incubated in constant light (LL) for 1–2 daysbefore transfer into specific protocols. Cool fluorescent white light or blue diodes(465 nm, Osram) were used as light sources, at an intensity of ca. 4 µE for either

CRYPTOCHROME

83

Table 5.1 Primers and primer sequences used to generate the ncry-knock-out-mutant.

primer name primer sequence

5’UTR-cry-fwd CTTGCCTCTTCCAACTTGAG5’UTR-cry-rev CTTGGACAACAGGTAAGG3’UTR-cry-fwd CCGCTTTCCTTACCAGCCTGCG3’UTR-cry-rev CCGCGCCTCCGTGCTCAAATAChygR-fwd GAGGTCGACAGAAGATGATATTGhygR-rev GAGTCTAGAAAGAAGGATTACCTC

one. All experiments were performed in temperature-controlled rooms set at aconstant temperature of 25ºC. Each experiment was additionally monitored fortemperature and light using Data Loggers (HOBOwarePro 2.3.1).

Experiments included incubation in constant darkness (DD/25ºC) to assessfree running period, entrainment in light cycles with alternating 12 hours of lightand darkness (12/12 LD), 2/10 LD and 6/6 LD. In some protocols, fluence titra-tion was included (0.004 to 4 µE). Another form of titration was entrainment ina photoperiodic series, from 2/22 LD to 22/2 LD, with white or blue light(465nm).

All race tubes were marked every other day to monitor conidiation rhythmsand growth rates. The tubes were scanned after completion of the experiment.The scans were saved as pict files, which were modified to increase contrast. Alldata from the imported pictures was analyzed using Chrono 6.7.1m 68K/PPC andChronOS X 0.9.3 (copyright Till Roenneberg (Roenneberg & Taylor, 2000)).

CHAPTER 5

84

5'UTR-cry

to generate overhangs in hygR-cassette

HygR 3'UTR-cry

PCR

PCR

3'UTR-cry HygR 5'UTR-cry

5'UTR-cry + HygR HygR + 3'UTR-cry

into Neurospora crassatransform

homologous recombination

digest fragments andligate

cry-gene replaced by hygR-cassette

Figure 5.1 Generation of a cryptochrome-mutant in Neurospora crassa. X indicates wherehomologous recombination events are likely to occur. The shaded area indicates where anoverlap is generated.

Results

The motif search within the nCRY protein sequenceSearching for motifs in the protein sequence of the nCRY yielded the followingresults (see Table 5.2): ● The unprocessed precursor of nCRY is 745 amino acids long with a molecular

weight of 81412 Da. The predicted protein is 723 aa long. ● A DNA-photolyase-homology domain is located in amino acid position 1-202. Also phosphorylation and glycosylation sites can be found: ● An N-glycosylation site is predicted at position 386-389.● There are numerous predicted phosphorylation sites, namely targets of Casein

kinase II, cAMP- and cGMP-dependent protein kinase, Protein kinase C andTyrosine kinase.

● The search yielded a putative photoreceptor function according to similarity.Similarity comparison predicts binding of 1 FAD (in aa-position 255-563) and1 5,10-methenyltetrahydrofolate (MTHF) per subunit and a potential mito-chondrial precursor location (position 1-22).

CRYPTOCHROME

85

Table 5.2 Motifs in the nCRY protein sequence. All data has been obtained viahttp://www.broad.mit.edu/annotation/genome/neurospora/Home.html),(http://expasy.org/prosite/) and (http://www.expasy.org/uniprot/Q7SI68).

Motif Position (amino acids)

DNA-photolyase domain 1-202FAD-binding 255-563N-glycosylation 386-389, 433-436CKII-phosphorylation 17-20, 77-80,105-108, 123-126, 144-147, 222-225,

276-279, 371-374, 421-424, 530-533cAMP-phosphorylation 571-574Protein kinase C (PKC) phosphorylation 177-179, 276-278, 290-292, 303-305, 338-340,

421-423, 432-434, 628-630, 654-656Tyrosine kinase phosphorylation 149-156N-myristoylation 199-204Glycine-rich region 594-722Histidine-rich region 637-655Serine-rich region 582-643Ice nucleation protein repeat 429-444

Phenotype of the cryptochrome-k.o.-mutant:IN CONSTANT DARKNESS (DD)In the animal kingdom, CRY functions either as a transcription factor or photore-ceptor. In plants, it is a photoreceptor. In all cases, it is an integral part of regu-lating the circadian clock. Thus, we subjected the cry-KO-mutant cry- to standardcircadian protocols (DD/25ºC and 12/12 LD). In constant darkness, the free-running period (t) of the mutant cry- is not different from the background straincry+ (22.2 h in cry+ vs. 22.0 h in cry-; Fig. 5.2).

Also, visual inspection of race tubes yielded no difference. We furthermorecompared the amplitudes of the rhythm in cry+ and cry- and did not find adisparity in the mutant.

IN LIGHT-DARK CYCLES IN THE CONTEXT OF 24 HOURS

Entrainment in 12/12 LD. The next standard protocol used was 12/12 LD(25ºC). In entrainment with 12/12 LD (Fig. 5.3) the cry-mutant is delayed byabout one hour in its conidiation (p- value= 0.0027).

Entrainment in a photoskeleton cycle (2/10/2/10 LD). Since the standard proto-cols revealed no or small phenotypic differences we subjected mutant and „wildtype“ cry+ to a photoskeleton cycle of 2/10 LD. We hypothesized that by

CHAPTER 5

86

cry+ cry–

τ = 22.2 h τ = 22.0 h

Figure 5.2 cry+ and cry– in DD/25ºC: Race tubes were kept in constant darkness andtemperature (25ºC) to measure period lengths. Double plots and race tubes are shown foreach strain. The double blots and τ have been calculated as an average from 12 race tubesper strain using Chrono 6.7.1m (Roenneberg and Taylor, 2000).

changing zeitgeber strength, we could exaggerate the phenotypic differencebetween mutant and wild type (Daan & Pittendrigh, 1976). In a skeletonphotoperiod protocol of 2/10LD the cry–mutant is significantly delayed by almost3 hours (42 º, p-value= 0.0357) (Fig. 5.5 and 5.6).

CRYPTOCHROME

87

cry+ cry–

Φon = 101.4° Φon = 113.5°

Figure 5.3 cry+ and cry– in LD 12/12 (25ºC), shown as average double plots (n = 12):Race tubes have been kept in constant temperature (25ºC) at alternating 12 hours of dark-ness and light (fluorescent white light, 4µE). Standard errors (in º) are 2.0 for cry+ and 2.5for cry–. Light is shown as white background, darkness in grey.

0

dens

ity (

coni

datio

n)

180 360phase (°)

cry+

cry–

**

Figure 5.4 In 12/12 LD at 25ºC, the cry-K.O.-strain cry– is delayed for less than one hour(12º) (p-value= 0.0027). The line plot shows averaged conidiation rhythms (n = 12) forcry+ and cry–. Light is shown as white background, darkness in grey.

CHAPTER 5

88

0 180 360phase (°)

0 180 360phase (°)

cry+ cry–

Φon = 82.1° Φon = 124.3°

Figure 5.5 cry+ and cry– in LD 2/10/2/10 (25ºC), shown as double plots: Race tubes havebeen kept in constant temperature (25ºC) at alternating 10 hours of darkness and 2 hoursof light (fluorescent white light, 4µE). The difference in Φon is almost 3 hours (42º) andsignificant (p-value = 0.0357). Standard errors are 10.7º for cry+ and 3.0º for cry–. Light isshown as white background, darkness in grey.

0

dens

ity (

coni

datio

n)

180 360phase (°)

cry+

cry–

***

27090

Figure 5.6 In a skeleton photoperiod protocol, 2/10LD at 25ºC, the cry-K.O.-strain cry– isdelayed for about 3 hours (p-value = 0.0357). The line plot shows averaged data for cry+

and the average of 3 microconidial derivatives of cry–. Light is shown as white background,darkness in grey.

Entrainment in a frequency-demultiplication protocol 6/6 LD at fluences from0.004 to 4 µE. Weak zeitgebers have been used previously to accentuate differ-ences in phase relationships (Pittendrigh & Daan, 1976c). We hypothesized thatwe would find a bigger phenotypic difference in the mutant when the zeitgeberstrength was decreased. We applied a frequency demultiplication protocol (6/6LD) in combination with different fluences from 0.004 to 4 µE (Fig. 5.7, only0.004 and 4 µE are shown). The lowest fluence yielded the biggest phenotypicdifference: whereas cry+ shows two peaks of different amplitude per day, thecry–mutant shows two peaks of similar height.

Entrainment in photoperiods from 2/22 LD to 22/2 LD. We also entrained cry– byscanning through from short to long photoperiods. In light cycles ranging fromvery short photoperiods (2/22 LD) to very long photoperiods (22/2 LD) usingwhite light, the delay of the cry-mutant compared to cry+ is bigger in longphotoperiods (more than 50% light). In short photoperiods (< 50% light) theentrainment of cry– and cry+ is locked to midnight, whereas in long photoperiodsit shifts towards dawn (Fig. 5.8).

CRYPTOCHROME

89

0

–100

–200

200

100

12 24 36 48time (h)

00.004 µE

4 µE

–100

–200

200

100

0

0 12 24 36 48time (h)

cry+ cry–

Figure 5.7 cry+ (left panels) and cry– (right) in 6/6 LD at different fluences of 4 (top) and0.004µE (bottom). Shown are the averages (black lines) of 6 replicates across two days asline plots. Light is shown as white background, darkness in grey.

Entrainment using white light contrasts that with spectrally defined blue light(465nm; Fig. 5.9). In blue light, entrainment of the mutant resembles cry+ inlong and short photoperiods, in that both strains are entrained relative to dusk inshort photoperiods and relative to dawn in long photoperiods. In short photope-

CHAPTER 5

90

LL

50%

DD

phot

oper

iod

12–12 0

cry +

cry –

Figure 5.8 cry+ and cry– in photoperiods using white light (shown as light grey back-ground). The points represent the onset of conidiation (in hours) in relation to midnight (indi-cated on the y-axis as 0, white vertical line). Each row represents an experiment whereas thephotoperiod (in h) increases from top (DD, 0 hours of light) to bottom (LL, 24 hours of light).The white horizontal line indicates 12/12 LD. The black lines are drawn arbitrarily.

LL

50%

DD

phot

oper

iod

12–12 0

cry +

cry –

Figure 5.9 cry+ and cry– in photoperiods using blue light (465nm, 4µE, shown as light greybackground). The points represent the onset of conidiation (in hours) in relation tomidnight (indicated on the y-axis as 0, white vertical line). Each row represents an experi-ment whereas the photoperiod (in h) increases from top (DD, 0 hours of light) to bottom(LL, 24 hours of light). The white horizontal line indicates 12/12 LD (50%). The black linesare drawn arbitrarily.

riods (up to 12/12 LD) the delay in the mutant increases with increasingphotoperiod (so that it is largest at around 10/14 and 12/12 LD). When morethan 50% of the cycle is light-exposed, the delay in cry– vs cry+ is smallercompared to short photoperiods, unlike in white light photoperiods of the samelength (see Fig. 5.8). All in all, especially short blue photoperiods (<50% light)result in a larger delay in entrainment than does white light. The mutant isdelayed significantly (p-value = 0.0052) for more than 1 hour in these condi-tions (short blue photoperiods compared to short white photoperiods).

Discussion

This work describes our analysis of the Neurospora cry gene. Based on itspredicted amino acid sequence, it is most similar to CRY proteins found inanimals, all of which have defined functions within circadian clocks. Although wepredict that this will primarily occur via photoreception in the case of nCRY (ithas important co-factor binding sites), it could also be integrated in the clockgene network via phosphorylation. Phosphorylation represents one of the keyshared features of clock molecular networks (Merrow et al, 2006b). nCRY haspotential targets for CKII and other kinases, which could here – as has beendemonstrated for FRQ - regulate phosphorylation and ensuing degradationprocesses in a circadian manner. We have produced antibody to nCRY which willbe used to investigate this possibility.

On the phenotypic level we investigated several properties of the ncry-mutant. Unlike in mice, where a k.o. of cry1 shortens and a k.o. of cry2 lengthensthe free running period (τ) (van der Horst et al, 1999), we did not find a differ-ence in period of the ncry mutant compared to the background strain bd.Furthermore, we found only a one hour delayed phase in the mutant in 12/12LD (p= 0.0027). Weaker zeitgebers spread out entrained phases in a population(Daan & Pittendrigh, 1976; Roenneberg & Merrow, 2003). We applied in thefollowing skeleton photoperiods (e.g. 2/10 LD), frequency demultiplicationprotocols and did a fluence titration, where we used light intensities of 0.004,0.04, 0.4 and 4 µE in 6/6 LD. Then, we assessed a series of systematic photope-riods from 2/22 LD to 22/2 LD. In summary, cry– shows a more obvious mutantphenotype in 2/10 LD, with a phase delay of 2 hours compared to cry+. In 6/6LD, we see the biggest difference in the cycle with the lowest light intensity:whereas cry+ shows alternating high and low amplitudes of conidiation in each12h bin, the peaks in the mutant are all of the same magnitude. This findingmight speak for stronger masking effects due to higher sensitivity to light as azeitgeber in the cry-mutant. In the photoperiodic series using white light we

CRYPTOCHROME

91

observed an increased phase delay of the cry-mutant compared to cry+ in longphotoperiods (>50% light). In white light entrainment is linked to midnight inshort photoperiods (<50% light), as shown by our group before (Merrow et al,1999a; Tan et al, 2004)Rémi, 2007). In long photoperiods with white light,entrainment in both strains shifts towards dawn.

As all known light responses in Neurospora crassa have a peak in the bluespectrum at 465nm and since it has been shown that also Arabidopsis CRY has anabsorption maximum at approximately this wave length (Ahmad et al, 2002) wetried to find out whether a ncry-mutant would be impaired in responses to bluelight of 465nm. In the first light cycle applied, 8/16 LD, we found a highly signif-icant delay of 3 hours. The difference in entrainment in short vs long photope-riods (shown in white photoperiods here and also in our group before, e.g. Rémi,2007) can be observed in blue photoperiods, as well. Here it is even morepronounced, in that cry+ and cry– entrain relative to dusk in short photoperiodsand relative to dawn in long photoperiods. This behaviour could speak for intra-cellular morning- and evening oscillators, that integrate dusk and dawn signalsdifferently, as already proposed by our group (Rémi, 2007). Importantly, themutant is on average more than 1 hour phase-delayed in short photoperiods withblue light compared to photoperiods using white light. The important question iswhether the observed phase differences in white and blue photoperiods are dueto quantity or quality of the light applied. Reciprocity experiments (here fluencerate and time are varied to yield the same response) and experiments applyingcombinations of wave lengths could help in discovering the reasons for theobserved effects.

The comparison of the ncry-mutant with the vvd-knockout mutant is inter-esting, in that both mutants show similar properties, namely impaired entrain-ment but no difference in t (Shrode et al, 2001). As already mentioned in chap-ters 3 and 4 of this thesis and as suggested by Michael et al. (Michael et al,2003), the traits period and phase will sometimes be under the control ofdifferent genes. This might explain why in many cases (see chapter 3 and 4)phase and period are not correlated and why we find in some cases - includingthe ncry- and vvd-mutants - no difference in period compared to the wild type buta difference in phase. One could imagine also, that mutations in input or outputgenes might not affect the central oscillator (and hence have no impact onperiod) but lead to a delayed or advanced phase. If one believes that entrain-ment reflects the function of the circadian system in nature, then looking at τeffects in mutants is not the only reliable method for classifying ‘core’ clock genefunctions. Given that entrainment with weak or partial zeitgebers resulted in amore obvious phenotype, we suggest that standard protocols, e.g. 12/12 LD andDD/25ºC, are not (always) sufficient to characterize – or, even more important,

CHAPTER 5

92

unveil - clock mutants. In a recent publication Brown et al (Brown et al, 2008)report, that fibroblast cultures from extreme human chronotypes, so-called„larks“ and „owls“, show the same free running period in almost 50% of thecases. It was hypothesized that this observation might be due to differences inamplitude of the free running rhythm. This could be the case also in the ncrymutant and could be a further point of investigation.

Based on the mutant chronotype of cry–, we propose that the cry-mutant is aclock mutant, likely through involvement of nCRY in circadian photoreception. Inaddition, non-photoreceptive functions, as e.g. in mammals, are possible.

Conclusion and future outlook

In this study a cry mutant in Neurospora has been described in various physiolog-ical protocols as a potential clock mutant. In this mutant phase delays areobserved, despite a free-running period that is equal to the ‚wild type’. To date,the field is still wide open for work on nCRY. Future work might concentrate ondifferent read-outs of the clock, especially the molecular characterization of thecry-mutant. In this respect it is of great interest to study interacting partners,rhythmic expression, possible de-/phosphorylation and effects of combinedknock-outs together with other photoreceptors or photoreceptor candidates.

References

CRYPTOCHROME

93

Ahmad M, Cashmore AR (1993) HY4 gene of A. thaliana encodes a protein with characteris-tics of a blue-light photoreceptor. Nature 366(6451): 162-166

Ahmad M, Grancher N, Heil M, Black RC, Giovani B, Galland P, Lardemer D (2002) Actionspectrum for cryptochrome-dependent hypocotyl growth inhibition in Arabidopsis. PlantPhysiol 129(2): 774-785

Bieszke JA, Braun EL, Bean LE, Kang S, Natvig DO, Borkovich KA (1999) The nop-1 gene ofNeurospora crassa encodes a seven transmembrane helix retinal-binding protein homolo-gous to archaeal rhodopsins. Proc Natl Acad Sci U S A 96(14): 8034-8039

Borkovich KA, Alex LA, Yarden O, Freitag M, Turner GE, Read ND, Seiler S, Bell-Pedersen D,Paietta J, Plesofsky N, Plamann M, Goodrich-Tanrikulu M, Schulte U, Mannhaupt G,Nargang FE, Radford A, Selitrennikoff C, Galagan JE, Dunlap JC, Loros JJ, Catcheside D,Inoue H, Aramayo R, Polymenis M, Selker EU, Sachs MS, Marzluf GA, Paulsen I, DavisR, Ebbole DJ, Zelter A, Kalkman ER, O'Rourke R, Bowring F, Yeadon J, Ishii C, Suzuki K,Sakai W, Pratt R (2004) Lessons from the genome sequence of Neurospora crassa:tracing the path from genomic blueprint to multicellular organism. Microbiol Mol BiolRev 68(1): 1-108

Brown SA, Fleury-Olela F, Nagoshi E, Hauser C, Juge C, Meier CA, Chicheportiche R, DayerJM, Albrecht U, Schibler U (2008) The period length of fibroblast circadian gene expres-sion varies widely among human individuals. PLoS Biol 3(10): e338

CHAPTER 5

94

Cashmore AR, Jarillo JA, Wu YJ, Liu D (1999) Cryptochromes: blue light receptors for plantsand animals. Science 284(5415): 760-765

Ceriani MF, Darlington TK, Staknis D, Mas P, Petti AA, Weitz CJ, Kay SA (1999) Light-dependent sequestration of TIMELESS by CRYPTOCHROME. Science 285(5427): 553-556

Daiyasu H, Ishikawa T, Kuma K, Iwai S, Todo T, Toh H (2004) Identification of cryptochromeDASH from vertebrates. Genes Cells 9(5): 479-495

Darlington TK, Wager-Smith K, Ceriani MF, Staknis D, Gekakis N, Steeves TD, Weitz CJ,Takahashi JS, Kay SA (1998) Closing the circadian loop: CLOCK-induced transcriptionof its own inhibitors per and tim. Science 280(5369): 1599-1603

Elvin M, Loros JJ, Dunlap JC, Heintzen C (2005) The PAS/LOV protein VIVID supports arapidly dampened daytime oscillator that facilitates entrainment of the Neurosporacircadian clock. Genes Dev 19(21): 2593-2605

Franchi L, Fulci V, Macino G (2005) Protein kinase C modulates light responses in Neuro-spora by regulating the blue light photoreceptor WC-1. Mol Microbiol 56(2): 334-345

Froehlich AC, Liu Y, Loros JJ, Dunlap JC (2002) White Collar-1, a circadian blue lightphotoreceptor, binding to the frequency promoter. Science 297(5582): 815-819

Froehlich AC, Noh B, Vierstra RD, Loros J, Dunlap JC (2005) Genetic and molecular analysisof phytochromes from the filamentous fungus Neurospora crassa. Eukaryot Cell 4(12):2140-2152

Galagan JE, Calvo SE, Borkovich KA, Selker EU, Read ND, Jaffe D, FitzHugh W, Ma LJ,Smirnov S, Purcell S, Rehman B, Elkins T, Engels R, Wang S, Nielsen CB, Butler J,Endrizzi M, Qui D, Ianakiev P, Bell-Pedersen D, Nelson MA, Werner-Washburne M,Selitrennikoff CP, Kinsey JA, Braun EL, Zelter A, Schulte U, Kothe GO, Jedd G, Mewes W,Staben C, Marcotte E, Greenberg D, Roy A, Foley K, Naylor J, Stange-Thomann N,Barrett R, Gnerre S, Kamal M, Kamvysselis M, Mauceli E, Bielke C, Rudd S, Frishman D,Krystofova S, Rasmussen C, Metzenberg RL, Perkins DD, Kroken S, Cogoni C, Macino G,Catcheside D, Li W, Pratt RJ, Osmani SA, DeSouza CP, Glass L, Orbach MJ, Berglund JA,Voelker R, Yarden O, Plamann M, Seiler S, Dunlap J, Radford A, Aramayo R, Natvig DO,Alex LA, Mannhaupt G, Ebbole DJ, Freitag M, Paulsen I, Sachs MS, Lander ES, NusbaumC, Birren B (2003) The genome sequence of the filamentous fungus Neurospora crassa.Nature 422(6934): 859-868

Guo H, Yang H, Mockler TC, Lin C (1998) Regulation of flowering time by Arabidopsisphotoreceptors. Science 279(5355): 1360-1363

Heintzen C, Loros JJ, Dunlap JC (2001) The PAS protein VIVID defines a clock-associatedfeedback loop that represses light input, modulates gating, and regulates clock resetting.Cell 104(3): 453-464

Kavakli IH, Sancar A (2002) Circadian photoreception in humans and mice. Mol Interv 2(8):484-492

Krishnan B, Levine JD, Lynch MK, Dowse HB, Funes P, Hall JC, Hardin PE, Dryer SE (2001)A new role for cryptochrome in a Drosophila circadian oscillator. Nature 411(6835):313-317

Merrow M, Brunner M, Roenneberg T (1999) Assignment of circadian function for theNeurospora clock gene frequency. Nature 399(6736): 584-586

Merrow M, Mazzotta G, Chen Z, Roenneberg T (2006) The right place at the right time:regulation of daily timing by phosphorylation. Genes Dev 20(19): 2629-2623

Michael TP, Salome PA, Yu HJ, Spencer TR, Sharp EL, McPeek MA, Alonso JM, Ecker JR,McClung CR (2003) Enhanced fitness conferred by naturally occurring variation in thecircadian clock. Science 302(5647): 1049-1053

CRYPTOCHROME

95

Nagashima K, Matsue K, Konishi M, Iidaka C, Miyazaki K, Ishida N, Kanosue K (2005) Theinvolvement of Cry1 and Cry2 genes in the regulation of the circadian body temperaturerhythm in mice. Am J Physiol Regul Integr Comp Physiol 288(1): R329-335

Pittendrigh CS, Daan S (1976) A functional analysis of circadian pacemakers in nocturnalrodents. IV. Entrainment: Pacemaker as clock. J Comp Physiol A 106: 291-331

Rémi J (2007) The circadian surface of Neurospora crassa- From physiology to molecularmechanisms Doctoral Thesis University of Munich, LMU

Roenneberg T, Merrow M (2003) The network of time: understanding the molecular circa-dian system. Curr Biol 13(5): R198-207

Roenneberg T, Taylor W (2000) Automated recordings of bioluminescence with special refer-ence to the analysis of circadian rhythms. Methods Enzymol 305: 104-119

Sancar A (2003) Structure and function of DNA photolyase and cryptochrome blue-lightphotoreceptors. Chem Rev 103(6): 2203-2237

Schafmeier T, Haase A, Kaldi K, Scholz J, Fuchs M, Brunner M (2005) Transcriptional feed-back of Neurospora circadian clock gene by phosphorylation-dependent inactivation ofits transcription factor. Cell 122(2): 235-246

Schwerdtfeger C, Linden H (2003) VIVID is a flavoprotein and serves as a fungal blue lightphotoreceptor for photoadaptation. EMBO J 22(18): 4846-4855

Shimura M, Ito Y, Ishii C, Yajima H, Linden H, Harashima T, Yasui A, Inoue H (1999)Characterization of a Neurospora crassa photolyase-deficient mutant generated by repeatinduced point mutation of the phr gene. Fungal Genet Biol 28(1): 12-20

Shrode LB, Lewis ZA, White LD, Bell-Pedersen D, Ebbole DJ (2001) vvd is required for lightadaptation of conidiation-specific genes of Neurospora crassa, but not circadian conidia-tion. Fungal Genet Biol 32(3): 169-181

Tan Y, Merrow M, Roenneberg T (2004) Photoperiodism in Neurospora crassa. J BiolRhythms 19(2): 135-143

van der Horst GT, Muijtjens M, Kobayashi K, Takano R, Kanno S, Takao M, de Wit J, VerkerkA, Eker AP, van Leenen D, Buijs R, Bootsma D, Hoeijmakers JH, Yasui A (1999)Mammalian Cry1 and Cry2 are essential for maintenance of circadian rhythms. Nature398(6728): 627-630

Van Gelder RN (2002) Tales from the crypt(ochromes). J Biol Rhythms 17(2): 110-120Van Gelder RN, Gibler TM, Tu D, Embry K, Selby CP, Thompson CL, Sancar A (2002)

Pleiotropic effects of cryptochromes 1 and 2 on free-running and light-entrained murinecircadian rhythms. J Neurogenet 16(3): 181-203

Van Gelder RN, Sancar A (2003) Cryptochromes and inner retinal non-visual irradiancedetection. Novartis Found Symp 253: 31-42; discussion 42-55, 102-109, 281-104

Vogel HJ (1956) A convenient growth medium for Neurospora (Medium N). Microbiol GenetBull 13: 42-43

Yager LN, Lee HO, Nagle DL, Zimmerman JE (1998) Analysis of fluG mutations that affectlight-dependent conidiation in Aspergillus nidulans. Genetics 149(4): 1777-1786

Entrainment of the Neurospora circadian clock

M. Merrow, C. Boesl, J. Ricken, M. Messerschmitt,M. Goedel and T. Roenneberg

Based on a publication in Chronobiology International 30 (5), (2006): 553-555

6CHAPTER

ABSTRACT

Neurospora crassa has been systematically investigated for circadian entrainment behavior.Many aspects of synchronization can be investigated in this simple, cellular system, rangingfrom systematic entrainment and drivenness to masking. Clock gene expression duringentrainment and entrainment without clock genes suggest that the known transcription/translation feedback loop is not alone responsible for entrainment in Neurospora.

.

Introduction

The circadian clock is a self-sustained biological oscillator with a period of 24 hin constant conditions. Sets of clock genes have been identified in animals,plants, fungi, and cyanobacteria, functioning as a transcription/translation nega-tive feedback loop. Their mutation often results in a change in the free-runningcircadian rhythm. Circadian clocks in nature are, however, rarely subjected to theconstant conditions that allow a free-running oscillation. They are normallyexposed to a rhythmic environment, so that appropriate signals (zeitgebers),such as light, temperature, or occasionally even social cues, feed into the clockand entrain its oscillation to the 24 h day. The phase relationships of oscillatingprocesses can be expected to be novel according to the entraining condition.They can move to different times of day or be suppressed depending on the struc-ture of the cycle (Roden et al., 2002). Yet, entrainment is clearly the most rele-vant state for an organism’s survival and also the state that was subjected toselection in the course of evolution (Roenneberg and Merrow, 2002a, 2002b).Because of the high complexity of circadian systems, both on the molecular-cellular as well as on the systemic level in higher plants and animals, we used arelatively simple model system, the filamentous fungus Neurospora crassa tostudy the molecular basis of circadian entrainment. Neurospora is a haploid, fila-mentous fungus with a sequenced genome of only 40 million base-pairs, anno-tated to about 10,000 genes (Galagan et al., 2003). A circadian rhythm inasexual spore (conidium) formation was first observed by Pittendrigh andcoworkers (1959). In constant darkness, the period is about 22 h; in constantlight, at any level greater than that approximating moonlight, conidia formationis arhythmic (Sargent et al., 1956). Conidiation was used to generate a panel ofclock mutants, making Neurospora the second system for molecular approachesto circadian rhythms after Drosophila (Feldman and Hoyle, 1973; Konopka andBenzer, 1971).

The Neurospora clock gene frequency (frq) was cloned and used to demon-strate the concept of negative feedback in clock regulatory loops, by showing thatFRQ-protein over-expression shuts down transcription from the endogenous frqlocus (Aronson et al., 1994; McClung et al., 1989). Both the lack of functional frqRNA (Aronson et al., 1994a, 1994b; Loros and Feldman, 1986) and its constitu-tive over-expression result in arhythmicity. The activators of frq transcriptioninclude the blue light photoreceptor WC-1 and its partner WC-2 (Figure 6.1)(Ballario et al., 1996; Froehlich et al., 2002; He et al., 2002; Linden and Macino,1997). The mechanism of negative feedback is likely not to act via direct interfer-ence with the transcription factor complex on the promoter. Rather, FRQ regu-lates the phosphorylation state of WC-1 and WC-2, which controls their activity

CHAPTER 6

98

(Schafmeier et al., 2005). This may also be the mechanism by which FRQ regu-lates WC-1 levels (Lee et al., 2000; Merrow et al., 2001), giving the appearanceof both positive and negative feedback loops among one set of molecules. Anadditional photoreceptor, VIVID (VVD) is found in the cytoplasm (Schwerdtfegerand Linden, 2003), and it regulates the clock in entrainment, even though thefreerunning period of a vvd mutant is not different from that of a wild-type strainunder numerous culture conditions (Elvin et al., 2005; Heintzen et al., 2001;Shrode et al., 2001).

We have set out to systematically describe the entrainment properties ofNeurospora using many of the classical circadian protocols that have been usedwith other species for the purposes of comparison and validation of theNeurospora model system.

Entrainment with light

Entrainment is characterized by a stable phase reference point in relationship toan entraining cycle (e.g., light/dark or warm/cold). A reference point or phase inthe biological oscillation is chosen (i.e., core body temperature nadir or mela-

ENTRAINMENT

99

FRQ

FRQPh

VVD VVDPh

vvdWC-1?, WC-2?

frqWC-1Ph, WC-2Ph

frqWC-1, WC-2

Figure 6.1 Molecular clock network in Neurospora. A transcription-translation negativefeedback loop is described, whereby FRQ protein feeds back negatively on its own RNAexpression. The transcriptional activators of frq and other light-induced genes include WC-1and WC-2, whose activity is modified according to phosphorylation state, apparently themechanism by which FRQ exerts negative feedback. In theory, it is possible the transcriptionfactor complexes with mixed phosphorylation states could result in fine-tuning of transcrip-tional regulatory responses (shown here as WC-1? and WC-2? on the vvd promoter). Theseveral kinases and phosphatases and post-transcriptional regulation that are part of clockregulation are omitted here.

tonin onset in humans, or conidiation onset in Neurospora) and the differencebetween this internal time (phase) and a chosen reference of the external, zeit-geber cycle (lights on, lights off, midnight, etc.) is called the phase of entrain-ment. Phase can be expressed in real h or in degrees (with 360º representing afull zeitgeber cycle no matter how long or short its period is in real h). In thecase of Neurospora, with the recent introduction of rigorous quantificationmethods (Roenneberg and Taylor, 2000), virtually any point in the conidiationcycle can be used as phase reference, allowing critical evaluation of the entirewaveform, as it appears when entrained under different zeitgeber conditions.The waveform of the conidiation rhythm discussed here can be well described byusing four standard phase reference points: onset, peak, offset, and trough. In thecase of conidiation, onset (the upward transition through the non-rhythmictrend) has proven to be the most reliable marker for phase of entrainment(Roenneberg et al., 2005) so we will refer only to onset of conidiation here.

Entrained phase can be determined in 24 h cycles as well as in shorter orlonger zeitgeber cycles (e.g., T = 18 to T = 26h). In non-24h T-cycles, thebiological clock is expected to entrain earlier in long cycles and later in shortcycles, as shown decades ago, e.g.,with lizards and hamsters (Hoffmann, 1963;Pittendrigh and Daan, 1976). In symmetrical T-cycles with alternating light (L)and dark (D) conditions (LD, e.g., 50 % of each cycle in L and in D), Neurosporaapparently breaks this rule in that conidiation onset lies 7 h after dusk, irrespec-tive of cycle length (Figure 6.2A) (Merrow et al., 1999). This finding suggestedan hourglass system, with light driving the formation of the conidial band ratherthan it being controlled by an entrained biological oscillator–a puzzling result fora model circadian system.

We therefore probed the system with alternative entrainment protocols, i.e.,by systematically changing the duration of light (photoperiod) within a T = 24 h.When the phase of entrainment (judged by onset of conidiation) was measuredin long and short photoperiods, it always appeared around the middle of thenight (Figure 6.2B; Tan et al., 2004a). Unlike a fixed relationship to dawn ordusk, a fixed relationship to midnight (independent of night-length) means thatthe phase of entrainment, in reference to both dawn or dusk, changes systemati-cally with photoperiod.

Thus, the Neurospora clock shows entrainment under these conditions ratherthan drivenness (as described above).

We have additionally investigated molecular aspects of entrainment by light,focussing on FRQ for several reasons: induction of its RNA by light has beenequated with phase resetting and entrainment (Collett et al., 2002; Crosthwaiteet al., 1995; Elvin et al., 2005; Liu, 2003), and our work had shown that frq nullmutants fail to entrain to light (Merrow et al., 1999). When Neurospora is trans-

CHAPTER 6

100

ferred from darkness to light, both frq RNA and FRQ protein are rapidly induced(Collett et al., 2002; Crosthwaite et al., 1995, 1997). FRQ protein levels take 4 to6 h until they reach a plateau in continuous light, while RNA levels are down-regulated following a peak at 30 min and are maintained at an adapted level,

ENTRAINMENT

101

D

C

A

T = 24

R

T = 22

T = 20

LD = 8:16

B LD = 12:12

LD = 16:8

LD = 20:4

LD = 14:10

LD = 4:20

T = 26

E

T = 22

T = 22

RP P

R RP P

R RP P

T = 12

Figure 6.2 Summary of light and temperature entrainment protocols with the band strain ofNeurospora. In panels A, B, and C, grey areas represent a dark incubation period; the openbox is light. In panels D and E, the grey area represents cool temperature; whereas, the openarea concerns the warm phase of the cycle. Panels A, B, D, and E use a cartoon of the coni-dial band to illustrate entrainment of banding. A. T cycles with light show an onset of coni-diation that occurs the same number of hours after darkness, independent of cycle length,thus showing driven, non-entrained states. B. In cycles of different photoperiods but T = 24h, conidiation onset is typically at midnight, indicating that both dawn and dusk signals areintegrated for entrainment. C. R represents when RNA is induced; P shows when protein isinduced; a shadowed R and P indicate when RNA protein levels decrease, respectively.Although RNA levels respond predictably and directly according to light conditions, FRQprotein levels increase either rapidly or in an attenuated fashion when lights come on. Theprotein decreases at midnight, as conidiation is beginning. D. Entrainment of conidiation intemperature cycles of different lengths is systematic, in that it is shows a later phase in shortcycles, and an earlier one in long cycles. E. Neurospora frequency demultiplies in T = 12 htemperature cycles.

which is still higher than levels in dark-grown tissue. Similar experiments wereperformed with the reverse protocol, transferring Neurospora tissue from light todarkness, and these two single transition protocols were purported to explainentrainment (Collett et al., 2002; Liu, 2003). However, if frq or FRQ are involvedin systematic circadian entrainment that occurs in T = 24 h photoperiod cycles,then they should be expressed with different kinetics in different cycles. Wefound this to be true for the protein but not the RNA (Figure 6.2C; Tan et al.,2004b). The frq RNA-kinetics are independent of photoperiod (and night length,scotoperiod). RNA is rapidly induced by the dawn light and photo-adapts duringthe day; it decreases rapidly after nightfall and then—but only in nights longerthan 6 to 8h— gradually increases during the scotoperiod. Thus, all aspects ofRNA expression appear driven by non-circadian responses to the light environ-ment. Because drivenness can be a special case of entrainment by a very strongsignal that always resets the clock to a certain phase (Roenneberg et al., 2005), itwould be more appropriate to state that frq RNA is masked by light (see below).The protein, however, is only induced rapidly (like it was in the single-releaseexperiments) in short photoperiods. Its expression is substantially delayed in longphotoperiods in spite of the rapid dawnlinked RNA-induction, indicating thatpost-transcriptional regulation of FRQ expression carries key information forcircadian entrainment. The systematic responses to different photoperiods on themolecular level in Neurospora invite the inquiry as to whether systematic entrain-ment according to season is meaningful in the Neurospora life cycle. We deter-mined the quantitative yield of three light-regulated processes and found theywere all specifically controlled by photoperiod (Tan et al., 2004a). Conidiation(asexual propagation) is most abundant in LD cycles of 12 : 12; more sexualspores are produced in 14 : 10 cycles; and mycelial carotenoids increase over abroad range of photoperiods from about 10 to 20 h. In all cases, longer photope-riods decrease the output. In addition, these responses are disrupted in clockmutant strains, proving that these are not simply irradiance responses but repre-sent photoperiodism in Neurospora that is somehow tied to the circadian clock.

To summarize, we can conclude the following about entrainment of theNeurospora clock by light and darkness: although it appears to be driven insymmetrical LD T-cycles, circadian entrainment is apparent when differentphotoperiods are used in the context of 24 h cycles.

Systematic circadian entrainment can be seen at the level of FRQ protein,while frq RNA passively reacts to light (is masked), so that transcriptional regula-tion can be ruled out as a player of the entrainment mechanism.The benefit of entrainment for fitness exists both on the daily level (anticipationof changing environmental conditions) as well as on the seasonal level (e.g., bythe time-of-year-specific enhancement of reproduction).

CHAPTER 6

102

Entrainment with temperature

We investigated entrainment with another zeitgeber, namely temperature(Merrow et al., 1999). Although circadian systems are compensated for differenttemperature levels, they do respond to temperature changes; thus, clocks canalso entrain to temperature as a zeitgeber. In symmetrical T-cycles, using alterna-tions of 22 and 27ºC, entrained phase is earlier in long and later in short cycles.Clock mutant strains with short or long free-running periods, respectively, alsopredictably and consistently entrain earlier or later (Figure 6.2D). When cyclesare shortened to about half of the free-running rhythm (e.g., T=12 h for the wildtype strain), a single conidiation bout occurs each 24 h (i.e., every second cycle;a so-called ‘frequency demultiplication’; Figure 6.2E). This demultiplication is anindication of a robust circadian oscillator.

Thus, the Neurospora clock performs as predicted for a (biological) oscillatorin temperature cycles. Using onsets as phase reference points, it has been shownby a number of laboratories that even clock null strains (at the frq locus) showsystematic entrainment in these T-cycles (Merrow et al., 1999; Pregueiro et al.,2005; Roenneberg et al., 2005) This is an exciting discovery, suggesting a multi-oscillator circadian system in Neurospora, like that seen in humans, rodents, flies,and plants (Aschoff et al., 1967; Grima et al., 2004; Honma et al., 1983;Johnson, 2001; Stoleru et al., 2004). Thus, the study of even a simple fungus(with neither organs nor a brain) shows the circadian system is a complex mech-anism that consists of multiple oscillators, implying that this is an importantadaptive feature of circadian clocks.

Experiments in constant conditions demonstrate that temperature regulatesfrq RNA splicing (Diernfellner et al., 2005). Thus, there are at least two levels—transcriptional, regulating the timing of RNA expression, and post-transcrip-tional, regulating splicing—to control for different amounts of FRQ proteinaccording to temperature signals (Liu et al., 1998). At the molecular level,temperature entrainment also contrasts light entrainment. Here, frq RNA isclearly not driven, appearing early in a long and later in a short cycle, and thisindicates that it does not respond like an hourglass timer (Merrow et al., 1999).

Masking

The behavior of frq RNA when light is used as a zeitgeber (Figure 6.2C) suggests‘masking’ on the molecular level. Masking is an acute, non-circadian effect on thesystem that can, nonetheless, be induced by a zeitgeber signal. A commonexample in mice is that they typically stop running when lights are turned on,

ENTRAINMENT

103

regardless of the time of day. This can confound straightforward determination ofentrained phase, requiring release of the animals from the zeitgeber cycle intoconstant darkness, which allows an estimation of when they would have startedrunning if their activity had not been acutely inhibited by light. Masking demon-strates that zeitgebers have effects other than dedicated circadian ones. So,although frq RNA is a component of the clock, it can be masked by light.Conidiation in Neurospora can also show masking, as demonstrated in responseto temperature (Pregueiro et al., 2005). Such effects are seen in the wild typestrain, and can become even more prominent in the frq-less mutants (Merrow etal., 1999; Roenneberg et al., 2005). Changing zeitgeber strength is one of thebest protocols to distinguish between entrainment (phase changes) and masking(phase remains the same) (Roenneberg et al., 2005).

Discussion

The use of Neurospora to characterize entrainment principles has several advan-tages. To fully describe and eventually understand entrainment, a large numberof experiments must be performed, which systematically scan different cyclelengths, amplitudes, and light or temperature portions (photo- and thermope-riod) of the zeitgeber in different clock mutants. These can be readily done inNeurospora, which is a powerful molecular genetic model system as well as aneconomical (non-animal) system. Comparison of entrainment at the level ofphysiology and gene expression shows that regulation of frq RNA by the circadiansystem occurs in temperature but not in light cycles. Temperature cycles alsodemonstrate circadian clock characteristics (an entrainable, frequency de-multi-plying oscillator) in frq null mutants, indicating that the exact function of theFRQ-WC transcription/translation feedback loop within the Neurospora clockmust be reconsidered. Experiments (Merrow et al., 1999; Merrow et al., 2001)and modelling (Roenneberg and Merrow, 1998, 1999, 2002a) indicate that inputpathways into the circadian clock are both interfaces to the environment andintegral components of the rhythmgenerating mechanism, so-called “zeit-nehmers” (German for time taker, Roenneberg et al., 1998) that are, themselves,under circadian control. The frq/FRQ oscillator is part of the light input; whentransducing light information, frq RNA is predominantly controlled by the inputsignal, while it shows its circadian regulation when the system is entrained viadifferent inputs, e.g., by temperature. Renewing the conceptual view of the FRQ-WC loop within the circadian system does not diminish its dominant role in clockfunction, e.g., by determining chronotype and thereby phase relationships ingeneral.

CHAPTER 6

104

We used a mathematical model to simulate a circadian system that is com-posed of a network of feedback loops (Roenneberg and Merrow, 2002a). Theindividual feedbacks, when isolated, do not show circadian properties; however,the intact network does. Some could be involved in driving outputs, others inprocessing a specific zeitgeber (the zeitnehmer loops). A zeitnehmer feedbacksupplies rhythmic input, even in constant conditions (comparable to animalschanging retinal light exposure by closing their eyes as part of the circadianlycontrolled behavior in constant light). Each of the feedbacks in the network isessential for the entire system, but experimental in silico mutagenesis suggeststhat they would be discovered more (close to or meditating zeitgeber input) orless often (distant from zeitgeber input) in mutant screens (Merrow andRoenneberg, 2005).

If we take Neurospora as a model system, then a logical extension is to ask:What is the zeitnehmer in other clock model systems? An intriguing examplerecently surfaced, with photoreceptor mutant mice (lacking melanopsin androds) showing large changes in entrained phase, like frq mutants have shown(Mrosovsky and Hattar, 2005). So, at the level of the organism, photoreceptorsengaged with the circadian system can regulate chronotype, a clear example ofinputs determining phase. The mammalian circadian system is hierarchical in thesense that many cells have been demonstrated to oscillate as cellular clocks, evenwhen cultured as single-cell suspensions (Balsalobre et al., 1998; Welsh et al.,1995). If they are each a clock, then a constructive approach is to ask: Whichcellular clock components serve as zeitnehmers? In some cases, what we think ofas central clock genes may be functioning in this role.

References

ENTRAINMENT

105

AcknowledgementsWe thank the organizers of the European Pineal and Biological Rhythms Society Conferencefor an excellent meeting and for the opportunity to prepare this review. Our work issupported by the Deutsche Forschungsgemeinschaft, the Nederlandse Organisatie voorWetenschappelijk Onderzoek, the European commission, the Dr. Meyer-Struckman-Stiftung,and the Daimler-Benz Stiftung

Aronson, B.D., Johnson, K.A., Dunlap, J.C. (1994a). The circadian clock locus frequency: asingle ORF defines period length and temperature compensation. Proc. Natl. Acad. Sci.USA. 91:7683–7687.

Aronson, B.D., Johnson, K.A., Loros, J.J., Dunlap, J.C. (1994b). Negative feedback defininga circadian clock: autoregulation of the clock gene frequency. Science 263:1578–1584.

Aschoff, J., Gerecke, U., Wever, R. (1967). Desynchronization of human circadian rhythms.Jap. J. Physiol. 17:450–457.

Ballario, P., Vittorioso, P., Magrelli, A., Talora, C., Cabibbo, A.,Macino, G. (1996). Whitecollar-1, a central regulator of blue light responses in Neurospora, is a zinc fingerprotein. EMBO J. 15:1650–1657.

CHAPTER 6

106

Balsalobre, A., Damiola, F., Schibler, U. (1998). A serum shock induces gene expression inmammalian tissue culture cells. Cell 93:929–937.

Collett, M.A., Garceau, N., Dunlap, J.C., Loros, J.J. (2002). Light and clock expression of theNeurospora clock gene frequency is differentially driven by but dependent on WHITECOLLAR-2. Genetics 160:149–158.

Crosthwaite, S.K., Loros, J.J., Dunlap, J.C. (1995). Light-induced resetting of a circadianclock is mediated by a rapid increase in frequency transcript. Cell 81:1003–1012.

Crosthwaite, S.K., Dunlap, J.C., Loros, J.J. (1997). Neurospora wc-1 and wc-2: Transcription,photoresponses, and the origin of circadian rhythmicity. Science 276:763–769.

Diernfellner, A.C., Schafmeier, T., Merrow, M.W., Brunner, M. (2005). Molecular mecha-nisms of temperature sensing by the circadian clock of Neurospora crassa. Genes Dev.19:1968–1973.

Elvin, M., Loros, J.J., Dunlap, J.C., Heintzen, C. (2005). The PAS/LOV protein VIVIDsupports a rapidly dampened daytime oscillator that facilitates entrainment of theNeurospora circadian clock. Genes Dev. 19:2593–2605.

Feldman, J.F., Hoyle, M.N. (1973). Isolation of circadian clock mutants of Neurospora crassa.Genetics 75: 605–613.

Froehlich, A.C., Liu, Y., Loros, J.J., Dunlap, J.C. (2002). WHITE COLLAR-1, a circadian bluelight photoreceptor, binding to the frequency promoter. Science 297:815–819.

Galagan, J.E., Calvo, S.E., Borkovich, K.A., Selker, E.U., Read, N.D., Jaffe, D. et al. (2003).The genome sequence of the filamentous fungus Neurospora crassa. Nature422:859–868.

Grima, B., Chelot, E., Xia, R., Rouyer, F. (2004). Morning and evening peaks of activity relyon different clock neurons of the Drosophila brain. Nature 431:869–873.

He, Q., Cheng, P., Yang, Y., Wang, L., Gardner, K.H., Liu, Y. (2002). WHITE COLLAR-1, a DNAbinding transcription factor and light sensor. Science 297:840–843.

Heintzen, C., Loros, J.J., Dunlap, J.C. (2001). The PAS protein VIVID defines a clock-associ-ated feedback loop that represses light input, modulates gating, and regulates clockresetting. Cell 104:453–464.

Hoffmann, K. (1963). Zur Beziehung zwischen Phasenlage und Spontanfrequenz bei derendogenen Tagesperiodik. Z. Naturforschg. 18 b:154–157.

Honma, K., von Goetz, C., Aschoff, J. (1983). Effects of restricted daily feeding on freerunning circadian rhythms in rats. Physiol. Behav. 30:905–913.

Johnson, C.H. (2001). Endogenous timekeepers in photosynthetic organisms. Annu. Rev.Physiol. 63: 695–728.

Konopka, R., Benzer, S. (1971). Clock mutants of Drosophila melanogaster. Proc. Natl. Acad.Sci. USA. 68: 2112–2116.

Lee, K., Loros, J.J., Dunlap, J.C. (2000). Interconnected feedback loops in the Neurosporacircadian system. Science 289:107–110.

Linden, H., Macino, G. (1997). White collar 2, a partner in blue-light signal transductioncontrolling expression of light-regulated genes in Neurospora crassa. EMBO J.16:98–109.

Liu, Y. (2003). Molecular mechanisms of entrainment in the Neurospora circadian clock. J.Biol. Rhythms 18:195–205.

Liu, Y., Merrow, M., Loros, J.L., Dunlap, J.C. (1998). How temperature changes reset a circa-dian oscillator. Science 281:825–829.

Loros, J.J., Feldman, J.F. (1986). Loss of temperature compensation of circadian periodlength in the frq-9 mutant of Neurospora crassa. J. Biol. Rhythms 1:187–198.

McClung, C.R., Fox, B.A., Dunlap, J.C. (1989). The Neurospora clock gene frequency shares asequence element with the Drosophila clock gene period. Nature 339:558–562.

Merrow, M., Roenneberg, T. (2005). Enhanced phenotyping of complex traits with a circa-dian clock model. Meth. Enzymol. 393:251–265.

Merrow, M., Brunner, M., Roenneberg, T. (1999). Assignment of circadian function for theNeurospora clock gene frequency. Nature 399:584–586.

ENTRAINMENT

107

Merrow, M., Franchi, L., Dragovic, Z., Goerl, M., Johnson, J., Brunner, M., Macino, G.,Roenneberg, T. (2001). Circadian regulation of the light input pathway in Neurosporacrassa. EMBO J. 20:307–315.

Mrosovsky, N., Hattar, S. (2005). Diurnal mice (Mus musculus) and other examples oftemporal niche switching. J. Comp. Physiol. A Neuroethol. Sens. Behav. Physiol.191:1011–1024.

Pittendrigh, C.S., Bruce, V.G., Rosensweig, N.S., Rubin, M.L. (1959). Growth patterns inNeurospora crassa. Nature 184:169–170.

Pittendrigh, C.S., Daan, S. (1976). A functional analysis of circadian pacemakers innocturnal rodents. IV. Entrainment: Pacemaker as clock. J. Comp. Physiol. A106:291–331.

Pregueiro, A., Price-Lloyd, N., Bell-Pedersen, D., Heintzen, C., Loros, J.J., Dunlap, J.C.(2005). Assignment of an essential role for the Neurospora frequency gene in circadianentrainment to temperature cycles. Proc. Natl. Acad. Sci. USA. 102:2210–2215.

Roden, L., Song, H., Jackson, S., Morris, K., Carre, I.A. (2002). Floral responses to photope-riod are correlated with timing of rhythmic expression relative to dawn and dusk inArabidopsis. Proc. Natl. Acad. Sci. USA. 99:13313–13318.

Roenneberg, T., Merrow, M. (1998). Molecular circadian oscillators—an alternative hypoth-esis. J. Biol. Rhythms 13:167–179.

Roenneberg, T., Merrow, M. (1999). Circadian clocks and metabolism. J. Biol. Rhythms14:449–459.

Roenneberg, T., Merrow, M. (2002a). Life before the clock—modeling circadian evolution. J.Biol. Rhythms 17:495–505.

Roenneberg, T., Merrow, M. (2002b). “What watch? — such much!” — complexity andevolution of circadian clocks. Cell Tissue Res. 309:3–9.

Roenneberg, T., Taylor, W. (2000). Automated recordings of bioluminescence with specialreference to the analysis of circadian rhythms. Meth. Enzymol. 305:104–119.

Roenneberg, T., Merrow, M., Eisensamer, B. (1998). Cellular mechanisms of circadiansystems. Zoology 100:273–286.

Roenneberg, T., Dragovic, Z., Merrow, M. (2005). Demasking biological oscillators:Properties and principles of entrainment exemplified by the Neurospora circadian clock.Proc. Natl. Acad. Sci. USA. 102:7742–7747.

Sargent, M.L., Briggs, W.R., Woodward, D.O. (1956). Circadian nature of a rhythmexpressed by an invertaseless strain of Neurospora crassa. Plant Physiol. 41:1343–1349.

Schafmeier, T., Haase, A., Kaldi, K., Scholz, J., Fuchs, M., Brunner, M. (2005). Trans-criptional feedback of Neurospora circadian clock gene by phosphorylation-dependentinactivation of its transcription factor. Cell 122:235–246.

Schwerdtfeger, C., Linden, H. (2003). VIVID is a flavoprotein and serves as a fungal bluelight photoreceptor for photoadaptation. EMBO J. 22:4846–4855.

Shrode, L.B., Lewis, Z.A., White, L.D., Bell-Pedersen, D., Ebbole, D.J. (2001). vvd is requiredfor light adaptation of conidiation-specific genes of Neurospora crassa, but not circadianconidiation. Fung. Gen. Biol. 32:169–181.

Stoleru, D., Peng, Y., Agosto, J., Rosbash, M. (2004). Coupled oscillators control morningand evening locomotor behaviour of Drosophila. Nature 431:862–868.

Tan, Y., Merrow, M., Roenneberg, T. (2004a). Photoperiodism in Neurospora crassa. J. Biol.Rhythms 19: 135–143.

Tan, Y., Dragovic, Z., Roenneberg, T., Merrow, M. (2004b). Entrainment of the circadianclock: translational and post-translational control as key elements. Curr. Biol.14:433–438.

Welsh, D.K., Logothetis, D.E., Meister, M., Reppert, S.M. (1995). Individual neurons dissoci-ated from rat suprachiasmatic nucleus express independently phased circadian firingrhythms. Neuron 14:697–706.

Circadian entrainment:the rules of daily synchronizationof Neurospora crassa in temperature cycles

C. Madeti, T, Radic, T. Roenneberg and M. Merrow

Manuscript

7CHAPTER

ABSTRACT

One hypothesis concerning the circadian clock is that it serves to help organisms to antici-pate daily changes in their environment and to adjust to them. That this system is innatehas been proven by the fact that circadian rhythms continue unabated when entrainingstimuli (zeitgebers) are absent, i.e. they free-run with a period close to (but typically shorteror longer than) 24 hours. In natural conditions, however, zeitgebers coax a system to adjustto a 24h-rhythm, leading to stable entrainment. Zeitgebers include light, which is generallyconsidered the strongest of the signals for the circadian system, but also many other factors(that likely oscillate as a consequence of the light cycle), e.g. temperature, humidity, foodavailability, etc. The mechanisms generating rhythmicity are cell-based and rely on acomplex interplay of genes, their products, as well as posttranslational modifications. Clockgenes, e.g. frequency (frq) in the filamentous fungus Neurospora crassa, were identifiedbecause polymorphisms or mutations in their coding sequences or regulatory regions resultin a change of the free-running period. These mutations also change entrainment character-istics.Entrainment protocols usually utilize light as a zeitgeber. As a zeitgeber, it is easy to admin-ister and control, literally requiring a switch on or off. However, temperature cycles alsolead to stable entrainment, even though the free running period of a circadian system istemperature compensated. Even frq-null-strains entrain systematically, which has been usedto hypothesize a multi-oscillator-system in N. crassa. Here, we have assayed entrainment bytemperature in Neurospora using combinations of several endogenous and exogenousperiods with the goal of further elucidating the rules of entrainment. We aim to describe the‚entire surface’ of entrainment with temperature. Our data complements a similar experi-ment using light as a zeitgeber.

.

Introduction

Almost – if not - all living things on earth have it: a circadian clock, i.e. an innatesystem which predicts and adjusts an organism to daily, reccurring changes in theenvironment. Circadian clocks characteristically display oscillations with a periodclose to 24h, i.e. about a day (in Latin ‚circa diem’). In nature these oscillationsare synchronized with the help of zeitgebers (German for ‚time-giver’), includinglight, temperature, and food availability to yield a stably entrained phase.Circadian rhythms continue even without entraining stimuli, i.e. they free-run, aproperty which has been described already almost 300 years ago (De Mairan,1729). This self-sustainment was the basis on which it was proposed that thecircadian system is innate (Bünning, 1932). Since the early 1970’s clock geneshave been described in almost all phyla from Drosophila (Konopka & Benzer,1971) and Neurospora crassa (Feldman & Hoyle, 1973) to humans (Jones et al,1999b). By now we know that the circadian system is regulated by the interplayof many genes and their gene products (Roenneberg & Merrow, 2003;Shimomura et al, 2001). The machinery by which daily rhythms are generatedon a cellular level is characterized by complicated transcriptional-translationalautoregulatory feedback loops (Hardin et al, 1990), through which gene prod-ucts negatively feedback onto their own transcription, as well as positive feed-back loops, auxiliary loops and further modifications, e.g. by kinases (Roenne-berg & Merrow, 2005).

On the molecular level it is generally held that circadian clocks are reset bylight or other environmental signals via a change in the level of a key clockcomponent (Liu, 2003). In the filamentous fungus Neurospora crassa forexample, frq transcription is rapidly induced by light (Crosthwaite et al, 1995b),whereas in Drosophila, rapid degradation of the TIMELESS (TIM) protein by lightoccurs as part of entrainment (Young, 1998). The photoreceptor that is used forinduction of frq is WC-1 (WHITE-COLLAR 1), a blue light photoreceptor. Togetherwith WC-2, it forms a complex (the WHITE COLLAR COMPLEX or WCC) that is atranscriptional activator for light-inducible genes, such as frq (Froehlich et al,2002). The WCC exists in a small or large form, whereas in the dark the smallcomplex might be the one to activate frq-expression, while the large WCC leadsto transcriptional activation of frq in the light (Froehlich et al, 2002). Otherphotoreceptors and photoreceptor candidates exist in Neurospora (see introduc-tion of this thesis), some of which have been connected to the circadian clock,like VIVID (VVD), and others whose involvement has not yet been demonstrated(i.e. CRYPTOCHROME, PHYTOCHROME-1, PHYTOCHROME-2, NOVEL-OPSIN-1,OPSIN-RELATED-PROTEIN-1).

CHAPTER 7

110

The entrainment of Neurospora crassa using light as a zeitgeberIn nature, light and temperature changes are tightly linked, whereas in the labo-ratory they are generally separated when used as stimulus or zeitgeber for circa-dian experiments. As with most circadian systems, the zeitgeber response charac-teristics of Neurospora crassa were first established using light. When subjected toa single light pulse, the onset of conidiation (commonly used as the phasemarker for rhythmic behaviour in Neurospora) showed a systematic phase changein the following cycles, revealing a phase response curve typical of strong reset-ting (Crosthwaite et al, 1997b; Crosthwaite et al, 1995b; Dharmananda, 1980;Liu, 2003). That light can furthermore lead to highly systematic entrainment indifferent photoperiods has been shown by our laboratory a few years ago: with acycle length (T) of 24 hours, the laboratory standard strain entrains with conidia-tion beginning around midnight, when long, short and moderate photoperiodsare used (Tan et al, 2004). This occurs in conditions of both bright and dim light(3000 or 30nE). As a follow-up, a multi-dimensional protocol was applied,whereby T and the photoperiod within a cycle were varied. As a third dimensionthe inherent free-running period (hereafter shown as τ, ‘tau’) was also variedwithin these protocols. The period was varied by using wild type Neurospora andtwo frq-mutants, one with a short and one with a long τ. The results showsystematic rules of entrainment (Rémi, 2007, see Figures in the Results section).According to these rules, the relationship of T and τ is a determinant of phase,with a weaker zeitgeber (a shorter proportion of light in a cycle) leading to plas-ticity in the entrained phase. Furthermore, entrainment to either dawn or dusksignals may indicate intracellular ME-oscillators in this simple cellular circadiansystem (Rémi, 2007). We have taken advantage of this insight to add power toour functional genomics approaches to finding clock mutants (see chapter 7).

The entrainment of Neurospora crassa using temperature as a zeitgeberThe first systematic investigations into circadian responses to temperature as azeitgeber in Neurospora were performed by Van Gooch, who looked at straight-forward phase resetting of a free running rhythm (Gooch et al, 1994).Entrainment with temperature is also highly systematic, but will most probablyfollow different (molecular) rules. The Neurospora circadian clock can entrain totemperature pulses or cycles, whereas high temperatures correspond to a lightphase and low temperatures to darkness (Dharmananda, 1980; Francis &Sargent, 1979). In symmetrical T-cycles, where cold and warm phases are ofequal length, it was shown that the phase of entrainment is earlier in long andlater in short cycles (Merrow et al., 1999). Clock mutants displaying shorter orlonger free-running periods compared to the ‘wild type’ (bdA) entrain earlier orlater in a predictable and systematic manner (i.e., earlier or later). When temper-

TEMPERATURE CYCLES

111

ature cycles with a length of about half of the free-running rhythm (e.g. T=12hfor bdA with a τ of 22 hours) are applied, only one conidiation bout occurs inevery second cycle (e.g. for T=12 each 24 hours, a phenomenon called‘frequency demultiplication’), speaking for a robust oscillator mechanism(Merrow et al, 2006a).

Interestingly, entrainment with temperature has been used to infer a multi-oscillator clock system in Neurospora because entrainment occurs even in theabsence of one of the central clock components, the FRQ-protein (Merrow et al,1999a; Pregueiro et al, 2005; Roenneberg et al, 2005), or also without FRQ andWC-1 (Lakin-Thomas, 2006a). Thus, even in this simple cellular model organismcomplex mechanisms comparable to those described in photosynthetic organ-isms, flies and mammals (Aschoff, 1967; Johnson, 2001) exist as part of thecircadian clock mechanism.

Taken together, the experiments with light and temperature revealed thefollowing ‘rules’ for Neurospora (Merrow et al, 1999a; Onai & Nakashima, 1997;Roenneberg et al, 2005; Tan et al, 2004):● For entrainment with light, when cycle length (T) = τ, onset of conidiation

occurs at around midnight in all photoperiods● For entrainment with light, when T < τ in short photoperiods, onset of coni-

diation is dawn-locked● For entrainment with light, when T > τ in short photoperiods, onset of coni-

diation is dusk-locked● For entrainment with light, independent of T/τ in long photoperiods, onset of

conidiation is linked to midnight● For entrainment with temperature, when T≠ 24h, entrainment is highly

systematic, with conidiation occurring later in short cycles, and earlier in longcycles.

So far, not much is known about entrainment with both light and tempera-ture as zeitgebers. When light and temperature are applied together, theentrained phase reflects the strength of the zeitgeber amplitude. It follows thetemperature cycle if its amplitude is high, and it follows the light cycle if itsamplitude is lower (Liu 2003; Roenneberg & Merrow, 2001).

As already mentioned earlier, Neurospora has also been used to elucidatemolecular mechanisms of entrainment with light, leading to the proposal thattracking frq RNA might explain this phenomenon (Liu, 2003). However, it turnsout that FRQ protein correlates with conidiation – specifically, the degradation ofFRQ coincides with the onset of conidiation - whereas frq RNA is masked in light(Tan et al., 2004; Rémi J, 2007). This contrasts earlier results in temperaturecycles, where the frq RNA follows conidiation and is therefore not masked

CHAPTER 7

112

(Merrow et al, 1999). Furthermore, there are at least two possibilities to controlthe abundance of FRQ-protein in a temperature-dependent manner: at the levelof transcription (by regulated RNA-expression) and at the post-translational levelvia temperature-dependent splicing (Diernfellner et al, 2007; Diernfellner et al,2005; Liu et al, 1998). In this way the amounts of FRQ protein are regulateddepending on temperature signals (Liu et al, 1998). Also, as already pointed outbefore, frq RNA is clearly not driven as it is in light cycles (Tan et al, 2004; RémiJ, 2007). In short cycles it appears later and in long cycles it can be observedearly, suggesting that its mechanism does not work like an ‘hour glass’, but ratherlike an entrained system (Merrow et al, 1999).

To address the question of whether Neurospora shows a systematic change inphase angles (i.e. does it show entrainment or drivenness?) in temperature cyclesand whether these are different from entrainment with light, a series of system-atic temperature cycles were executed. Comparable to the multi-dimensionallight experiments mentioned before (Rémi, 2007), here three dimensions werealso varied (therefore the experiments were termed a ‘surface’), namely cyclelength, thermoperiods and τ. More precisely, the wild type strain (bd) and twofrq-allelic mutants, one displaying a shorter and the other a longer τ than bd,were assessed in thermoperiods from 16 to 84% in cycles with a length of 16, 22or 26 hours.

Material and methods

strainsAll mutant strains used are mutants on the background of the standard labora-tory strain ‚band’ (bdA, FGSC #1858). bd carries a mutation in ras-1 whichconfers a slower growth rate compared to wild type strains and a clear expressionof the circadian conidiation rhythm due to a reduced sensitivity to reactiveoxygen species (ROS) which accumulate in race tubes over the course of anexperiment (Belden et al, 2007). bd has a t of ca. 22 hours in complete darknessand 25ºC.

The mutant strains included in the temperature surface experiments were: ● bdA (FGSC #1858)● bd- frq1A (hereafter called frq1), a short period mutant (τ = 16.5h) and● bd- frq7A (hereafter called frq7), a long period mutant (τ = 29h) which has

lost its temperature-compensation. Both mutants are the results of singlepoint mutations (G to A transition) in the frq gene (Merrow & Dunlap, 1994)

TEMPERATURE CYCLES

113

Growth conditions (media, light)Temperature cycles were created in custom-made incubators, which circulatewarm or cold water in two waterbaths alternately, thereby ensuring 100%humidity and gradual temperature transitions. For example, to increase thetemperature from 22 to 27ºC, 90% of the warm temperature was attained in 50min (in a room with constant 18ºC). To cool down from 27 to 22°C the systemrequires 110 min for 90% of the change. All experiments were conducted incomplete darkness.

All strains were assassed on ‘race tubes’. Race tubes are hollow glass tubes(40 cm long, 12 mm diameter) with both ends bent up by 45°. A set of six repli-cate race tubes each were filled with 10 ml of molten race tube media (1X Vogel’ssolution (Vogel, 1956), 0.5% Arginine, 10 µl/100ml Biotin, 2% Agar) with either0% or 0.3% glucose (as indicated). One of these ‘six packs’ was used for eachstrain in each condition. The inoculated tubes were placed in continuous light atroom temperature until germination was observed. Then, the strains weresynchronized by the release from constant light into the experimental conditions.The growth front of the growing mycelia was marked directly on the glass tubebefore the tubes were transferred into the temperature cycles. Marking wasperformed at least every other day throughout the experiment. Additionally, theexact time of marking was noted in a mark time protocol). Since the growth rateis supposedly constant, period and phase can be calculated from the position andtiming of the conidial bands relative to one another and relative to the dailymarked growth fronts.

When the growth front reached the end of the agar, the tubes are scannedand saved as a PICT file and further analyzed with the Chrono Program, version6.4m (Roenneberg and Taylor, 2000).

All the strains were inoculated as three replicates on medium without glucoseor 0.3% glucose (adding a small amount of sugar increases the conidial densityand could give therefore more stable results in some of the strains).

Entrainment conditionsTemperature cycle lengths of 16h, 22h, and 26h were chosen for this surfaceprotocol. Various proportions of the warm phase in the cycle (thermoperiods)were varied, as was the inherent τ. The respective thermoperiods applied in eachcycle were 16, 25, 33, 40, 50, 60, 67, 75 and 84 %. Each cycle was divided intotwo phases, with (i) a warm phase corresponding to 27 °C and (ii) a cold phasecorresponding to 22 °C. Each cycle started with the cold phase once the race tubeswere put into the water baths. τ was varied by utilizing three different strains ineach of the cycles: bd (τ in DD/25ºC = 22h) frq7 (τ = 29h) and frq1 (τ = 16.5h).All in all, for each strain 27 conditions were applied (see Table 7.1).

CHAPTER 7

114

AnalysisUsing the Chrono Program (Version 6.4m, (Roenneberg & Taylor, 2000)), theonset of conidiation (Φon, ‘phi on’ relative to warm-off) for each strain and thecorresponding standard error of the mean (SEM) were calculated. Then, for all ofthe strains and for all cycle lengths, the onset of conidiation referring to themiddle of the cold phase was calculated. This phase marker is comparable tomidnight was chosen in order to be able to compare a) the cycles with each otherand b) the results from entrainment by temperature cycles to entrainment bylight cycles (established by J. Rémi in (Rémi, 2007))

Results

Previously, we demonstrated highly systematic entrainment using light (Rémi J,2007). The circadian surface with light was described with experiments thatvaried exogenous period (T), endogenous period (t; by utilizing frq-allelicmutants with differing FRPs) and photoperiod. A similar experiment wasdescribed once, wherein symmetrical T-cycles were used to probe the inductionof diapause via photoperiods in the European corn borer (Beck SD, 1962). In ourstudies, the proportion of the zeitgeber is also varied.

TEMPERATURE CYCLES

115

Table 7.1 The entrainment conditions used for the circadian surface with temperature. Theleft row shows the percentage of warm phase in each cycle from 16% on top to 84% at thebottom, the proportion of cold phase is indicated as the second number (from 84% on thetop to 16% at the bottom). The columns indicate from left to right the three strains used(top row), frq1, bd and frq7 in three different cycle lengths (16, 22 and 26 hours), each. Theconditions applied per strain are indicated by 1–27, whereas same numbers mean the samecondition.

τ = 16.5 h (frq1) τ = 22 h (bd) τ = 29 h (frq7)

% 16h 22h 26h 16h 22h 26h 16h 22h 26h

16/84 1 10 19 1 10 19 1 10 1925/75 2 11 20 2 11 20 2 11 2033/67 3 12 21 3 12 21 3 12 2140/60 4 13 22 4 13 22 4 13 2250/50 5 14 23 5 14 23 5 14 2360/40 6 15 24 6 15 24 6 15 2467/33 7 16 25 7 16 25 7 16 2575/25 8 17 26 8 17 26 8 17 2684/16 9 18 27 9 18 27 9 18 27

The results are shown for each of the strains separately. Each cycle was dividedinto two phases, warm phase (27 °C, shown as light grey background in thefollowing figures) and cold phase (22 °C, shown as dark grey triangle in thefigures).

In the bd strain (Fig 7.1), in T=22h (where T equals τ) entrainment is lockedto midnight in temperature cycles. In T = 16h (where T is shorter than τ) inshort thermoperiods, the onset of conidiation is linked to dawn. The onset ofconidiation shifts thereafter in long thermoperiods into the middle of the warmphase. In the case where T is longer than τ, the onset of conidiation is equallylined up with the middle of the cold phase, but it is phase advanced relative tothe shorter, T=22h cycle.

In frq1 (Fig 7.2), in T = 16h, where T equals τ, entrainment is locked tomidnight and midcold. In T = 22h and 26h, where T is longer than τ, firstly ashift towards an earlier phase can be seen with the onset of conidiation consis-tently lined up with the middle of the cold phase.

In frq7 (Fig 7.3), in T=16h, where T is about half (56%) of τ, frq7 showsfrequency demultiplication with one conidiation bout per two cycles (32 hours,therefore not included in Figure 7.6). In T= 26h, where T equals τ, in short ther-moperiods (less than 50% warm phase) the onset of conidiation is locked to themiddle of the cold phase. In T=22, where T is shorter than τ, the onset of conidi-ation in frq7 seems (except for one point) equally lined up with the middle of thecold phase.

CHAPTER 7

116

27°C

50%

22°C

ther

mop

erio

d

T=16 T=22 T=26

Figure 7.1 The surface with temperature for the ‘wild type’, bd (τ = 22h). Onset of conidia-tion related to the middle of the cold phase, shown as dots in T-cycles with a length of 16(left graph), 22 (middle graph) and 26 (right graph) hours. Each cycle is divided into twophases, warm phase (corresponding to 27 °C, shown as white background in the figures)and cold phase (22 °C, shown as a grey triangle in the figures), the experiments are shownas a continuum from top (100% cold, 22ºC) to bottom (100% warm, 27ºC), so the fifthpoint in each graph is the experimental set-up with 50% of warm and cold phase. Standarderrors are shown as horizontal bars smaller than most points).

Discussion

In conclusion, the circadian surface with temperature yielded the following‘rules’: ● When T = τ, the onset of conidiation in bd and frq1 entrains parallel to the

middle of the cold phase. In frq7, entrainment relates to the middle of thecold phase only in cycles with short thermoperiods (less than 50% of warmphase in a cycle) and is linked to the warm-cold transition (dusk) in longthermoperiods (less than 50% of warm phase in a cycle) .

● If T> τ, onsets of conidiation in bd and frq1 are locked parallel to the middleof the cold phase, but in a longer T the phase becomes earlier

TEMPERATURE CYCLES

117

27°C

50%

22°C

ther

mop

erio

d

T=16 T=22 T=26

Figure 7.2 The surface with temperature for frq1, the short period mutant (τ = 16.5h).Onset of conidiation related to the middle of the cold phase, shown as dots in T-cycles witha length of 16 (left graph), 22 (middle graph) and 26 (right graph) hours. Standard errorsare shown as horizontal bars (smaller than most points).

27°C

50%

22°C

ther

mop

erio

d

T=16 T=22 T=26

Figure 7.3 The surface with temperature for frq7, the long period mutant (τ = 29h). Onsetof conidiation related to the middle of the cold phase, shown as dots in T-cycles with alength of 22 (middle graph) and 26 (right graph) hours. The results for T = 16h are notincluded, since frq7 shows frequency-multiplication here. Standard errors are shown as hori-zontal bars (smaller than most points).

● If T< τ, the onset of conidiation appears before the cold-warm transition inshort thermoperiods and switches to after the cold-warm transition when theproportion of warm phase is more than 40%.

Comparing entrainment with light and temperature: two surfacesThe ‘wild type’ bd (τ = 22h) in a light and temperature surface:

Figure 7.4 compares the ‘wild type’ strain bd (τ= 22h) in light and tempera-ture surfaces. In T=22h, where T equals τ, entrainment is locked to midnight inboth light and temperature cycles. In T = 16h, where T is shorter than τ, and inshort photo-/thermo-periods, the onset of conidiation is linked to dawn. In thetemperature surface, as in the light surface, the onset of conidiation shows aphase shift in long photoperiods. In all T=16h experiments, the phase is later inthe temperature cycles than in light cycles. In the case, where T is longer than τ,entrained phase is dusk-linked in short photoperiods and a midnight-linked inlong photoperiods. The situation is not so clear for the temperature surface.Here, the onset of conidiation is generally lined up parallel to midnight/midday.Thus, the phase of entrainment is similar between light and temperature cyclesexcept in short T cycles.

CHAPTER 7

118

LL

50%

DD

phot

oper

iod

T=16 T=22 T=26

27°C

50%

22°C

ther

mop

erio

d

Figure 7.4 The comparison of surfaces with light (lower sequence by Rémi J, 2007) andwith temperature (upper sequence) for bd.

The short period mutant frq1 (t = 16.5h) in a light and temperature surface

Figure 7.5 compares light and temperature surfaces for the short period mutant,frq1. In T = 16h, where T equals τ, entrainment parallels midnight when eitherlight or temperature is used as a zeitgeber. In T = 22h and 26h, where T islonger than τ, the results for the light surface show a progressive shift towards anearlier phase in T = 22h and in T = 26h. In T = 26h, a dusk-link in shortphotoperiods becomes a midnight-lock in long photoperiods. Although a similarshift towards an earlier phase can be seen in long cycles for the temperaturesurface, T=22h is similar to T= 26h. Here, the onset of conidiation is parallel tomid-cold in T = 22h and 26h. In all T’s, the entrained phase in light is later thanin temperature. Furthermore, the range of entrained phase is far greater in thelight surface than in the temperature surface for this strain.

The long-period mutant frq7 (τ = 27h) in a light and temperature surface:Figure 7.6 shows the comparison between light and temperature surfaces for thelong period mutant, frq7. In T= 26h, where T equals τ, the onset of conidiationparallels midnight, when entrained with light, but this occurs only in short ther-moperiods (less than 50% warm phase) in the temperature surface. In T=22h,

TEMPERATURE CYCLES

119

LL

50%

DD

phot

oper

iod

T=16 T=22 T=26

27°C

50%

22°C

ther

mop

erio

d

Figure 7.5 Comparison of surfaces with light (lower sequence, by Rémi J, 2007) and withtemperature (upper sequence) for frq1.

where T is shorter than τ, a clear dawn-link occurs in short photoperiods,whereas in long photoperiods entrainment is relative to midnight/midday. Incontrast to this observation, the onset of conidiation in frq7 in the temperaturesurface seems (except for one point) consistently lined up with mid-cold inT=22. In light cycles where T is approximately equal to τ, entrainment is onlymodestly later than in temperature cycles. In some of the short T cycles, entrain-ment is earlier in light cycles than in temperature cycles.

In summary, entrainment with light compared to temperature as a zeitgeber,shows some similarities in the basic properties. This trend is strongest in the ‘wildtype’ strain bd. Beyond the wild type strain, the entrained phases in T=22conform to circadian theory, in that phase of entrainment in temperature cyclesrelates to period length in DD (Roenneberg & Merrow, 2003). frq1 entrainsearlier than bd, and frq7 entrains later than bd (see Fig. 7.7). The patterns ofentrainment of these tau mutants vary considerably between the light andtemperature protocols (frq1 entrains consistently earlier, frq7 later in temperaturecyles). This is probably explained by their different light PRCs (frq1 PRC strongerthan bd, frq7 weaker than bd, Dharmananda), but until we have temperaturePRCs for the same strains, it is difficult to certify this.

CHAPTER 7

120

LL

50%

DD

phot

oper

iod

T=16 T=22 T=26

27°C

50%

22°C

ther

mop

erio

d

Figure 7.6 Comparison of surfaces with light (upper sequence, by Rémi J, 2007) and withtemperature (lower sequence) for frq7.

In the period mutants frq1 (short) and frq7 (long) the most striking differ-ences relative to entrainment in light are seen in short thermoperiods: whereas inthe corresponding photoperiods dawn- or dusk –linked phases occur in the lightcycles, in the short thermoperiods almost all points are lined up parallel tomidnight.

Given the complexity of the network between environment and the circadianclock it is obvious to think that a selective advantage might be conferred by an‘optimal’ interplay of both (or all) zeitgebers (Dunlap, 1999; Hotta et al, 2007;Young & Kay, 2001). That zeitgeber cycles exert evolutionary pressure has beenshown for example in cyanobacteria and plants. In cyanobacteria mutants withdifferent free running periods were kept in LD-cycles of different lengths.Consistently, the strain with the τ closest to the cycle length was the one tooutgrow the others with the non-matching period (Ouyang et al, 1998). Thesame phenomenon, termed circadian resonance, was found in Arabidopsis, wherestrains with periods matching the environmental rhythm had enhanced fitnesstraits (measured in biomass, photosynthesis and competitive advantage) (Doddet al, 2005).

The results of the surface with light (Rémi J, 2007, see Results section) andtemperature in Neurospora crassa could be interpreted also along this line, wherean entrainment linked to midnight should increase fitness because conidiation(and thus nuclear replication) occurs in the middle of the night (possibly togenerate enough carotenoid-containing conidia before sunrise, where UV-damage is likely).

TEMPERATURE CYCLES

121

LL

50%

DD

phot

oper

iod

bdfrq1

frq7

Figure 7.7 Thermoperiods (27/22ºC, with a cycle length T = 22h) with increasing warmproportions from top (100% cold, in grey) to bottom (100% warm, in white) of bd, frq1 andfrq7 in comparison. Phase of onset of conidiation (in hours) referring to the middle of thecold phase is chosen. Standard errors are shown as horizontal bars

Little is known about the signaling pathways involved in transducing thermalsignals to the clock and also about the molecular machinery therein. A simplesystem like Neurospora crassa may be ideal to study the complexity of entrain-ment according to multiple zeitgebers, as is typical in a natural situation.

References

CHAPTER 7

122

Aschoff J (1967) Human circadian rhythms in activity, body temperature and other func-tions. Life Science and Space Research: 159-173

Beck, SD (1962) Photoperiodic induction of diapause in an insect. Biol Bull 122(1): 1-12Belden WJ, Larrondo LF, Froehlich AC, Shi M, Chen CH, Loros JJ, Dunlap JC (2007) The

band mutation in Neurospora crassa is a dominant allele of ras-1 implicating RASsignaling in circadian output. Genes Dev 21(12): 1494-1505

Bruce VG, Weight F, Pittendrigh CS (1960) Resetting the sporulation rhythm in Piloboluswith short light flashes of high intensity. Science 131: 728-730

Bünning E (1932) Über die Erblichkeit der Tagesperiodizität bei den Phaseolus Blättern. Jbwiss Bot 81: 411-418

Crosthwaite SK, Dunlap JC, Loros JJ (1997) Neurospora wc-1 and wc-2: transcription,photoresponses, and the origins of circadian rhythmicity. Science 276(5313): 763-769

Crosthwaite SK, Loros JJ, Dunlap JC (1995) Light-induced resetting of a circadian clock ismediated by a rapid increase in frequency transcript. Cell 81(7): 1003-1012

De Mairan JJdO (1729) Observation botanique. Histoir de l'Academie Royale des Science: 35-36Dharmananda S (1980) Studies of the circadian clock of Neurospora crassa: light-induced

phase shifting. University of California, Santa Cruz, Santa Cruz, CA, Diernfellner A, Colot HV, Dintsis O, Loros JJ, Dunlap JC, Brunner M (2007) Long and short

isoforms of Neurospora clock protein FRQ support temperature-compensated circadianrhythms. FEBS Lett 581(30): 5759-5764

Diernfellner AC, Schafmeier T, Merrow MW, Brunner M (2005) Molecular mechanism oftemperature sensing by the circadian clock of Neurospora crassa. Genes Dev 19(17):1968-1973

Dodd AN, Salathia N, Hall A, Kevei E, Toth R, Nagy F, Hibberd JM, Millar AJ, Webb AA(2005) Plant circadian clocks increase photosynthesis, growth, survival, and competitiveadvantage. Science 309(5734): 630-633

Dunlap JC (1999) Molecular bases for circadian clocks. Cell 96: 271-290Francis CD, Sargent ML (1979) Effects of Temperature Perturbations on Circadian

Conidiation in Neurospora. Plant Physiol 64(6): 1000-1004Froehlich AC, Liu Y, Loros JJ, Dunlap JC (2002) White Collar-1, a circadian blue light

photoreceptor, binding to the frequency promoter. Science 297(5582): 815-819Gooch VD, Wehseler RA, Gross CG (1994) Temperature effects on the resetting of the phase

of the Neurospora circadian rhythm. J Biol Rhythms 9: 83-94Hardin PE, Hall JC, Rosbash M (1990) Feedback of the Drosophila period gene product on

circadian cycling of its messenger RNA levels. Nature 343: 536-540Hotta CT, Gardner MJ, Hubbard KE, Baek SJ, Dalchau N, Suhita D, Dodd AN, Webb AA

(2007) Modulation of environmental responses of plants by circadian clocks. Plant CellEnviron 30(3): 333-349

Johnson CH (2001) Endogenous timekeepers in photosynthetic organisms. Annu Rev Physiol63: 695-728

Jones CR, Campbell SS, Zone SE, Cooper F, DeSano A, Murphy PJ, Jones B, Czajkowski L,Ptacek LJ (1999) Familial advanced sleep-phase syndrome: A short-period circadianrhythm variant in humans. Nat Med 5(9): 1062-1065

Lakin-Thomas PL (2006) Circadian clock genes frequency and white collar-1 are not essen-tial for entrainment to temperature cycles in Neurospora crassa. Proc Natl Acad Sci USA103(12): 4469-4474

TEMPERATURE CYCLES

123

Liu Y (2003) Molecular mechanisms of entrainment in the Neurospora circadian clock. J BiolRhythms 18(3): 195-205

Liu Y, Merrow M, Loros JJ, Dunlap JC (1998) How temperature changes reset a circadianoscillator. Science 281(5378): 825-829

Merrow M, Boesl C, Ricken J, Messerschmitt M, Goedel M, Roenneberg T (2006)Entrainment of the Neurospora circadian clock. Chronobiol Int 23(1-2): 71-80

Merrow M, Brunner M, Roenneberg T (1999) Assignment of circadian function for theNeurospora clock gene frequency. Nature 399(6736): 584-586

Merrow MW, Dunlap JC (1994) Intergeneric complementation of a circadian rhythmicitydefect: phylogenetic conservation of structure and function of the clock gene frequency.EMBO J 13: 2257-2266

Onai K, Nakashima H (1997) Mutation of the cys-9 gene, which encodes thioredoxin reduc-tase, affects the circadian conidiation rhythm in Neurospora crassa. Genetics 146(1):101-110

Ouyang Y, Andersson CR, Kondo T, Golden SS, Johnson CH (1998) Resonating circadianclocks enhance fitness in cyanobacteria. Proc Natl Acad Sci USA 95(15): 8660-8664

Pregueiro AM, Price-Lloyd N, Bell-Pedersen D, Heintzen C, Loros JJ, Dunlap JC (2005)Assignment of an essential role for the Neurospora frequency gene in circadian entrain-ment to temperature cycles. Proc Natl Acad Sci USA 102(6): 2210-2215

Rémi J (2007) The circadian surface of Neurospora crassa- From physiology to molecularmechanisms Doctoral Thesis University of Munich, LMU

Roenneberg T, Dragovic Z, Merrow M (2005) Demasking biological oscillators: propertiesand principles of entrainment exemplified by the Neurospora circadian clock. Proc NatlAcad Sci USA 102(21): 7742-7747

Roenneberg T, Merrow M (2003) The network of time: understanding the molecular circa-dian system. Curr Biol 13(5): R198-207

Roenneberg T, Merrow M (2005) Circadian clocks - the fall and rise of physiology. Nat RevMol Cell Biol 6(12): 965-971

Roenneberg T, Taylor W (2000) Automated recordings of bioluminescence with special refer-ence to the analysis of circadian rhythms. Methods Enzymol 305: 104-119

Shimomura K, Low-Zeddies SS, King DP, Steeves TD, Whiteley A, Kushla J, Zemenides PD,Lin A, Vitaterna MH, Churchill GA, Takahashi JS (2001) Genome-wide epistatic interac-tion analysis reveals complex genetic determinants of circadian behavior in mice.Genome Res 11(6): 959-980

Tan Y, Merrow M, Roenneberg T (2004) Photoperiodism in Neurospora crassa. J BiolRhythms 19(2): 135-143

Vogel HJ (1956) A convenient growth medium for Neurospora (Medium N). Microbiol GenetBull 13: 42-43

Young MW (1998) The molecular control of circadian behavioral rhythms and their entrain-ment in Drosophila. Annu Rev Biochem 67: 135-152

Young MW, Kay SA (2001) Time zones: a comparative genetics of circadian clocks. NatureReviews Genetics 2: 702-715

Time bandit: the band mutant holds up the wild type

M. Merrow, C. Madeti and T. Roenneberg

Manuscript

8CHAPTER

ABSTRACT

One of the fundamental properties of cells is temporal regulation by the circadian clock.Molecular clock components have been identified in animals, plants, fungi and prokaryotes,revealing transcriptional negative feedback loops. These components must necessarily betightly linked with input pathways (e.g., transducing light signals) to communicate informa-tion about environmental time-of-day. Neurospora crassa is a classic model system forresearch on the circadian clock and its molecular mechanisms. Practically all of the work onthe Neurospora clock utilized a mutant (bd) which was discovered for its clear daily growthpatterns. The location of bd mutant has recently been identified by Belden et al. to lie in theras-1 gene. Because the mutation is also known to modify light regulated gene expression, itcould potentially alter the circadian phenotype. Belden et al. determined that it has nodiscernable effect on circadian oscillations in constant darkness – despite decreasing growthrate and increasing sporulation.

.

Circadian clocks evolved such that they regulate cellular metabolism to functionoptimally within the day and to anticipate daily changes in environmental quali-ties (e.g., light, temperatures or nutrients). A post hoc analytical study recentlyshowed the scale of this regulation in proposing that the expression of practicallyall genes is modulated by the circadian clock in eukaryotic cells (Ptitsyn et al.2007), as had already been shown for the prokaryote, Synechococcus (Liu et al.1995). Whereas an intact clock confers an adaptive advantage (Yan et al. 1998),defective clocks may lead, for example, to increased incidences of certain cancers(Fu et al. 2002). Living against the clock, by experimental induction of chronicjetlag in mice (Davidson et al. 2006) or in humans, when social schedules areincompatible with individual circadian time (Wittmann et al. 2006), may chal-lenge longevity and health. Chronobiologists are, thus, racing to understand howthe circadian clock – a fundamental characteristic of all cells - is working on themolecular level and how it is regulating metabolism.

The Neurospora circadian clock

Research into circadian mechanisms – like in other fields – utilizes preferredmodel systems for both historical and methodological reasons. Mammals havelong been represented by hamsters, due to their remarkably precise wheel-running rhythms (Davis 1980). In the genetic era of clock research, mice havewon the competition over hamsters (Vitaterna et al. 1994). As with other geneticand behavioral breakthroughs, extraordinary insights concerning clocks havecome from Drosophila (Konopka and Benzer 1971; Hardin et al. 1990). Circadianrhythms in plants were initially studied in many different species, ranging frombeans (Bünning and Moser 1973) to Madagascan shrubs (Engelmann et al.1961), but are now predominantly investigated in Arabidopsis (Park et al. 1999).Much of our early knowledge about cellular clocks stems from studying photo-synthesizing unicellular algae, such as Gonyaulax (Hastings and Sweeney 1959),Euglena (Bruce and Pittendrigh 1958), or Chlamydomonas (Bruce 1972) but theeukaryotic unicells have now been overshadowed by the remarkable work on theprokaryote Synechococcus (Kondo et al. 1994; Nakajima et al. 2005).

The fungus Neurospora crassa was introduced to circadian research because ofits conspicuous daily growth patterns (Sargent et al. 1956; Pittendrigh et al.1959). Neurospora has many requisite advantages for an ideal model system,including rapid growth and a small, haploid genome, tractability for geneticstudies and it is completely safe to work with (non-pathogenic and no teeth). Ithas no brain to confuse pacemaker and organ-specific clocks, and defies theconfusion of sex and development; although it technically does develop and can

CHAPTER 8

126

have sex, the former results in a terminal tissue left behind a leading growthfront, and the latter is a relatively rare condition. This fungus is an excellentsystem combining genetic and biochemical approaches, and new tools are contin-uously being developed (a recent functional genomics project aims to makeknockouts of every annotated ORF, a task that will promote Neurospora to theupper echelon of model systems beyond circadian research (www.dartmouth.edu/~neurosporagenome/).

Neurospora has been used to establish many of the principles of molecularmechanisms of cellular clocks (Lakin-Thomas et al. 1990; Lakin-Thomas andBrody 2004; Brunner and Schafmeier 2006). Its circadian clock is measured byfollowing outputs including the production of asexual spores (the conspicuousdaily growth patterns) or gene expression, protein levels or enzyme activity. As inall organisms – from Synechococcus to humans, Neurospora’s circadian system andits outputs are synchronized (entrained) to precisely 24 h by environmentalcycles (e.g., light or temperature). The clock was shaped through evolution bylight-dark cycles, which are the most reliable environmental signal for synchro-nizing daily rhythms. It is the clock – in combination with environmental signals– that makes (most of) us diurnal and most rodents nocturnal. Although entrain-ment is the natural state of circadian clocks, they are often investigated experi-mentally using the common, yet remarkable phenotype of continuing their rhyth-micity in constant conditions (with no decrease in amplitude). These free-running rhythms often deviate from 24 h. In the case of Neurospora, rhythmssuch as spore production occur once per 22 h in constant darkness.

Molecular clock mechanisms – from the slant of Neurospora

A transcriptional negative feedback loop is responsible for a daily increase anddecrease in clock regulated gene expression. The components of this networkhave been determined with the toolbox of the modern geneticist, showing theproducts of the frequency (frq) and the white collar genes (wc-1 and wc-2) servingrepression and activating roles, respectively (Dunlap and Loros 2004). Notably, inaddition to it's role as a transcription factor, WC-1 is a blue light photoreceptor(He et al. 2002). Thus, both light and clock information appear to go via WC-1on their way to gene regulation. To complicate matters further, the negativeelement, FRQ, is also necessary for the clock to “see” light (Merrow et al. 1999)and, in addition, maintains levels of its own activator, WC-1. Such complicationsare typical of clock and similar networks and other circadian systems have simi-larly constructed feedback loops, but animals, plants, fungi and prokaryotespossess unique sets of clock genes (Young and Kay 2001).

TIME BANDIT

127

Another characteristic of most circadian systems is that they appear to beconstructed from multiple oscillators – even those in simple unicells (Roennebergand Morse 1993). This network quality was established for Neurospora in severalways. (i) Residual oscillations in the circadian range – as well as with longerperiods – can be recorded in clock gene mutants (Loros and Feldman 1986;Lakin-Thomas and Brody 2000; Dragovic et al. 2002). (ii) Mutants of character-ized clock genes (e.g., frq) can still be entrained (e.g., to temperature cycles)with circadian characteristics (Roenneberg et al. 2005). (iii) Circadian oscilla-tions of isolated RNAs can be measured in apparently arrhythmic clock mutants(Correa et al. 2003).

The bandit mutation

Neurospora is an ideal 'simple' system for studying the complex genetic trait ofcircadian rhythmicity. The molecular mechanisms responsible for circadianrhythm generation have – at least in part – been worked out, and we have aglimpse of how multiple oscillators come together to form a molecular clock-network. Yet, a shadow of doubt hangs over the system because virtually allcircadian experiments have been performed in a mutant background, called band(bd, referring to the enhanced banding in spore formation relative to the wildtype strain(s); Fig. 8.1). Although model systems for genetic research are oftenstunted versions of their wild type cousins, the bd mutation has been a particularconcern because it has major effects on the cells' biology, leading – in addition toenhanced sporial banding – to substantial decreases in growth rate (Sargent et al.1966). Furthermore, the bd mutation enhances transcription of some genes inresponse to light (Arpaia et al. 1993). This was especially important due to theinseparable roles of clock and light input in Neurospora. In order to lift theshadow, the gene responsible for the bd mutation had to be identified and themechanisms underlying its effects on banding and light expression elucidated.

The post-genomic era has now facilitated identification of the bd mutation,showing that it lies in the ras-1 gene (Belden et al. 2007). The RAS protein iswell characterized in other contexts and thus allows for experimentation todetermine how bd (now ras-1bd) effects Neurospora’s circadian system. RAS wasfirst identified as an oncogene (it was identified in viral isolates from rat sarcomatumors) (Diaz-Flores and Shannon 2007). Mutations in RAS are responsible for adaunting variety of cancers, making it an attractive target for developing anti-cancer therapies. RAS is bound to the cell membrane and catalyzes the conver-sion of GTP to GDP – it is a G protein. GTPbinding converts the protein to itsactive state which signals to downstream targets via hydrolysis to GDP. GTP

CHAPTER 8

128

binding is effected by a family of facilitating proteins, suggesting a network ofregulatory mechanisms. In theory, any one of these components (the G protein,the facilitators or the GTP, the acceptor of the signal) is regulated independently,thus lending specificity – also in the temporal domain – to a potent signal trans-duction mechanism. The location of RAS, in the cell membrane, suggests theirrole in sensing the extra-cellular environment (Neurospora tissue is a syncitiumreadily allowing inter-cellular signaling). Circadian clocks can be synchronized toregular changes of temperature (Merrow et al. 1999; Brown et al. 2002) or nutri-tion (Roenneberg and Rehman 1996), using these cues to reliably predict thetime of the environmental day.

RAS-1 is constitutively expressed in Neurospora under constant conditions,both in wildtype and in the ras-1bd mutant (Belden et al. 2007), unlike so manyother cellular components (e.g., (Correa et al. 2003)). Although this is a firstcrucial test, RAS-1 could still have circadian impact. Several key clock compo-nents, including WC-2 and casein kinase Ia, are expressed at constant levels(Crosthwaite et al. 1997; Görl et al. 2001). Alternatively, RAS-1 could use a clockregulated facilitator protein and become temporally regulated.

In another approach, the ras-1bd mutation was phenocopied using reducingagents and genetics showing that these manipulations also do not change thefree running period of spore formation. Thus, concerning free running rhyth-micity, there seems to be little if any effect on circadian timing. Despite this,when molecular components of the clock pathway were measured, most compo-nents (except for ras-1 itself) were found at unusual levels (see Table 8.1). For

TIME BANDIT

129

grow

th ra

te (

% r

el. t

o bd

)

200

90

150

bd

Figure 8.1 Daily spore formation by the bd mutant (top) and 3 wild type strains, 74- OR23-1A, FGSC #8802 and FGSC #8860. We have adapted the race tube assay (omitting glucosefrom the media) for improved banding of wild type isolates. These strains were chosen todemonstrate the variety of circadian phenotypes that are recovered from natural isolates.One of them is adequate, one shows no regular banding (evaluated either by eye or digi-tally) and one of these strains (8860) looks as though it bands more robustly than the bdstrain. Their growth rate relative to bd is indicated on the right. 74-OR23 was collected inthe U.S., and 8802 and 8860 are from India. (Thanks to D. Jacobson for supplying theIndian strains.)

vvd RNA (VVD is a light signaling modifier as well as yet another photoreceptor),levels are lower and later in the ras-1bd mutant. The RNAs of the clock genes frqand wc-1, and the output gene fluffy, all show higher levels. In light of the appar-ently normal free-running rhythms in darkness, there seems to be compensationon the level of RNA regulation within the circadian network. The results also maypoint to our ignorance on what role RNA regulation plays in circadian rhythmgeneration (Yang and Sehgal 2001).

Entrainment – the key property of the circadian clock

RAS-1bd apparently has no effects on free-running rhythms in Neurospora. Butthe circadian clock rarely, if ever, gets a chance to display its capacity to run freein nature – it is normally entrained which makes investigation of whether themutation affects entrainment crucial, especially since this mutant leads to excep-tionally high levels of light-induced gene expression (Arpaia et al. 1993).Although entrainment per se was not investigated by Belden et al., they did inves-tigate how the clock and the expression of its output genes respond to light. Theresults produced puzzles which still remain to be solved. As in darkness, wc-1RNA is expressed at elevated levels in the ras-1bd mutant compared to wild typein response to light, whereas the WC-1 protein is decreased. Given the normalcircadian rhythmicity in constant darkness – despite altered clock gene RNAexpression (see above) – these results clearly demonstrate that RNA levels arenot a good gauge for function. But, even regulation of the protein does not tell

CHAPTER 8

130

constant darkness 4h lightexposureRNA level Rhithmic?

first 12h only

first 12h only first day only

RNA wc-1

frq

fluffy

vvd

Protein WC-1

Table 8.1 RNA and protein levels in the ras-1bd strain, relative to wild type. See Belden et alfor methodological details; upward and downward arrows indicate increases and decreases,respectively, observed in the mutant strain relative to the wild type. Lack of rhythmicity isshown as a straight line, whereas circadian rhythms are indicated by the squiggle.

the full story, shifting the focus to post-translational modification as a criticalfactor in the circadian clock: as for many other transcription factors, WC-1activity depends on phosphorylation state and sub-cellular localization(Schafmeier et al. 2006). Its de-phosphorylated nuclear form is most effective asa transcriptional activator, so if this is unaffected in ras-1bd, then it could – quan-titatively - maintain its job in circadian regulation of frq transcription, even asoverall amounts have fallen.

The fact that reduction of a key component in the so-called core clock ofNeurospora does not have an effect on precise timing shows that we have to startthinking more about entire networks of molecular loops. Indeed, when wc-1 isexpressed constitutively from an inducible promoter, all three so-called centralclock proteins – WC-1, WC-2 and FRQ - are expressed at abnormally high levels,yet the free running period is perfectly normal (Cheng et al. 2001). Apparently,there is also compensation for clock protein levels among the central clockcomponents.

Although Belden et al. investigated light induced gene expression in ras-1bdversus wild type, a remaining conundrum is what the mutation means for entrain-ment. The process of entrainment also refers to differences between individuals.In a population there is typically a normal distribution of entrained phases(chronotypes), ranging from extreme early to extreme late types with all the restof the individuals somewhere in between. So if the bd mutation caused a shiftwithin this chronotype distribution, then it would be a clock mutant – even whenthe free running period was identical to wild type. Circadian clock theory (basedon both oscillator theory and wet experiments) associates phase of entrainment(chronotype) with individual differences in free running period. By this conven-tion, the ras-1bd mutant would be expected to have a phase like any wild typestrain with the same free running period. There are exceptions to this rule,reflecting the nature of the molecular clock as complex networks (e.g., see (Spoel-stra et al. 2004; Merrow et al. 2005) for discussion). Thus without more elaborateexperimentation on the effects of the bd mutation on the important state of thecircadian system, entrainment, the verdict of whether or not it is a clock gene isstill open. But even if it turns out that this mutant affects chronotype, the shadowof chasing a vigorously banding ghost can be swept aside. The new results showthat the story of the Neurospora clock does not have to be re-written because allthat remains is whether we have gained our current knowledge about this fasci-nating model system from a Neurospora lark or a Neurospora owl

TIME BANDIT

131

AcknowledgementsOur work is supported by the Dutch Science Foundation (NWO), the University ofGroningen, the Hersenstichting Nederland, the German Science Foundation (DFG), theEuropean Comission (EUCLOCK) and the Daimler-Benz Foundation (CLOCKWORK).

References

CHAPTER 8

132

Arpaia, G., Loros, J.J., Dunlap, J.C., Morelli, G., and Macino, G. 1993. The interplay of lightand the circadian clock. Plant Physiol 102: 1299-1305.

Belden, W.J., Larrondo, L.F., Froehlich, A.C., Shi, M., Chen, C.H., Loros, J.J., and Dunlap,J.C. 2007. The band mutation in Neurospora crassa is a dominant allele of ras-1 impli-cating RAS signaling in circadian output. Genes & development 21(12): 1494-1505.

Brown, S.A., Zumbrunn, G., Fleury-Olela, F., Preitner, N., and Schibler, U. 2002. Rhythms ofmammalian body temperature can sustain peripheral circadian clocks. Curr Biol 12:1574–1583.

Bruce, V. and Pittendrigh, C.S. 1958. Resetting the Euglena clock with a single light stim-ulus. Am Nat 92: 294-306.

Bruce, V.G. 1972. The biological clock in Chlamydomonas reinhardi. Genetics 70:537-548.Brunner, M. and Schafmeier, T. 2006. Transcriptional and post-transcriptional regulation of the

circadian clock of cyanobacteria and Neurospora. Genes & development 20(9): 1061-1074.Bünning, E. and Moser, I. 1973. Light induced phase shifts of circadian leaf movements of

Phaseolus: comparison with the effects of potassium and of ethyl alcohol. Proc Natl AcadSci USA 70: 3387-3390.

Cheng, P., Yang, Y., and Liu, Y. 2001. Interlocked feedback loops contribute to the robustnessof the Neurospora circadian clock. Proceedings of the National Academy of Sciences of theUnited States of America 98(13): 7408-7413.

Correa, A., Lewis, Z.A., Greene, A.V., March, I.J., Gomer, R.H., and Bell-Pedersen, D. 2003.Multiple oscillators regulate circadian gene expression in Neurospora. Proc Natl Acad SciUSA 100(23): 13597–13602.

Crosthwaite, S.K., Dunlap, J.C., and Loros, J.J. 1997. Neurospora wc-1 and wc-2: Trans-cription, photoresponses, and the origin of circadian rhythmicity. Science 276: 763-769.

Davidson, A.J., Sellix, M.T., Daniel, J., Yamazaki, S., Menaker, M., and Block, G.D. 2006.Chronic jet-lag increases mortality in aged mice. Curr Biol 16(21): R914-916.

Davis, F.C., & Menaker, M. 1980. Hamsters through time's window: Temporal structure ofhamster locomotor rhythmicity. Am J Physiol 239: R149-R155.

Diaz-Flores, E. and Shannon, K. 2007. Targeting oncogenic Ras. Genes & development21(16): 1989-1992.

Dragovic, Z., Tan, Y., Görl, M., Roenneberg, T., and Merrow, M. 2002. Light reception andcircadian behavior in 'blind' and 'clock-less' mutants of Neurospora crassa. EMBO J21(14): 3643-3651.

Dunlap, J.C. and Loros, J.J. 2004. The Neurospora Circadian System. J Biol Rhythms 19:414-424.

Engelmann, W., Karve, A., and Schoser, G. 1961. Initation of rhythmical petal movements inKalanchoe blossfeldiana by transfer from continuous darkness to continuous light or viceversa. 56: 700-711.

Fu, L., Pelicano, H., Liu, J., Huang, P., and Lee, C.C. 2002. The circadian gene Period2 plays animportant role in tumor suppression and DNA damage responses in vivo. Cell 111: 41-50.

Görl, M., Merrow, M., Huttner, B., Johnson, J., Roenneberg, T., and Brunner, M. 2001. APEST-like element in FREQUENCY determines the length of the circadian period inNeurospora crassa. EMBO J 20(24): 7074-7084,.

Hardin, P.E., Hall, J.C., and Rosbash, M. 1990. Feedback of the Drosophila period geneproduct on circadian cycling of its messenger RNA levels. Nature 343: 536-540.

Hastings, J.W. and Sweeney, B.M., ed. 1959. The Gonyaulax clock. A.A.A.S. Prss.,Washington D.C.

He, Q., Cheng, P., Yang, Y., Wang, L., Gardner, K.H., and Liu, Y. 2002. White collar-1, a DNAbinding transcription factor and a light sensor. Science (New York, NY 297(5582): 840-843.

Kondo, T., Tsinoremas, N.F., Golden, S.S., Johnson, C.H., Kutsuna, S., and Ishiura, M. 1994.Circadian clock mutants of cyanobacteria. Science 266: 1233-1236.

TIME BANDIT

133

Konopka, R. and Benzer, S. 1971. Clock mutants of Drosophila melanogaster. Proc Natl AcadSci USA 68: 2112-2116.

Lakin-Thomas, P.L. and Brody, S. 2000. Circadian rhythms in Neurospora crassa : Lipid defi-ciencies restore robust rhythmicity to null frequency and white-collar mutants. Proc NatlAcad Sci USA 97: 256-261.

-. 2004. Circadian rhythms in microorganisms: new complexities. Annu Rev Microbiol 58:489-519.

Lakin-Thomas, P.L., Coté, G.G., and Brody, S. 1990. Circadian rhythms in Neurospora crassa:Biochemistry and Genetics. Crit Rev Microbiol 17: 365- 416.

Liu, Y., Tsinoremas, N.F., Johnson, C.H., Lebedeva, N.V., Golden, S.S., Ishiura, M., andKondo, T. 1995. Circadian orchestration of gene expression in cyanobacteria. GenesDevelop 9: 1469-1478.

Loros, J.J. and Feldman, J.F. 1986. Loss of temperature compensation of circadian periodlength in the frq-9 mutant of Neurospora crassa. J Biol Rhythms 1: 187- 198.

Merrow, M., Brunner, M., and Roenneberg, T. 1999. Assignment of circadian function for theNeurospora clock gene frequency. Nature 399: 584-586.

Merrow, M., Spoelstra, K., and Roenneberg, T. 2005. The circadian cycle: daily rhythms frombehaviour to genes. EMBO reports 6(10): 930-935.

Nakajima, M., Imai, K., Ito, H., Nishiwaki, T., Murayama, Y., Iwasaki, H., Oyama, T., andKondo, T. 2005. Reconstitution of circadian oscillation of cyanobacterial KaiC phospho-rylation in vitro. Science 308: 414-415.

Park, D.H., Somers, D.E., Kim, Y.S., Choy, Y.H., Lim, H.K., Soh, M.S., Kim, H.J., Kay, S.A.,and Nam, H.G. 1999. Control of circadian rhythms and photoperiodic flowering by theArabidopsis GIGANTEA gene. Science 285: 1579-1582.

Pittendrigh, C.S., Bruce, V.G., Rosensweig, N.S., and Rubin, M.L. 1959. Growth patterns inNeurospora crassa. Nature 184: 169-170.

Ptitsyn, A.A., Zvonic, S., and Gimble, J.M. 2007. Digital signal processing reveals circadian base-line oscillation in majority of mammalian genes. PloS computational biology 3(6): e120.

Roenneberg, T., Dragovic, Z., and Merrow, M. 2005. Demasking biological oscillators: prop-erties and principles of entrainment exemplified by the Neurospora circadian clock.Proceedings of the National Academy of Sciences of the United States of America 102(21):7742-7747.

Roenneberg, T. and Morse, D. 1993. Two circadian oscillators in one cell. Nature 362: 362-364.Roenneberg, T. and Rehman, J. 1996. Nitrate, a nonphotic signal for the circadian system.

FASEB J 10: 1443-1447.Sargent, M.L., Briggs, W.R., and Woodward, D.O. 1966. Circadian nature of a rhythm

expressed by an invertaseless strain of Neurospora crassa. Plant Physiol 41: 1343-1349.Schafmeier, T., Kaldi, K., Diernfellner, A., Mohr, C., and Brunner, M. 2006. Phosphorylation-

dependent maturation of Neurospora circadian clock protein from a nuclear repressortoward a cytoplasmic activator. Genes & development 20(3): 297-306.

Spoelstra, K., Albrecht, U., van der Horst, G.T.J., Brauer, V., and Daan, S. 2004. Phaseresponses to light pulses in mice lacking functional per or cry genes. J Biol Rhythms19(6): 518-529.

Vitaterna, M.H., King, D.P., Chang, A.-M., Kornhauser, J.M., Lowrey, P.L., McDonald, J.D.,Dove, W.F., Pinto, L.H., Turek, F.W., and Takahashi, J.S. 1994. Mutagenesis and mappingof a mouse gene, Clock, essential for circadian behavior. Science 264: 719-725.

Wittmann, M., Dinich, J., Merrow, M., and Roenneberg, T. 2006. Social jet-lag: sleep, well-beingand stimulus consumption of different chronotypes. Chronobiol Int 23(1&2): 497-509.

Yan, O.Y., Andersson, C.R., Kondo, T., Golden, S.S., Johnson, C.H., and Ishiura, M. 1998.Resonating circadian clocks enhance fitness in cyanobacteria. PNAS 95(15): 8660-8664.

Yang, Z. and Sehgal, A. 2001. Role of molecular oscillations in generating behavioralrhythms in Drosophila. Neuron 29(2): 453-467.

Young, M.W. and Kay, S.A. 2001. Time zones: a comparative genetics of circadian clocks.Nature Reviews Genetics 2: 702-715.

Cellular clocks: circadian rhythms in primary humanfibroblasts

M. Merrow, C. Boesl and T. Roenneberg

Published in J. Biosci. 30(5), (2005): 553-555

9CHAPTER

Almost 300 years ago, a French astronomer made the observation that daily leafmovement continues even when a plant is kept in constant darkness (De Mairan1729). These so-called circadian rhythms exist at all levels of biology, rangingfrom gene expression to complex behaviours. They are controlled by a cellularclock that has been observed in organisms from all phyla. The underlying molec-ular mechanism has been described using genetic strategies that identified a setof clock genes that function in a transcriptional regulatory loop (Young and Kay2001). Mutations in any of these genes can cause disruption in some facet ofcircadian timing, and micro-array studies suggest that – in complex organisms –most cells are capable of generating circadian oscillations (Panda et al 2002).Thus, circadian clocks are cell-based.

Circadian clocks are entrained to exactly 24 h in nature, where organisms usevarious cues from the environment (zeitgebers) that cycle reliably and thusprecisely represent the rotation of the Earth. The best understood and perhapsthe strongest zeitgeber is light, which changes systematically in intensity and inspectral quality over the course of each day, in addition to the changing ratio oflight and darkness over the course of the year. Similar to physical oscillators,biological clocks will entrain differently depending on their period and ampli-tude. A relationship between external (entraining) and internal (circadian)period has been noted in animals (Hoffmann 1963; Pittendrigh and Daan 1976)and – at least in young adults – in humans (Duffy and Czeisler 2002). In general,a long free running circadian period entrains late in the day relative to a shorterone. Spore formation in fungal clock mutants with a short period occurs earlierin the night compared to wild type strains, and hamsters are active earlier orlater in the 24 h light-dark cycle according to their free running period(Pittendrigh and Daan 1976; Merrow et al 1999). Importantly, the palette ofclock-regulated physiologies also reflects the phase of entrainment. The implica-tions of physiological chronotype are substantial, ranging from optimizingmedical treatment to quality of life and shift work. To understand chronotype weneed to first understand free running rhythms, and then what happens whenthey are entrained.

Obviously, in humans, circadian rhythms are typically observed in theentrained state. Pioneering experiments did study free running rhythms inhumans, showing an approximate 25 h period in many parameters, includingsleep/wake cycles and core body temperature. The complexity of a circadiansystem was apparent in this early work when activity and temperature rhythmsdissociated, resulting in two free-running rhythms in a single individual (Aschoff1965). More recently, human rhythms are studied in forced desynchrony proto-cols, that estimate the period to be closer to 24 h (Czeisler et al 1999). What isapparent to chronobiology researchers is the cumbersome nature of this sort of

CHAPTER 9

136

human experimentation, requiring sequestration of subjects for days, weeks ormonths in temporal isolation.

This is at the least expensive and it can also be psychologically challengingfor subjects. Thus, more efficient tools to study human rhythms are overdue.

A recent publication in PLoS describes such a tool, namely a luminescentreporter that is used to follow cellular oscillations (Brown et al 2005). Theauthors built on their previous work showing that fibroblast cell lines in tissueculture could be synchronized such that they display coordinated circadian regu-lation of gene expression (Balsalobre et al 1998). Here (Brown et al 2005),primary human cells were transformed with a lentiviral vector to insert a fusionof the clock gene BMAL1 promoter and the coding region of firefly luciferase. Therecipient cells were derived either from skin biopsies (fibroblasts), hair roots(keratinocytes) or peripheral blood (monocytes).

In all cases, circa-24 h oscillations in BMAL1-driven luminescence wereobserved from the cultures, but the fibroblast system is – at least at this stage –the most robust. The authors compared different fibroblasts cultures from thesame individual and from different individuals, showing that intra-individualdifferences are smaller than inter-individual differences. Thus, the individual cellsappear to represent the characteristics of the donor’s clock. The critical questionis: can this tool replace some aspect of the temporal isolation experiments onhumans? All indications are that this will indeed be the case. Importantly, adistribution of cellular free running periods was observed amongst fibroblaststested from 19 individuals. The range was, however, larger than would beexpected according to the before mentioned isolation experiments, with periodsrunning from 23 h to over 26 h. To put this into perspective, the authors investi-gated fibroblast rhythms in mice and compared them to behavioural rhythms.Interestingly, while in all cases a correlation between cellular and behaviouralperiods was seen, the cell phenotype was more extreme. For instance, a mutantmouse with a short, 23.4 h period in running wheel activity showed a very short20 h oscillation in gene expression. This is not entirely surprising because it isapparent that the circadian rhythm of a tissue is more robust and consolidatedthan at the level of its dissociated cells (Welsh et al 1995; Vansteensel et al2003).

The circadian system is a collection of oscillators – within the body, withinorgans and tissues, and even perhaps at the molecular level within individualcells. The tools described by Brown et al (2005) open up exciting new possibili-ties for investigations into the molecular biology and genetics of clocks at thelevel of the cell as well as the system. Any number of other clock gene- or clock-regulated promoters can be employed to explore molecular clock regulation.Non-invasive sources of cells will surely be developed. Cells from different tissues

CELLULAR CLOCKS

137

can be investigated, to determine developmental and epigenetic effects on theclock genotype or even one day to use the circadian biology as a tool fordetecting pathologies at the tissue level. For basic research on human chronobi-ology, the new tools will significantly accelerate our understanding of how clockgenotype and cell-based rhythms shape the chronotype distribution in humanbehaviour.

References

CHAPTER 9

138

AcknowledgementsOur work is supported by the Deutsche Forschungsgemeinschaft, the NederlandseOrganisatie voor Wetenschappelijk Onderzoek, the Dr Meyer-Struckmann-Stiftung, and bythe EU (BrainTime).

Aschoff J 1965 Circadian Rhythms in Man; Science 148 1427–1432Balsalobre A, Damiola F et al 1998 A serum shock induces gene expression in mammalian

tissue culture cells; Cell 93 929–937Brown S A, Fleury-Olela F et al 2005 The Period Length of Fibroblast Circadian Gene

Expression Varies Widely among Human Individuals; PLoS 3 e338Czeisler C A, Duffy J F et al 1999 Stability, precision, and near-24-hour period of the human

circadian pacemaker; Science 284 2177–2181De Mairan J J d O 1729 Observation botanique; Histoire de l’Academie Royale des Science

35–36Duffy J F and Czeisler C A 2002 Age-related change in the relationship between circadian

period, circadian phase, and diurnal preference in humans; Neurosci. Lett. 318 117–120Hoffmann K 1963 Zur Beziehung zwischen Phasenlage und Spontanfrequenz bei der endo-

genen Tagesperiodik; Z. Naturforschg. 18b 154–157Merrow M, Brunner M et al 1999 Assignment of circadian function for the Neurospora clock

gene frequency; Nature (London) 399 584–586Panda S, Antoch M P et al 2002 Coordinated transcription of key pathways in the mouse by

the circadian clock; Cell 109 307–320Pittendrigh C S and Daan S 1976 A functional analysis of circadian pacemakers in nocturnal

rodents. IV. Entrainment: Pacemaker as clock; J. Comp. Physiol. A106 291–331Vansteensel M J, Yamazaki S et al 2003 Dissociation between circadian Per1 and neuronal

and behavioural rhythms following a shifted environmental cycle; Curr. Biol. 131538–1542,

Welsh D K, Logothetis D E et al 1995 Individual neurons dissociated from rat suprachias-matic nucleus express independently phased circadian firing rhythms; Neuron 14697–706

Young M W and Kay S A 2001 Time zones: a comparative genetics of circadian clocks; Nat.Rev. Genet. 2 702–715

CELLULAR CLOCKS

139

Summary

10CHAPTER

A few weeks ago, I sat at the dinner table with my parents and my husband,when my mother said: ‘This morning, I thought the sun was shining at me. Butwhen I opened my eyes I realized it was your father, reading with the lights on -at 5 o’clock in the morning!’. Then she continued counting up all the things myfather does in the ‘too early morning hours’ while she is still sleeping. The list ranfrom reading and housework to morning sports. This was the point for myhusband to introduce his experiences with his new wife waking up at 4:30 in themorning. Both husband and mother agreed in their criticism of this behavior.This anecdote illustrates two characteristics of human circadian systems: first,not all people have the same timing of their sleep-wake cycles, and second, thistiming is – in some aspects – inherited.

Considering the human sleep/wake cycle in particular and daily rhythms ingeneral, these two observations are demonstrated properties of circadian clocks.Specifically, they are inherited and show interindividual characteristics. Otherfeatures of circadian systems include stability over a broad range of environ-mental conditions (i.e., temperature compensation), self-sustainment and preci-sion. The adaptive capacity of circadian systems is reflected in a survival advan-tage according to clock characteristics (DeCoursey et al, 2000; Johnson &Golden, 1999)

Since the early descriptions of leaf movements by Androsthenes 2000 yearsago, a circadian clock has been described in organisms from almost all phyla,even unicellular algae and prokaryotes. The work presented in this thesis givesspecial emphasis on the circadian clock in the fungus Neurospora crassa, termed‘the organism behind the molecular revolution’ (Perkins, 1992). We use a simplecellular system to elucidate basic clock mechanisms that are comparable to thosein more complex organisms.

From chronoecology to the molecular mechanismsof the circadian clock

An important part of chronobiology is the field called chronoecology. It concernscircadian biology with a view to “the real world“ (Roenneberg et al, 2005).Chapters 2 and 3 are examples for how chronoecology is executed withNeurospora crassa.

Until very recently, N. crassa was thought to be an inhabitant of tropics andsubtropics. In such an environment, a system to measure changes in photope-riods would be unneccessary. This may be supported by the fact that - in nature -one substrate for Neurospora growth is underneath the bark of burnt trees. In2002, however, strains of N. discreta were found as far North as Alaska (Jacobson

CHAPTER 10

142

et al, 2004). In the following year, a collection trip (EuroNeuro2003) in Europeyielded N. crassa, N. dicreta, N. sitophila and N. tetrasperma, confirming thatNeurospora inhabits temperate climates (Jacobson et al, 2006). Novel growthsubstrates for Neurospora, namely on the bark of burnt trees rather than under-neath it, were identified on the European collection trip, also. These findingssuggest the ecological question: does Neurospora crassa display photoperiodism?In our group this question was addressed (Tan et al, 2004), showing thatNeurospora displays an integration mechanism to measure the length of day andnight (Roenneberg et al, 2004). Furthermore, it was shown that photoperiodicresponses are abolished in strains that carry a mutation in the clock gene frequency,pointing to the connection between the circadian clock and photoperiodism.

To investigate photoperiodism, natural populations from various latitudes area prized study object. Chapter 2 of this thesis describes the results of a Neuro-spora crassa collection trip in 2003 in Europe and compares the strain prevalenceand growth patterns with those of the previously known strains. From thiscomparison, we have a clearer picture on the phylogentic basis of the strains inthe collections that are available for researchers. The species from Europe fallinto the clade ‚NcB’ and have changed the descriptive biogeography. Further-more, we learned about the distribution of Neurospora, eventually finding italmost everywhere we looked where fires had occured. Although the climate,latitude and vegetation are similar between western North America and Europe,the natural substrate and growth pattern is dissimilar. When the collections inEurope are added to those from temperate North America and the older collec-tions from the tropics and subtropics, they form a worldwide collection ofNeurospora species that await evolutionary and ecological experiments in thefuture.

In chapter 3, the worldwide Neurospora crassa collection is used in chrono-ecology studies. The entire collection was screened for several circadian pheno-types. The relatedness of circadian traits (like free running period and phase)with the latitude-of-origin has been shown in several previous studies (Michael etal, 2007; Michael et al, 2003; Pittendrigh & Takamura, 1989; Sawyer et al,1997). The results from our ‚chronotyping’ of Neurospora wild type strains showthat latitudinal clines exist for some but not all phenotypes. We find that phase ofentrainment in light correlates with latitude-of-origin, whereas phase of entrain-ment in temperature does not.

Another result shows that the phase-period-rule applies in some cases but notin others. More precisely, entrainment with light does not correlate with period,but entrainment with temperature does. The findings could reflect the qualitiesof temperature as a zeitgeber (how it is perceived and transduced) in theNeurospora clock. It has been speculated before that temperature might be an

SUMMARY

143

even stronger zeitgeber than light for the circadian clock in this fungus (Liu et al,1998), but this was subsequently shown to depend on zeitgeber strength: in highamplitude temperature cycles (20ºC to 30ºC) and concurrent, antiphase lightcycles, the fungus entrains to temperature. But with a 5 degree amplitude cycleand concurrent light cycles (this time the cycles had different lengths, or T’s),entrainment follows the light cycle (Roenneberg & Merrow, 2001). From theexperiments described here, however, we can only speculate about zeitgeber prop-erties. An important difference between higher latitudes and lower latitudes is thatthe proportion of twilight is much higher in the North (see (Daan & Aschoff,1975). The earlier phases observed in the wild type strains from higher latitudescould enable these strains to perceive even dim light when twilight starts.

I conclude that the latitude of origin is reflected in the way an organismentrains to a stimulus, as revealed by an earlier phase in strains from a higherlatitude. However, the geographical origin is not reflected in the period length, sowe cannot simply deduce earlier entrainment from shorter periods (as e.g. inhumans (Duffy et al, 2001; Jones et al, 1999; Pittendrigh & Daan, 1976a;Pittendrigh & Daan, 1976b)).

These data demonstrate the principle that important clock characteristicssuch as period and chronotype can be encoded at numerous locations in thecircadian system. Presumably a functional change in a so-called core clock genewould change the free running period (changes in zeitnehmers can also changeperiod (Roenneberg & Merrow, 1998)). Here, it is not a requisite change inperiod that leads to a difference in chronotype (in entrainment with light). Thus,the data point to alterations either in inputs to the oscillator or to changes incoupling of outputs to the oscillator as a source of chronotype variability(Roenneberg et al, 1998).

Using the large collection of wild type strains, we demonstrated thatentrained phase or chronotype can vary independently of period. In thepreceding paragraph, I speculated on how this is informative, with respect toconstruction of the clock system. Considering that all clock mutants described sofar in Neurospora crassa have alterations in free running period, we hypothesizethat mutant screens in entrainment should reveal additional clock genes (Goedel,unpublished data). An alternative approach for uncovering gene-phene relation-ships is Quantitative Trait Locus (QTL) analysis. Chapter 4 utilizes selected wildtype strains to this end. A set of up to 500 progeny from a cross between twowild type strains, one from the Caribbean and the other from India, were chrono-typed and subjected to a QTL-analysis. The aim was to correlate genotypes tocircadian phenotypes (free running period and phase of entrainment). The corre-lation will eventually lead to new genes involved in circadian biology.Importantly, in a proof of principle experiment, we find the frequency (frq) gene

CHAPTER 10

144

cosegregating with circadian period. frq is one of the first clock genes discovered,with mutant alleles leading to an altered free running period or arhythmicity.Another gene, period-4 (prd-4), is a candidate gene linked to circadian tempera-ture compensation. prd-4 is indispensable for temperature compensation, alsoconferring a short period in k.o.-mutants. It was also found in an independentQTL-analysis for circadian clock genes in Neurospora (Kim et al, 2007). Wefurthermore found interesting candidates that have not been described previouslyin connection with the circadian clock. Two of them are hsp-80 and a largeregion, DA-25, that could both be targets of further research.

Another way to find novel clock genes is a functional genomics approach. Bythis method, candidate genes are identified and the phenotypic effects of aknock-out or of over-expression on the circadian system are determined. From atheoretical point of view, genes that are part of the input or output pathways ofthe circadian clock are extremely interesting. As mentioned earlier, some muta-tions in the so-called core clock genes will change the free running period orresult in arhythmicity. Mutations in the input or output pathways, however,might result in chronotype changes (Roenneberg et al, 1998). In chapter 5 welooked at the effects of a defect in the Neurospora cryptochrome (ncry) gene, aputative blue light photoreceptor identified in the genome project (Galagan et al,2003). I made a knock-out-mutant and found that it had a normal free runningperiod in constant darkness. However, under entrainment, the mutant shows adelayed phase (compared to the parental strain bd). The delay is morepronounced in long photoperiods, where more than 50% of the cycle has a lightsignal. The photoperiod-dependent entrainment phenotype is absent when bluelight is used, suggesting saturation of the dose-response mechanism. In shortphotoperiods using blue light (up to 10h of blue light per 24h) the delay in themutant compared to the background strain is even bigger compared to shortphotoperiods with white light. Whether the observed effects are due to qualita-tive or quantitative aspects of light cannot be concluded from the experimentspresented here. All in all, the cry-mutant could be more light-sensitive comparedto a wild type strain. A lack of CRY in N. crassa leads almost exclusively to adelayed phenotype. Since cryptochromes in other species are blue light photore-ceptors, we predict that ncry is part of the input pathway of the circadian clock.This could also explain why period is not changed in the ncry-mutant. Based onthe observation that a cry knock-out shows a late chronotype, we suggest thatcryptochrome is a clock gene in Neurospora crassa. These experiments in general–and unexpectedly- serve as a proof of principle concerning the relationship ofperiod and phase. Mutations in genes like frq show a strict relationship (e.g., seechapter 7); mutations in genes like cry do not. This feature may be useful inelaborating molecular mechanisms of clocks.

SUMMARY

145

The entrainment of the lab strain bd – as opposed to the wild type strains andprogeny of these – is described in chapter 6. The most important result reviewedhere is that frq RNA is regulated in a circadian manner in temperature cycles butnot in light cycles. Importantly, even frq null strains display an entrainable oscil-lator, as also shown in other studies (Merrow et al, 1999; Pregueiro et al, 2005;Roenneberg et al, 2005), speaking for a multi-oscillatory system.

Chapter 7 contains another contribution to understanding the rules ofentrainment and how they reveal circadian mechanisms. Entrainment in temper-ature cycles is described in a series of systematic physiological experiments.Temperature entrainment shows many similarities compared to entrainmentusing comparable T-cycles with light as a zeitgeber (Rémi, 2007). In T=22h, bdentrains with the onset of conidiation at midnight. The short period mutantentrains earlier and the long period mutant entrains later. The differencesbetween entrainment with temperature versus light (see also chapter 3 and 4)may be due to differences in the properties of the zeitgebers and how thesesignals are processed (for further discussion see final paragraph of this chapter).

The final two chapters deal with the issue of model organisms in circadianresearch. What are their strengths and weaknesses? What have we learned fromthem and what open issues remain?

Chapter 8 is a review on the cloning of the mutation behind the ‚band’ (bd)strain. A recent publication reveals that bd – the standard lab strain forNeurospora chronobiology - carries a mutation in ras-1 (Belden et al, 2007). Ras-1was discovered as an oncogene in rats (Diaz-Flores and Shannon 2007).Although it is generally accepted that the clock is not compromised in bd, amutated ras-1 leads –in the case of N. crassa- to elevated or decreased levels inthe RNA or protein level of some clock components, which implies that ras-1might be involved in the circadian biology of Neurospora crassa. However, theimportant experiments using entrainment in bd compared to the real wild typestrain 74OR are still missing.

How circadian signals are transduced within higher eukaryotic systems (e.g.tissues, organs or a body) can best be described by using other model systems.Chapter 9, a review chapter, leads away from N. crassa to another model,namely human fibroblasts described in a publication by Brown et al. (Brown etal, 2005). Here, a luminescent reporter is used as a tool to show cellular oscilla-tions. By now, this technique has been improved further (Brown et al, 2008), sothat there is hope that psychologically stressful temporal isolation and constantroutine experiments on humans can be avoided. To date, fibroblasts are indispen-sable and the tool of choice to study the complexity of cellular circadian clockswithin the mammalian body. However, there are still some issues to be solved,such as whether the rhythms are directly comparable to behavioural rhythms. In

CHAPTER 10

146

a recent publication by the same authors (Brown et al, 2008) the phase-period-rule is challenged once more. Here, fibroblast cultures having the same periodhave been found to display different phases. Rather, the amplitude of therhythms has been suggested to be correlated with phase. N. crassa could be anideal candidate to test this hypothesis further. Much of the work usingNeurospora was done before mammalian tissue culture systems were available. Todate, it is still an excellent cellular eukaryotic model system that can help tounderstand basic properties of circadian clocks on the physiological and molecularlevel.

To sum up, three major points arise from my work. The first refers to a resultthat came up several times (Chapter 3, 4 and 5), namely that phase and periodare not always linked in a direct relationship. In nature, selection can work onlyupon phase of entrainment (Roenneberg & Merrow, 2007). Both internal andexternal factors influence entrainment, such as sensitivity to the zeitgeber, themake-up of the transducing pathways, how oscillators and in- and outputs arecoupled, etc. (Roenneberg & Merrow, 2007). In humans, for example, chrono-type is strongly dependent on light schedules, sex and age. All in all, we shouldrethink the phase-period rule (when studying living organisms and especially ingenetically heterogeneous populations) and rather think in terms of complexityof factors controlling chronotype.

Secondly, the rules of entrainment using temperature as a zeitgeber wereestablished and are ready to be taken to the next level. In our experiments, inshort days the phases are dusk-linked in long T-cycles and dawn-linked in shortT-cycles, when light is used as a zeitgeber. When temperature is used as a zeit-geber, however, entrainment parallels midnight independent of cycle length.Furthermore, in T=22, the phase of entrainment in temperature cycles relates toperiod length in DD, but the patterns of entrainment of the used period mutantsvary considerably between the light and temperature protocols, in that frq1

(short period) entrains consistently earlier, and frq7 (long period) later intemperature cyles compared to light cycles. Whether this behaviour is due todifferent PRCs in temperature or light still has to be shown.

The natural zeitgeber landscape is highly complex and includes both temper-ature and light, and many other signals, as well (for example, nutrient cycles canentrain mammalian tissues (Stokkan et al, 2001)). As mentioned before, espe-cially in humans, many factors influence chronotype. Here, simple model organ-isms like N. crassa might help in unravelling contributions of single factors versuscombinations of two or three zeitgebers at a time. The insights gained might beapplicable also in fibroblast tissue cultures, thus moving up the ladder ofcomplexity. The entrainment surfaces that have been established will allow amore thorough understanding of how the clock entrains in reality.

SUMMARY

147

As some of the most important findings, I show novel data concerningentrainment in a cryptochrome mutant in N. crassa using white vs blue light. Inmy study, the impact of light quality on the circadian clock has been investigatedin a simple fungus. In almost all photoperiodic experiments described here, thecryptochrome mutant displays a delay in phase of entrainment compared to thebackground strain, with long photoperiods (more than 12h light per 24h cycle)leading to a stronger delay. In blue light, on the other hand, short photoperiodscause a bigger phase delay in short photoperiods with less than 12h light per 24hcycle. The similar delays in long photoperiods using white or blue light mightspeak for a saturated response. To date, however, we cannot yet say whether theresults are due to qualitative or quantitative effects of the light sources used.Although we cannot yet fully explain the findings, they serve as a template forexperimental protocols in subsequent functional genomic approaches. One of thebiggest interests in circadian research is understanding the workings of the circa-dian clock in natural conditions. To this end, our future goal is to elaborate rulesthat can be further applied to human circadian behaviour, as well. Eventually,this knowledge will be used to improve working conditions, e.g. during shiftwork, or in senior citizens’ homes.

References

CHAPTER 10

148

Belden WJ, Larrondo LF, Froehlich AC, Shi M, Chen CH, Loros JJ, Dunlap JC (2007) Theband mutation in Neurospora crassa is a dominant allele of ras-1 implicating RASsignaling in circadian output. Genes Dev 21(12): 1494-1505

Brown SA, Fleury-Olela F, Nagoshi E, Hauser C, Juge C, Meier CA, Chicheportiche R, DayerJM, Albrecht U, Schibler U (2005) The period length of fibroblast circadian gene expres-sion varies widely among human individuals. PLoS Biol 3(10): e338

Daan S, Aschoff J (1975) Circadian Rhythms of locomotor activity in captive birds andmammals: their variations with season and latitude. Oecologia 18: 269-316

DeCoursey PJ, Walker JK, Smith SA (2000) A circadian pacemaker in free-living chipmunks:essential for survival? J Comp Physiol [A] 186(2): 169-180

Duffy JF, Rimmer DW, Czeisler CA (2001) Association of intrinsic circadian period withmorningness-eveningness, usual wake time, and circadian phase. Behav Neurosci115(4): 895-899

Galagan JE, Calvo SE, Borkovich KA, Selker EU, Read ND, Jaffe D, FitzHugh W, Ma LJ,Smirnov S, Purcell S, Rehman B, Elkins T, Engels R, Wang S, Nielsen CB, Butler J,Endrizzi M, Qui D, Ianakiev P, Bell-Pedersen D, Nelson MA, Werner-Washburne M,Selitrennikoff CP, Kinsey JA, Braun EL, Zelter A, Schulte U, Kothe GO, Jedd G, Mewes W,Staben C, Marcotte E, Greenberg D, Roy A, Foley K, Naylor J, Stange-Thomann N,Barrett R, Gnerre S, Kamal M, Kamvysselis M, Mauceli E, Bielke C, Rudd S, Frishman D,Krystofova S, Rasmussen C, Metzenberg RL, Perkins DD, Kroken S, Cogoni C, Macino G,Catcheside D, Li W, Pratt RJ, Osmani SA, DeSouza CP, Glass L, Orbach MJ, Berglund JA,Voelker R, Yarden O, Plamann M, Seiler S, Dunlap J, Radford A, Aramayo R, Natvig DO,Alex LA, Mannhaupt G, Ebbole DJ, Freitag M, Paulsen I, Sachs MS, Lander ES, NusbaumC, Birren B (2003) The genome sequence of the filamentous fungus Neurospora crassa.Nature 422(6934): 859-868

SUMMARY

149

Jacobson DJ, Dettman JR, Adams RI, Boesl C, Sultana S, Roenneberg T, Merrow M, DuarteM, Marques I, Ushakova A, Carneiro P, Videira A, Navarro-Sampedro L, Olmedo M,Corrochano LM, Taylor JW (2006) New findings of Neurospora in Europe and compar-isons of diversity in temperate climates on continental scales. Mycologia 98(4): 550-559

Johnson CH, Golden SS (1999) Circadian programs in cyanobacteria: adaptiveness andmechanism. Annu Rev Microbiol 53: 389-409

Jones CR, Campbell SS, Zone SE, Cooper F, DeSano A, Murphy PJ, Jones B, Czajkowski L,Ptacek LJ (1999) Familial advanced sleep-phase syndrome: A short-period circadianrhythm variant in humans. Nat Med 5(9): 1062-1065

Kim TS, Logsdon BA, Park S, Mezey JG, Lee K (2007) Quantitative trait loci for the circa-dian clock in Neurospora crassa. Genetics 177(4): 2335-2347

Merrow M, Brunner M, Roenneberg T (1999) Assignment of circadian function for theNeurospora clock gene frequency. Nature 399(6736): 584-586

Michael TP, Park S, Kim TS, Booth J, Byer A, Sun Q, Chory J, Lee K (2007) Simple sequencerepeats provide a substrate for phenotypic variation in the Neurospora crassa circadianclock. PLoS ONE 2(8): e795

Michael TP, Salome PA, Yu HJ, Spencer TR, Sharp EL, McPeek MA, Alonso JM, Ecker JR,McClung CR (2003) Enhanced fitness conferred by naturally occurring variation in thecircadian clock. Science 302(5647): 1049-1053

Perkins DD (1992) Neurospora: the organism behind the molecular revolution. Genetics130(4): 687-701

Pittendrigh CS, Daan S (1976a) A functional analysis of circadian pacemakers in nocturnalrodents: V. Pacemaker structure: a clock for all seasons. J Comp Physiol A 106: 333-355

Pittendrigh CS, Daan S (1976b) A functional analysis of circadian pacemakers in nocturnalrodents. IV. Entrainment: Pacemaker as clock. J Comp Physiol A 106: 291-331

Pittendrigh CS, Takamura T (1989) Latitudinal clines in the properties of a circadian pace-maker. J Biol Rhythms 4(2): 217-235

Pregueiro AM, Price-Lloyd N, Bell-Pedersen D, Heintzen C, Loros JJ, Dunlap JC (2005)Assignment of an essential role for the Neurospora frequency gene in circadian entrain-ment to temperature cycles. Proc Natl Acad Sci USA 102(6): 2210-2215

Roenneberg T, Dragovic Z, Merrow M (2005) Demasking biological oscillators: propertiesand principles of entrainment exemplified by the Neurospora circadian clock. Proc NatlAcad Sci USA 102(21): 7742-7747

Roenneberg T, Merrow M (1998) Molecular circadian oscillators: an alternative hypothesis.J Biol Rhythms 13(2): 167-179

Roenneberg T, Merrow M, Eisensamer B (1998) Cellular mechanisms of circadian systems.Zoology 100: 273-286

Roenneberg T, Tan Y, Dragovic Z, Ricken J, Kühnle T, Merrow M (2004) Chrono-ecologyfrom fungi to humans. In Working in the 24 hour environment - gene mechanisms andhuman life, Honma K, Honma S (eds). Sapporo, in press: Hokkaido Univ. Press

Sawyer LA, Hennessy JM, Peixoto AA, Rosato E, Parkinson H, Costa R, Kyriacou CP (1997)Natural variation in a Drosophila clock gene and temperatur compensation. Science278: 2117-2120

Stokkan KA, Yamazaki S, Tei H, Sakaki Y, Menaker M (2001) Entrainment of the circadianclock in the liver by feeding. Science 291: 490-493

Tan Y, Merrow M, Roenneberg T (2004) Photoperiodism in Neurospora crassa. J BiolRhythms 19(2): 135-143

Nederlandse samenvatting

Een paar weken geleden, zat ik met mijn ouders en mijn echtgenoot aan tafel,toen mijn moeder zei: ‘Vanochtend, dacht ik dat de zon scheen, maar toen ikmijn ogen opende realiseerde ik me dat je vader om 5 uur in de ochtend met hetlicht aan het lezen was! ‘Toen somde ze alle dingen op die mijn vader in de ‘tevroege ochtenduren’ doet terwijl zij nog slaapt. De lijst liep van lezen tot huis-houdelijk werk aan ochtendsporten.

Dit was het moment voor mijn echtgenoot om te beginnen over zijn ervaringmet het ontwaken om 4:30 in de ochtend van zijn nieuwe vrouw. Zowel moederen echtgenoot stemde in met de kritiek van dit gedrag. Deze anekdote illustreerttwee kenmerken van het menselijke circadiane systeem: allereerst verschillenmensen in hun slaap waak ritme daarnaast is dit enigszins genetisch overdraag-baar.

Gezien het menselijke slaap waak patroon specifiek hun dagelijkse ritme zijndeze twee eigenschappen aangetoonde eigenschappen. De eigenschappen zijnovererfbaar en tonen overeenkomsten tussen mensen onderling. Anderekenmerken van circadiane systemen zijn stabiliteit binnen een brede variatie inomgevingscondities (bv. temperatuurcompensatie), zelfonderhoud en precisie.Het adaptieve voordeel van circadiane systemen is duidelijk aangetoond in expe-rimenten waarin een overlevingsvoordeel gerelateerd aan klokkarakteristiekenbleek te zijn (DeCoursey et al, 2000; Johnson & Golden, 1999)

De eerste beschrijving van de circadiane klok is reeds 2000 jaar geledenbeschreven in bladbewegingen door Androsthenes. Momenteel is deze circadianeklok in vrijwel alle levensvormen van eencellige algen tot prokaryoten gevonden.Het werk in dit proefschrift richt zich speciaal op de circadiane klok in deschimmel Neurospora crassa, ook wel genoemd ‘the organism behind the mole-cular revolution’ (Perkins, 1992). We gebruiken dit eenvoudige cellulaire systeemom basale klokmechanismen te bestuderen die kunnen worden vergeleken metde mechanismen aanwezig in meer complexe mechanismen.

Van chrono-ecologie naar het moleculaire mechanismein de circadiane klok

Een belangrijk deel binnen de chronobiologie is het veld chrono-ecologie, hetbeschouwd circadiane biologie vanuit een natuurlijk perspectief “the real world”(Roenneberg et al, 2005). Hoofdstuk 2 en 3 geven voorbeelden hoe chrono-ecologie wordt bestudeerd met Neurospora crassa.

Recentelijk werd nog aangenomen dat Neurospora crassa voorkwam in detropen en subtropen. In deze gebieden variëren de dag en nacht lengtes (fotope-rioden) minimaal gedurende het jaar, (Daan & Aschoff, 1975). Dit suggereert dat

SAMENVATTING

152

een systeem dat veranderingen in daglengtes registreert onnodig is. Dit wordtondersteund door het feit dat, in de natuur, de typische voedingsbron voor degroei van Neurospora afkomstig is van verbrand gras en van achter de schors vanverbrande bomen. In 2002, zijn er zelfs noordelijk zoals Alaska stammen vanN. discreta gevonden (Jacobson et al, 2004). De ,EuroNeuro2003 collection trip’in Europa leverde N. crassa, N. dicreta, N. sitophila en N. tetrasperma op, watbevestigde dat Neurospora ook voorkomt in de meer gematigde klimaten (Jacob-son et al, 2006). Dit – samen met nieuw ontdekte voedingsbronnen, voorname-lijk op de schors van verbrande bomen – opent de vraag: Vertoond Neurosporacrassa fotoperiodisme? In onze groep werd deze vraag gesteld (Tan et al, 2004)en aangetoond dat Neurospora een geïntegreerd mechanisme heeft voor hetmeten van dag en nachtlengte (Roenneberg et al, 2004). Verder is aangetoonddat fotoperiodisme verdwijnt in stammen met een mutatie in het klokgenfrequency, dit geeft de connectie aan tussen de circadiane klok en fotoperiodisme.

Om fotoperiodisme te bestuderen zijn natuurlijke populaties van verschil-lende breedtegraden verzameld. Hoofdstuk 2 van dit proefschrift beschrijft deresultaten van een verzameling Neurospora crassa uit 2003 uit Europa en verge-lijkt deze stammen met reeds bekende stammen. Uit deze vergelijking hebben weeen beter beeld van fylogenetische basis. De toegevoegde soorten uit Europavallen in de ,NcB’ groep en hebben de beschrijvende biogeografie veranderd.Verder, hebben we geleerd over de distributie van Neurospora, die bijna overal opplekken waar bosbrand is geweest, kan worden aangetroffen. Ook al is hetklimaat, de breedtegraad en vegetatie vergelijkbaar tussen Noordwest Amerikaen Europa , de natuurlijke voedingsbronnen en groeipatronen zijn verschillend.Als we de collectie uit Europa samenvoegen met deze uit gematigd Noord-Amerika en de oudere collecties uit de tropen en de subtropen, dan vormen dezeeen wereldwijde collectie van Neurospora soorten die kunnen worden voor evolu-tionaire en ecologische experimenten.

In hoofdstuk 3 is de wereldwijde N. crassa collectie gebruikt voor chrono-ecologische studies. De relatie tussen circadiane eigenschappen (zoals vrijloopperiode en fase) en die van afstamming van een bepaalde breedtegraad is inmeerdere studies aangetoond (Michael et al, 2003; Pittendrigh & Takamura,1989). Onze resultaten bestaande uit chronotypering van Neurospora wildtypestammen uit verschillende circadiane protocollen en tonen aan dat gradiënt inbreedtegraad gelijk opgaat met het chronotype voor sommige maar niet allefenotypes.

Enkele genetische componenten van de circadiane klok zijn bekend. Omnieuwe klok genen te vinden zijn twee methoden toegepast: QTL (quantitativetrait loci) analyse, waarmee klok-gen kandidaten kunnen worden gevonden doorgezamenlijke chromosomale uitsplitsing met een circadiaan fenotype (hoofdstuk

SAMENVATTING

153

4), en door het muteren van klok-gen kandidaten, een functionele aanpak. Viadeze manier kunnen specifieke gene geïdentificeerd worden en de fenotypischeeffecten van een knock-out of over-expressie op het circadiane systeem wordengedetermineerd. Vanuit een theoretisch oogpunt, zijn genen die verondersteldworden deel uit te maken van in- output netwerk van de circadiane klok zeerinteressant. Zoals al eerder genoemd veroorzaken mutaties in de zogenoemde“kern klok genen” naar alle waarschijnlijk de verandering van de free-runningperiod of het resulteert in aritmiek. Echter mutaties in het in- of output netwerkkunnen resulteren in min of meer subtiele veranderingen (blijft in fase, zoalsbeschreven in Roenneberg et al, 1998). In hoofdstuk 5 hebben we keken naar deeffecten van een defect in het Neurospora cryptochrome (cry) gen, een veronder-steld blauw licht fotoreceptor geïdentificeerd in het “genome project” (Galaganet al, 2003). Ik heb een knock-out mutant gemaakt welke een normale free-running periode liet zien in constant donker. Echter, onder entrainment, laatdeze mutant een vertraagde fase zien (in vergelijking met de ouderlijke stam bd).De vertraging is duidelijk sterker aanwezig in een langere fotoperiode, waarbijmeer dan 50% van de cyclus blootgesteld is aan licht. Deze eigenschap is afwezigwanneer blauw licht wordt toegepast, dit suggereert verzadiging van het mecha-nisme wat reageert op licht blootstelling. In korte fotoperiodes waarbij gebruikwordt gemaakt van blauw licht ( waarbij tot 10 uur van de 24 uur blauw licht isgebruikt) in vergelijking met de mutant, ten opzichte van de controle stam, devertraging sterker aanwezig dan in aanwezigheid van wit licht.

Of de geobserveerde effecten ook daadwerkelijk het gevolg van kwalitatieveof kwantitatieve aspecten, als gevolg van het toegepaste licht, zijn kan nietgeconstateerd worden uit de hier gepresenteerde experimenten. Samenvattend,de cry-mutant zou meer lichtgevoelig kunnen zijn. Afwezigheid van CRY leidtbijna altijd tot een vertraagd fenotype. Omdat cryptochromen in andere soortenvan Neurospora blauw licht fotoreceptoren zijn, voorspellen wij dat cry deeluitmaakt van het input netwerk van de circadiane klok, dit zou tevens kunnenverklaren waarom de centrale oscillator niet wordt beïnvloed en waarom deperiode onveranderd is in de cry mutant. Op basis van de observaties dat cryknock-outs een vertraagt chronotype laten zien suggereren wij dat cryptochroomeen klok gen is van Neurospora crassa.

Entrainment van de lab-stam band (bd) – in tegenstelling tot de wildtypestammen en hun nageslacht word beschreven in hoofdstuk 6. Het belangrijksteresultaat dat hier word beschreven is dat frq RNA in een circadiane wijze wordgereguleerd in temperatuur cycli, maar niet in licht cycli. Belangrijk hierbij is, datzelfs frq-null stammen een entraineerbare oscillator laten zien, wat ook uit anderstudie is gebleken (Merrow et al 1999;Pregueiro et al, 2005; Roenneberg et al2005), wat pleit voor een multi-oscillatoir Systeem.

SAMENVATTING

154

Hoofdstuk 7 bevat een andere bijdrage aan het begrijpen van de regels vanentrainment en hoe deze het circadiane mechanisme blootleggen. Entrainmentbinnen temperatuur cycli wordt uitgelegd in een reeks van systematische fysiolo-gische experimenten. Temperatuur entrainment laat veel overeenkomsten zientussen vergelijkbare T-cycli waarbij licht gebruik wordt als zeitgeber (Remi,2007).

In T=22h, entraineert de band (bd) stam bij de start van sporenvorming ommiddernacht. De korte periodemutant entraineert vroeger en de lange periodemutant entraineert later. Het verschil tussen entrainment met temperatuurversus licht ( zie ook hoofdstuk 3 en 7) zou verklaart kunnen worden doorverschillen in de eigenschappen tussen de seingevers en hoe deze signaleringwordt verwerkt.

De laatste 2 hoofdstukken omvatten het probleem van model organismen incircadiaan onderzoek. Wat zijn de sterke en wat zijn de zwaktepunten van deverschillende modellen? Wat hebben we van ze geleerd en welke problemenstaan nog open.

Hoofdstuk 8 betreft een overzicht dat de klonering van de band (bd) stammutatie beschrijft. Een recente publicatie onthult dat bd – de standaard laborato-rium stam voor Neurospora chronobiologie – een mutatie draagt in het ras-1(Belden et al, 2007). Ras-1 is ontdenkt als oncogen in ratten (Diaz-Floren enShannon 2007). Ondanks het algemeen geaccepteerd is dat de klok nog intact isin bd, veroorzaakt een mutatie in ras-1, in het geval van N.crassa, tor verhoogdeof verlaagde eiwit of RNA concentraties bij sommige klokcomponenten, watinhoud dat ras-1 mogelijk betrokken is bij de circadiane biologie van N.crassa.Desondanks zijn de belangrijke experimenten betreffende entrainment in bd invergelijking met de wildtype stam 74OR nog niet uitgevoerd.

Hoe circadiane signalen worden doorgegeven in hogere eukaryotischesystemen (bijv. weefsel, organen of het lichaam) kan het best worden beschrevendoor gebruik van andere model systemen. In het overzicht hoofdstuk 9 verplaat-sten we ons van het N.crassa modelsysteem naar een ander systeem, namelijk hetdoor Brown ( Brown et al, 2005) beschreven humane fibroblast modelsysteem.In deze publicatie wordt gebruik gemaakt van een luminescente reporter omcellulaire oscillaties aan te tonen. Op dit moment is deze techniek verder ontwik-kelt en verbeterd ( Brown et al , 2008), met de hoop dat de vaak psychologische,tijdelijke isolatie en stress volle experimenten op mensen kunnen wordenvermeden. Fibroblasten zijn hedendaags onvervangbaar en de eerste keuze omde complexiteit van de cellulaire circadiane klokken in zoogdieren te bestuderen.Echter, enige overwegingen behoren gemaakt te worden, zoals of deze ritmesdirect vergelijkbaar zijn met gedragsritme is het organisme. In een recente publi-catie van dezelfde auteurs (Brown et al 2008) word de faseperiode wet nog een

SAMENVATTING

155

keer aan de tand gevoeld. Hier, word beschreven dat fibroblast cultures met dezelfde periode andere fasen laten zien. Echter word hier gesuggereerd dat deamplitudes van de ritmes zijn gecorreleerd aan de fasen. N.crassa zou een idealekandidaat zijn om deze hypothese verder te testen. Voordat zoogdier weefstelcultures beschikbaar waren is er veel gewerkt met Neurospora cultures. Tegen-woordig is dit nog steeds een uitstekend cellulair eukaryotisch model systeem datkan helpen om het inzicht in basale eigenschappen van circadiane klokken opfysiek en moleculair niveau.

Tot slotte zal ik drie hoofdpunten uitlichten. Het eerste heeft betrekking ophet resultaat dat verschillende malen naar voren komt (hoofdstukken 3,4 & 7).Namelijk: dat fasen en periode niet altijd in directe relatie aan elkaar gekoppeldzijn. Als resultaat hiervan zouden we faseperiode wet moeten heroverwegen (bijhet bestuderen van levende organismen) en ons richten op verschillende locatiesbinnen het complexe circadiane netwerk dat het chronotype bepaald. Tentweede, de wetten van entrainment, waar temperatuur als zeitgeber wordtgebruikt zijn al om geaccepteerd en klaar om naar een volgend niveau gebrachtte worden. Meer specifiek, de natuurlijke aspecten van zeitgeber hebben eenhoge complexiteit van bevat zowel temperatuur als licht (naast andere signalen).De basis die we hebben ontwikkeld zal ons helpen om een nog beter beeld vande werkelijke manier van klokentrainment te krijgen. Tenslotte, laat ik recentedata zien dat verkregen is tijdens onderzoek naar entrainment in een crypto-chroom mutant, waarbij we gebruik hebben gemaakt van wit versus blauw licht.Ondanks dat we onze bevindingen nog niet volledig kunnen verklaren, kunnenwe deze wel gebruiken als fundering voor experimentele protocollen in aanslui-tende functionele genomische doeleinden.

SAMENVATTING

156

Dankwoord/Dankwort/Acknowledgements

Es ist ein ziemlich langer Weg von der Entscheidung, eine Doktorarbeit schreibenzu wollen, über die Experimente, deren Auswertung zum Zusammenschreibenund zur endlichen Fertigstellung des Buches. Auf meinem Weg als Doktorandinhaben mich viele liebe Menschen begleitet, ohne deren Unterstützung Vielesnicht möglich gewesen wäre und die meine Zeit als Doktorandin zu etwasBesonderem für mich gemacht haben.

In front of all I am indebted to my ‘Doktor-Mutter’ Martha Merrow and my‘Doktor-Vater’ Till Roenneberg. You both have been much more than just supervi-sors for my doctorate. In you both, I found encouragement, support, enthusiasm,never ending ideas, and the trust in me as a student. I was able to experimentwith methods and ideas, just as one would wish for his time as a pHD-student.You also supported my moving to and back from the Netherlands, giving me theopportunity to smell a bit of the great scientific athmosphere in Groningen.

Prof. Ernst Pöppel möchte ich danken für die wissenschaftliche und persoen-liche Unterstützung während all der Zeit als Doktorandin. Das IMP (Institut fürMedizinische Psychologie) ist mir in den letzten Jahren ein Zuhause gewesen fürmeine Experimente und ein Ort, an dem ich mich einfach wohl gefühlt habe.

Serge Daan and Domien Beersma I would like to thank for supporting meduring my stay in Groningen and for the great scientific advice.

Menno Gerkema, Michael Brunner and Rolf Hoekstra accepted to be part ofthe reading committee. They furthermore provided valuable comments on thechapters in this thesis which helped me in many ways in improving this thesis.

David Jacobson took me and Shahana Sultana along on the EuroNeuro2003collection trip (it was a great week with a lot of travelling, searching and finallyfinding, and great fun!) and gave me a warm welcome with his family inCalifornia during the Asilomar Fungal Genetics Conference 2005.

Bill Schwartz I want to thank for an open ear and the neccessary support thathelped me in gaining confidence towards the end of my thesis.

David Jacobson, Namboori B. Raju and Pamarthi Maruti Mohan made thephotos of Neurospora shown in the introductions to some of the chapters andkindly agreed in me using them.

My time as a phD-student would not be the same without my wonderfulcollegues in Munich and Groningen:

Zuerst möchte ich Zdravko Dragovic, Jan Remi und Kruno Sveric für dielustige Zeit im Labor danken, ich durfte viel von Euch lernen.

Viel verdanke ich ausserdem meinen beiden ‘Paranymphen’, David Lenssenund Manfred Gödel. Ihr seid immer liebe Kollegen und Freunde gewesen, die zurSeite standen mit Hilfe beim Experimentieren sowie in anderen Lebenslagen,

ACKNOWLEDGEMENTS

158

bayrischen oder holländischen Schmankerln und einem offenen Ohr. Davidverdanke ich auch die Übersetzung der Zusammenfassung ins Holländische undunermessliche Hilfe bei der Organisation.

Dir, liebe Julia, Grottenolmlounge-Mitbewohnerin, Danke für die vielenlustigen, kollegialen Stunden und Deine Ratschläge!

Thomas Kantermann und ich saßen am Ende im selben Boot beimZusammenschreiben der Doktorarbeit. Er hat mir sehr viel durch Kommentareund Ratschläge zum Manuskript geholfen.

Den lieben anderen Münchner Kollegen möchte ich ebenso von Herzendanken für die schöne Zeit, die Kollegialität und Freundschaft: Karla Allebrandt,Julia Diegmann, Myriam Juda, Susanne und Thomas Kantermann, Tim Kuehnle,Ildiko Meny, Marlies Messerschmidt, Tanja und Tamara Radic, Silke Sondermaier,Astrid Stück, Shahana Sultana, Ying Tan, Celine Vetter.

Jeremias Gromotka hat die Blaulicht-Boxen gebaut, mit Hilfe derer ich meineMutante charakterisieren konnte.

Nicht vergessen will ich auch die fleissigen und zuverlässigen Studenten undStudentinnen, die in München und Groningen an meinen Projekten mitgeholfenhaben. Vor allem die Studentinnen aus Padua möchte ich hier erwaehnen, AnnaMarchetti, Elisabetta Trevellin, Miriam Franghini und Tanja Radic.

Ich danke auch unseren Laborassistentinnen Astrid Bauer und Vera Schiewe,die immer mit helfender Hand bereit standen, wenn mal Not am Mann war.

Ein herzliches Dankeschön an Helmut Klausner für seine hervorragendeArbeit. Er hat für uns nach Wunsch geschweisst, gelötet, gesaegt und geklebt,und viele unserer Experimente erst möglich gemacht.

Ausserdem möchte ich noch Danke sagen an die anderen guten Seelen amIMP, Ruth Hoffmann, Petra Carl, Christiane von Kentzingen und SusannePiccone, sowie Erika Branse aus der Personalverwaltung.

Thanks to all the dear collegues at the Chronobiology department of theUniversity of Groningen, especially the Merrow lab, for your friendship and ascientific and warm atmosphere.

Furthermore, I am indebted to Arjen Strijkstra, Margien Raven, Jasper andEdwin for doing a great job in translating parts of my thesis. Marlies Hof, thanks so much for all your help in filling forms, organizing thingsand having a nice chat whenever we meet.

Henk Visser, dank je wel voor je hulp!The credit for making my thesis a proper book goes to Dick Visser. Thanks for

the patience and excellent work you did.

ACKNOWLEDGEMENTS

159

Besonders danken möchte ich auch meiner Familie, v.a. meinen lieben Eltern undmeiner Schwester. Ihr wart immer für mich da, auch wenn ich mich entschloss,mal wieder mit dem Kopf durch die Wand zu gehen. Eure Liebe und die positiveArt, an Dinge heranzugehen, hat mir oft geholfen, auch tiefe Tiefen zu über-stehen und fest an ein gutes Ende zu glauben.

Mein Mann Sudhakar Madeti hat besonders viel zu dieser Arbeit beigetragen. Erhat mich mit allen ihm möglichen Mitteln unterstützt und auch in schwierigenSituationen immer an mich geglaubt.

ACKNOWLEDGEMENTS

160