Introduction
Circadian Rhythms: Function and Fundamental Properties
The environment on Earth is constantly changing. The day/night cycle imposes dramatic variations in light intensity and temperature on most organisms. Moreover, yearly cycles in day length and weather patterns add a layer of complexity to these daily rhythms. The ecological surroundings of most organisms are also rhythmic: food availability, predator activity, and mating opportunities usually show both daily and yearly variations. To cope with the day/night cycle, cyanobacteria, fungi, plants, and animals use circadian clocks (Dunlap, 1999). The word "circadian" is derived from the Latin circa diem, which means "about a day." Circadian clocks help organisms to optimize gene expression, protein activity, cell biochemistry, body physiology, and behavior with the day/night cycle to maximize their chances of survival. There are four fundamental properties that define circadian rhythms. Perhaps their most striking property is their "free running" nature, which means that they persist under constant conditions, without any environmental cues. A second canonical property is that these circadian rhythms have a period of approximately, but not exactly, 24 hours. The third property is that they can be synchronized by various environmental inputs, called Zeitgebers (a German term meaning "time-giver"). Light intensity is arguably the most important Zeitgeber for most organisms. Other Zeitgebers include temperature cycles, food availability, and social interactions. The fourth fundamental property of circadian rhythms is called temperature compensation. The period of arcadian rhythms stays very stable under a wide range of ambient temperatures, although the kinetics of most biochemical reactions increase with higher temperature. These four canonical properties define endogenous circadian rhythms (Pittendrigh, 1960).
Insects are the most widely distributed and diverse group of animals on the planet. They can be found in nearly all environments, including those exposed to extremely cold and warm temperatures. Insects are thus powerful models to study circadian rhythms at the cellular, physiological, behavioral, ecological, and evolutionary levels.
Circadian Behavioral Rhythms
Many insect behaviors are under circadian regulation. Locomotor activity is the most intensively studied circadian behavior in insects.Drosophila melano-gaster, for example, shows two peaks of locomotor activity around dawn and dusk, which are frequently called morning and evening peaks, respectively (Figure 1). These two peaks persist under constant darkness and constant temperature, and are thus circadian, although the morning peak tends to progressively weaken, for unclear reasons. We will discuss in detail the neural bases for these two peaks in section 15.4, and the molecular mechanisms underlying these rhythms in section 15.2.
Another well-studied insect circadian behavior is feeding (Abraham and Muraleedharan, 1990; Khan et al., 1997; Sothern et al., 1998; Das and Dimopoulos, 2008; Xu et al., 2008). Among feeding rhythms, blood-sucking might be the first that comes to mind. Indeed, female mosquitoes bite us preferentially at certain times of the day, and are vectors of devastating diseases (malaria and Dengue fever, for example). These rhythmic biting activities have long been reported for various mosquitoes (see, for example, Haddow, 1956; Standfast, 1967), and were more recently proven to be controlled by circadian clocks (Pandian, 1994; Das and Dimopoulos, 2008). Related to feeding behavior, foraging behavior can also be circadianly controlled — for example, in honeybees and bumblebees (Moore et al., 1989; Bloch, 2010; Stelzer et al., 2010).
Circadian clocks also influence reproduction. Timing of mating behavior might even contribute to reproductive isolation, since closely related Drosophila species prefer to mate at different times (Sakai and Ishida, 2001a, 2001b). In Drosophila melanogaster, male courtship is more frequent during the night and seems to drive rhythmic mating behavior (Fujii et al., 2007), although earlier studies indicate that the female circadian clock also affects mating rhythms (Sakai and Ishida, 2001a). Mating rhythms are also found in cockroaches and moths,for example (Silvegren et al., 2005; Rymer et al., 2007).
Figure 1 Drosophila circadian locomotor behavior and gene expression. (A) Actogram showing the average locomotor activity of a group of Drosophila males. Individual flies are placed in small glass tubes with sugar food and monitored under a 12 : 12 light : dark (LD) cycle for 3 days, and then released in constant darkness (DD). Gray shading indicates the dark phase of LD or DD. The actogram is double-plotted: each day (except the first) is plotted twice, first on the right half of the actogram, and then on the left half on the next line. Full and empty arrows indicate the morning and evening peaks of activity, respectively, which are controlled by the circadian clock. Zeitgeber Time (ZT) 0 corresponds to the beginning of the light phase. (B) Schematic representation of the expression pattern of two key circadian genes (per and tim) and their protein products. Gray lines represent mRNA levels, and black lines protein levels.
Interestingly, the production and release of, and response to, sex pheromones can be regulated by the circadian clock (Silvegren et al., 2005; Merlin et al., 2007; Krupp et al., 2008). Another well-characterized circadian reproductive behavior is oviposition (egg-laying rhythms; for review, see Howlader and Sharma, 2006). Interestingly, the neural substrate regulating oviposition rhythms in Drosophila melanogaster is distinct from those controlling circadian eclosion (see below) and adult locomotor rhythms (Howlader et al., 2006).
After egg-laying, important behaviors related to the transition from one developmental stage to another can be driven by the circadian clock. For instance, egg hatching is gated by the circadian clock in many insects (see, for example, Minis and Pittendrigh, 1968; Lazzari, 1991; Sauman and Reppert, 1996). Metamorphosis is also frequently circadian. During insect metamorphosis, ecdysteroid hormones are synthesized by the prothoracic gland and released into the hemolymph to trigger adult ecdysis. The prothoracicotropic hormone (PTTH) plays an important role in this process by stimulating the prothoracic gland. Circadian control of PTTH release, PTTH responsiveness, ecdysteroid synthesis, hemolymph ecdysteroid titer, and ecdysteroid receptor expression have been described in Rhodnius prolixus and other insects (Cymborowski et al., 1991; Vafopoulou and Steel, 1991, 1996; Pelc and Steel, 1997; Richter, 2001; for review, see Steel and Vafopoulou, 2006). Interestingly, in Drosophila the pro-thoracic gland shows rhythmic autonomous expressions of clock proteins (Emery et al., 1997). Drosophila eclo-sion occurs preferentially around dawn, and this eclosion rhythm is absent in classic clock gene mutants (Konopka and Benzer, 1971; Sehgal et al., 1994). Drosophila eclo-sion rhythms require both brain pacemaker neurons (the pigment dispersing factor (PDF) positive ventral lateral neurons — see section 15.4) and a functional clock in the prothoracic gland (Myers et al., 2003).
Another behavior that is modulated by the circadian clock in Drosophila is larval photophobicity (Mazzoni et al., 2005). Drosophila larvae (all stages except for late third-instar larvae) prefer to avoid exposure to light if possible. The clock modulates the photosensitivity of this behavior: larvae are more sensitive to light in the early morning.
Sophisticated behaviors such as time-compensated sun compass navigation require the circadian clock to correct the direction of flight as a function of time, because the position of the sun changes during the course of the day. This plays an essential role in long-distance foraging and migratory behavior — for example, in bees (Lindauer, 1960; Bloch, 2010) and Monarch butterflies (Mouritsen and Frost, 2002; Froy et al., 2003; Reppert et al., 2010). Strikingly, the waggle dance performed by forager honeybees to indicate the location of food to other foragers is also time-compensated to correct for changes in the solar azimuth (Lindauer, 1954).
Physiological and Metabolic Rhythms
Circadian clocks are found not only in the brain but also in inner organs such as the Malpighian tubules and the gut, and in external sensory organs (e.g., antennae, eyes, and proboscis) (Zeng et al., 1994; Emery et al., 1997; Giebultowicz and Hege, 1997; Plautz et al., 1997; Krishnan et al., 1999; Merlin et al., 2007, 2009; Ito et al., 2008; Uryu and Tomioka, 2010). Interestingly, these peripheral clocks function independently of the brain pacemaker neurons, since isolated organs show self-sustained molecular circadian rhythms in culture (Emery et al., 1997; Giebultowicz and Hege, 1997; Plautz et al., 1997; Merlin et al., 2009). Peripheral clocks affect local physiology. For example, Drosophila melanogaster shows robust rhythmic physiological responses to odor and taste stimuli (Krishnan et al., 1999; Krishnan et al., 2008; Chatterjee et al., 2010). Flies are more sensitive to odors at night, and to gustatory stimulation at dawn. Electro-physiological recordings reveal that spike amplitude is modulated by the circadian clock in gustatory and olfactory receptors (Krishnan et al., 2008; Chatterjee et al., 2010). The circadian clock also controls spike duration and frequency in gustatory neurons. The rhythmically expressed G Protein Receptors Kinase 2 (GPRK2) kinase plays a particularly important role in these processes (Tanoue et al., 2008; Chatterjee et al., 2010). In antennae, it regulates the localization of olfactory receptors to the dendritic membrane of the olfactory sensory neurons, and probably works similarly for the gustatory receptors as well.
Drosophila’s rate of survival is higher if an injection of pathogenic bacteria is performed during the night than during the day. Interestingly, this greater survival rate is correlated with a stronger induction of specific antimicrobial peptides at night (Lee and Edery, 2008).
In many pest insects, resistance to insecticides has been found to be dependent on the time of day (Sullivan et al., 1970; Pszczolkowski and Dobrowolski, 1999). Interestingly, genes involved in detoxification were found to be rhythmically expressed in whole-genome expression profiling studies (reviewed in Wijnen and Young, 2006). The activities of specific xenobiotic metabolizing enzymes were confirmed to be expressed rhythmically, which probably explains why the efficiency of some insecticides varies significantly during the day (Hooven et al., 2009).
These examples illustrate the profound effect that circadian clocks have on the physiology of insects.
Whole-genome expression studies have identified other physiological and metabolic processes that might be under circadian regulation in Drosophila (Claridge-Chang et al., 2001; McDonald and Rosbash, 2001; Ceriani et al., 2002; Y. Lin et al., 2002; Ueda et al., 2002; Boothroyd et al., 2007; Keegan et al., 2007). These studies have for the most part focused on head tissues. It would thus be interesting to extend these studies to specific organs such as the Malpighian tubules to better understand the role played by peripheral circadian clocks.
The Drosophila Circadian Pacemaker
This topic focuses heavily on Drosophila circadian rhythms. For more than half a century, Drosophila has played a particularly important role in the field of chro-nobiology. In the early 1950s, Pittendrigh and colleagues began to use Drosophila eclosion (adult ecdysis) as a readout to understand the fundamental properties of circadian clocks, such as their free-running nature, their entrain-ment, and the phenomenon of temperature compensation (for reviews, see Pittendrigh, 1960, 1993). In the late 1960s, Konopka and Benzer had the groundbreaking idea of screening for single gene mutants affecting Drosophila melanogasters timing of eclosion, with the conviction that such screens would ultimately lead to a mechanistic dissection of the circadian clock. In 1971, they reported the identification of three mutants affecting a single locus, period (per) (Konopka and Benzer, 1971). The perS, perL, and per0 mutations accelerated, slowed down, or completely eliminated rhythms of eclosion. Adult locomo-tor rhythms were also affected in a similar way, which indicated that the per gene was probably a central element of a circadian pacemaker responsible for many, if not all, circadian rhythms in the fruit fly. Adult locomo-tor activity progressively replaced eclosion as a read-out for circadian clocks in genetic screens, because it allows monitoring of individual flies, while eclosion can only be monitored at a population level. Moreover, sophisticated forward genetic approaches based on the use of the lucif-erase reporter gene controlled by per regulatory sequences (Brandes et al., 1996) have also successfully identified cir-cadian genes. Thus, many of the genes that will be discussed below were identified through forward genetics. In the past decade, the completion of the Drosophila genome sequence and the development of methods to monitor whole-genome expression patterns have allowed reverse genetic approaches to become important alternatives for circadian gene discovery. Another major development over this period has been the design of genetic tools to manipulate specific circadian tissues or cell types. This has particularly impacted our understanding of the neural circuits controlling Drosophila circadian locomotor.
The PER Transcriptional Feedback Loop
PERIOD (PER) At the end of 1984, two teams, led by Michael Young at Rockefeller University and Michael Rosbash and Jeffrey C. Hall at Brandeis University, reported the cloning of per (Bargiello et al., 1984; Reddy et al., 1984; Zehring et al., 1984). This was a heroic effort: fragments of DNA from the region of the X-chromosome to which the per gene had been mapped were systematically integrated into the per0 mutant fly genome until normal circadian rhythms were restored. The sequence of the per gene was at first uninformative, and PER function remained unclear for another decade. It was only 9 years later that a domain common to PER and two transcription factors (the PAS domain) was identified, indicating that PER might also be a transcription factor (Huang et al. , 1993). Interestingly, it was evident by that time that PER was somehow involved in a transcriptional feedback loop (Hardin et al., 1990, 1992) (Figure 2). This conclusion was based on three fundamental observations: first, per mRNA levels oscillate robustly during the course of the light/dark (LD) cycle, and also under constant darkness (DD) (Figure 1); second, the period of per mRNA oscillations, or their absence, parallels the circadian behavior phenotypes of perS, perL and per0 mutants; and third, this regulation is in great part transcriptional, since rhythmic reporter gene expression is obtained with per promoter sequences. In addition, PER acts as its own gene repressor, since per mRNA levels are low when PER is overexpressed (Zeng et al., 1994). This basic feedback mechanism of a key circadian gene on its own gene expression has proven to be crucial for all known eukaryotic circadian clocks (Dunlap, 1999).
Critically, PER protein levels, its phosphorylation state, and its nuclear localization are rhythmic as well (Siwicki et al., 1988; Zerr et al., 1990; Zwiebel et al., 1991; Edery et al., 1994a; Vosshall et al., 1994; Curtin et al., 1995)(Figure 1). These rhythms determine when and how efficiently PER can function as a repressor. This is discussed in detail below.
Figure 2 The PER circadian feedback loop in Drosophila. CLOCK (CLK) and CYCLE (CYC) are two transactivators that bind to the E-box of the tim and per promoter. PER and TIM first accumulate in the cytoplasm and then enter into the nucleus to block their own gene transcription.
TIMELESS (TIM) timeless (tim) — the second circadian gene identified in Drosophila — is also an integral part of the per negative transcriptional feedback loop (Sehgal et al., 1994, 1995; Myers et al, 1995; see also Figure 2). tim and per mRNA levels cycle with a similar phase and amplitude (Figure 1). In addition, tim mRNA cycling is disrupted by per mutations, and tim mutations affect per mRNA cycling. TIM actually promotes PER function: it binds to PER, stabilizes it, and contributes to PER’s translocation into the nucleus (Vosshall et al., 1994; Gekakis et al., 1995). In TIM’s absence, flies are arrhythmic. TIM is also an important target for light-input pathways (Hunter-Ensor et al., 1996; Lee et al., 1996; Myers et al, 1996; Zeng et al, 1996; Mazzoni et al., 2005), and thus connects the circadian pacemaker with the outside world (see section 15.3).
CLOCK (CLK) and CYCLE (CYC) The promoter elements that drive per and tim mRNA cycles contain E-box binding sites, which can be recognized by bHLH/PAS family members (Sogawa et al., 1995; Swanson et al., 1995; Hao et al., 1997; McDonald et al., 2001). However, although PER carries a PAS domain, it has no identifiable bHLH DNA binding domain (Huang et al., 1993), suggesting that it does not regulate transcription through direct binding to the per and tim promoters. The proteins that bind to these E-boxes are actually two bHLH-containing members of the PAS family: CLOCK (CLK) and CYCLE (CYC) (Allada et al., 1998; Darlington et al., 1998; Rutila et al., 1998) (Figure 2). In the dominant-negative Clkjrk and the null cycC0 mutants, per and tim mRNA levels are dramatically low, indicating that both CLK and CYC function as positive regulators of PER and TIM. CLK and CYC can indeed bind as heterodimers to E-boxes present in the per and tim genes (Lee et al., 1999), and transactivate a luciferase reporter under the control of a per or tim promoter in S2 cells (Darlington et al., 1998). (The S2 cell line is a commonly used embryonic Drosophila cell line, which has proved immensely useful to understand the molecular mechanisms underlying Drosophila circadian rhythms; see also below.) Furthermore, CLK/ CYC transactivation potential is strongly inhibited by PER and TIM. The PER feedback loop is therefore closed (Darlington et al., 1998). Interestingly, in mammals CLK, CYC, and PER homologs are also key components of the circadian pacemaker (Dunlap, 1999). The mammalian and fly clocks are thus based on similar mechanisms, although mammals and flies are separated by at least 600 million years of evolution.
Mechanisms of PER repression The PAS domain is responsible for homo and heterodimerization between proteins containing these domains (e.g., CLK/ CYC) (Huang et al., 1993). Since PER contains a PAS domain but no DNA binding domain, it would appear likely that PER interferes with CLK/CYC transcriptional activity by interacting with the PAS domains of these proteins. However, when the repression domain for PER was mapped using a transcriptional reporter assay in S2 cells, it was found to lie in the C-terminus of PER (Chang and Reppert, 2003) (Figure 3). This domain was called CCID (CLK/CYC inhibitory domain), and it has recently been shown that it contains a small domain required for PER interactions with CLK (Sun et al., 2010). The PAS domain is completely unnecessary for repression, at least in the context of this cell culture assay. What, then, is the function of the PAS domain? First, the PAS region and the C-domain located C-terminal of it are responsible for interactions with TIM, which is important for PER stability (Vosshall et al, 1994; Gekakis et al, 1995). Of note, the perL mutation is located near the N-terminus of the PAS domain. Second, PER can form homodimers through its PAS domain and the C-domain (Huang et al., 1993; Yildiz et al., 2005; Landskron et al., 2009), although in vivo most PER is associated with TIM (Zeng et al., 1996). It has recently been proposed that these homodimers have an important functional role in the circadian clock (Landskron et al., 2009).
Figure 3 PER functional domains and phosphorylation sites. The red box shows the SLIMB binding domain of PER, the grey box the DBT binding domain, and the black box the CLK binding domain (CBD). Three orange boxes indicate the PAS-A, PAS-B, and PAS-C domains, necessary for TIM binding. The blue box indicates the CCID (CLK-CYC inhibitory domain). Arrows indicate the phosphorylated amino acids with demonstrated function for circadian time-keeping; the kinases that modify these sites are also indicated.
Single amino acid mutants were specifically designed to disrupt PER homodimer formation, based on crystallographic studies. The M560D mutant was studied in great detail. Flies expressing only this mutant PER showed very poor PER protein cycling, and were frequently behaviorally arrhythmic (about 60% of flies) or showed a slight period lengthening. Importantly, while PER homodimer formation was disrupted, PER/TIM heterodimerization was unaffected by the M560D mutation. These results support the idea that PER homodimers are important for circadian rhythms. However, there is no way to be certain that only PER homodimer formation was affected by the M560D mutation. Interactions with kinases (see below) could for example have been altered. It is also not clear how PER homodimers would mechanistically affect the circadian pacemaker. Further studies are thus necessary to clarify the role of PER homodimers.
How, then, does PER repress CLK? It could somehow interfere with CLK/CYC activation, either by blocking the recruitment of co-activators or by recruiting co-repressors, or it could displace CLK from the E-box. Electrophoretic mobility shift assays (EMSAs) gave an early indication that the second hypothesis was correct. Indeed, CLK/CYC affinity for the E-box is reduced in the presence of PER/ TIM in vitro (Lee et al., 1999). Furthermore, occupancy of the E-box was more recently shown by chromatin immu-noprecipitation (Chip) assays to be rhythmic, with CLK/ CYC being more strongly bound to the per E-box when per gene transcription is high, and being displaced from DNA when PER is in the nucleus and represses transcription (Yu et al., 2006). A recent study actually indicates that the two mechanisms proposed above might be correct (Menet et al., 2010). Indeed, Chip assays show that PER is first associated with chromatin along with CLK/ CYC at a time when per transcription decreases. It must thus somehow interfere with the transactivation of DNA-bound CLK/CYC. Then, PER remains associated with CLK/CYC when it is detached from DNA (Figure 4).
In vivo, TIM contributes to PER repression, since it stabilizes PER, as discussed above. Does it also play an active role in PER repression? In cell culture, PER repression does not require TIM (Rothenfluh et al., 2000; Nawathean and Rosbash, 2004), although it has been reported that TIM can promote PER repression (Darlington et al., 1998).
Figure 4 Mechanisms of PER repression. DBT binds to PER and phosphorylates it in the cytoplasm. TIM binds to the complex and stabilizes PER. DBT, PER, and TIM then enter into the nucleus (although PER and TIM probably do not translocate together; see text for details). After binding to CLK, PER starts to repress its own transcription. CLK and CYC are subsequently removed from the E-box, and per and tim transcription is shut down. CLK is phosphorylated by an unknown kinase (empty triangle), recruited by DBT, and degraded.
TIM might also be important for the timing of PER entry into the nucleus, and therefore for its repressive potential. Initial studies indicated that PER and TIM probably translocate together in the nucleus (Vosshall et al., 1994; Hunter-Ensor et al., 1996; Saez and Young, 1996). However, under certain circumstances, TIM is dispensable for PER nuclear entry – for example, in a tim°:dbtf double mutant in which PER levels are quite high (Cyran et al., 2005). Moreover, Shafer et al. showed that in circadian neurons that control circadian behavior — the ventral lateral neurons (LNvs; see section 15.4) — PER enters into the nucleus before TIM (Shafer et al., 2002). However, these results do not exclude the possibility that a small fraction of TIM shuttles between the nucleus and the cytoplasm to promote PER localization. TIM could also promote PER nuclear entry from a cytoplasmic localization. Strikingly, recent results using YFP and CFP tagged PER and TIM and FRET (Foerster (or fluorescence) resonance energy transfer) to visualize in cell culture the formation of PER and TIM dimers, and their nuclear entry, actually confirm that PER and TIM enter into the nucleus independently of each other, rather than together (Meyer et al., 2006). However, prior to PER nuclear translocation, a PER/TIM dimer is observed in perinuclear foci, which indicates that TIM is indeed probably important as a cytoplasmic catalyst of PER nuclear entry. PER might then form homodimers after dissociating from TIM, since the aforementioned homodimer-ization mutant M560D shows defective nuclear entry (Landskron et al, 2009). But if TIM enters into the nucleus independently of PER, then what is its nuclear function? One possibility is that it could still help stabilize PER late in the circadian cycle, when it becomes hyper-phosphorylated. Recent results indicate that TIM might also help to stabilize PER—CLK interactions. Indeed, if TIM is present, a mutant PER protein missing the CLK interaction domain can still repress CLK/CYC in S2 cells. The same mutant protein associates with CLK in vivo in the dark, but when light is present this complex is dissociated andper/tim transcription increases (Sun et al., 2010). This is explained by the light-dependent degradation of TIM (see section 15.3). TIM cannot, however, entirely compensate for the absence of the CLK binding domain of PER, since flies expressing this mutant protein have a long-period phenotype.
As will be discussed in detail in section 15.2.2.1, PER repression requires the kinase DOUBLETIME (DBT) — with which it is tightly associated — and probably additional kinases as well. A crucial role of PER in repression is thus to bring these modifying enzymes into the vicinity of CLK/CYC to trigger repression.
CLOCKWORK ORANGE (CWO) As discussed above, DNA microarrays probing the expression of the entire Drosophila genome were used to identify genes regulated by the circadian clock (Claridge-Chang et al., 2001; McDonald and Rosbash, 2001; Ceriani et al., 2002; Y. Lin et al., 2002; Ueda et al., 2002). One of them was cg171°°. With the use of an inducible CLK—Glucocorticoid receptor fusion, it was found in cell culture and in fly heads to be a primary target of CLK (Kadener et al., 2007). Indeed, its promoter contains E-boxes similar to those of per and tim. CG17100 was renamed CLOCKWORK ORANGE (CWO) (Figure 5) because it is an Orange domain containing bHLH protein (Lim et al., 2007). CWO binds E-boxes, and represses both basal E-box-containing promoter activity and CLK-mediated transactivation in cell culture (Kadener et al., 2007; Lim et al., 2007; Matsumoto et al., 2007). Surprisingly, however, mRNA oscillations of most CLK/ CYC regulated genes show a blunted peak in cwo loss-of-function mutants (Kadener et al., 2007; Lim et al., 2007; Matsumoto et al., 2007; Richier et al., 2008). Peaks of mRNA cycling are reduced for per and tim, and for two other CLK/CYC targets that will be discussed in section 15.2.4 (pdple and vri). For some of these genes, trough levels are nevertheless increased as well. In sharp contrast, cwo mRNA is constantly high in cwo mutant flies. Taken together these results indicate that different CLK/CYC target genes are differentially sensitive to CWO. CWO might be predominantly a repressor in the case of its own gene regulation, while on other circadian genes it might both inhibit transcription at certain times of the cycle, and promote full activation at other times (Figure 5). It is conceivable that CWO contributes to the recruitment of activators that will help in opening chromatin, and promote the return of CLK/CYC on the promoter. However, how CWO is being displaced from chromatin is unclear. CWO protein levels do not cycle robustly (only about two-fold in the circadian neurons regulating behavior), but CWO could be modified by kinases, for example. Alternatively, it might be the state of CLK phosphorylation — regulated by PER and DBT (see section 15.2.2.1) — that determines its DNA binding affinity, and thus whether it is CLK/CYC or CWO that is bound to the promoter.
Behaviorally, severe hypomorphic cwo mutants, or even null mutants, are rhythmic. However, rhythms have a long period and amplitude is affected to diverse degrees, depending on the genetic manipulations that were used to reduce CWO levels (RNA interference (RNAi), P-element insertion, null mutation) (Kadener et al., 2007; Lim et al., 2007; Matsumoto et al., 2007; Richier et al., 2008). Most significantly, in null mutants rhythms were actually only slightly reduced in amplitudes, and persisted for at least 10 days (Richier et al, 2008). CWO therefore acts as a modulator of circadian rhythms, important for setting the proper pace of the circadian clock, but it is not required for circadian rhythms to occur and to persist.
Figure 5 Interlocked feedback loops. The PER feedback loop is modulated by cwo, which is itself regulated by this loop. CWO is a transcriptional repressor, but it also appears to be necessary for full activation of a subset of CLK/CYC regulated genes (dotted arrow). A second loop is attached to the PER loop. VRI and PDP1 generate circadian transcriptional rhythms peaking early during the light phase (left), in antiphase of those generated by the PER loop (right).
It should be noted that there are other members of the bHLH/Orange family. It is thus possible that one of them could partially substitute for CWO (Matsumoto et al., 2007), but whether this indeed happens has not yet been reported.
