A comparative view of insect circadian clock systems

Abstract

Recent studies revealed that the neuronal network controlling overt rhythms shows striking similarity in various insect orders. The pigment-dispersing factor seems commonly involved in regulating locomotor activity. However, there are considerable variations in the molecular oscillatory mechanism, and input and output pathways among insects. In Drosophila, autoregulatory negative feedback loops that consist of clock genes, such as period and timeless are believed to create 24-h rhythmicity. Although similar clock genes have been found in some insects, the behavior of their product proteins shows considerable differences from that of Drosophila. In other insects, mammalian-type cryptochrome (cry2) seems to work as a transcriptional repressor in the feedback loop. For photic entrainment, Drosophila type cryptochrome (cry1) plays the major role in Drosophila while the compound eyes are the major photoreceptor in others. Further comparative study will be necessary to understand how this variety of clock mechanisms derived from an ancestral one.

Introduction

Circadian rhythms of about 24-h periodicity are commonly observed in a variety of physiological functions of organisms from bacteria to humans. The rhythm is governed by a mechanism called the circadian system, generally including a circadian clock that generates a 24-h oscillation, a photoreceptor that is necessary for the clock to synchronize to the light–dark cycles (LD), and an output system that relays the clock information to various tissues to regulate their overt rhythmicity [1, 2].

The rhythm has some common, basic characteristics: it persists for long periods in constant conditions; its period is temperature-compensated and maintains a similar value over a wide range of temperatures; and it synchronizes to daily environmental cycles, using light and temperature as major synchronizing agents or zeitgebers. Insects provide good models for dissecting the circadian system both at the physiological and molecular levels. The circadian clock that regulates the overt behavioral rhythms has been localized in the brain. For example, the optic lobe is the clock locus for crickets, cockroaches, and beetles [3–5], while the central brain is known for moths, flies, and mosquitoes [6–8]. The compound eyes are known as the principal circadian photoreceptor for photic entrainment in most insects, but ocelli and the extraretinal photoreceptors also play a role in others (see [2]).

Recent studies have promoted our understanding of the circadian system at molecular and cellular levels. The circadian oscillatory mechanism has now been uncovered at the molecular level, and the circadian network for generating rhythmicity has been elucidated in detail in fruit flies [9–12]. Molecular and cellular dissection has also been started in other non-drosophilid insects, promoting the comparative and general formulation of the underlying mechanism [13]. This review will focus on our current understanding of the molecular and cellular mechanism of the insect circadian system.

The circadian clock network

Cellular organization of the clock has been extensively studied in Drosophila and other insects, using molecular probes, such as antibodies against clock gene products and reporter genes driven by clock gene promoter, etc. In Drosophila, there are about 150 cells in the brain that express the clock genes, and they are divided into seven subgroups [10, 14] (Fig. 1a). Three of them are located laterally between the optic lobe and the central brain, and are called lateral neurons (LNs). The ventrally located LNs consist of two subgroups of neurons with larger and smaller cell bodies and are called large-LNv (l-LNv) and small-LNv (s-LNv), respectively. The dorsally located LNs are called LNd. There are three groups of neurons in the dorsal region of the central brain called DN1, DN2, and DN3, respectively. The other group is located on the lateral posterior side of the central brain, and is called lateral posterior neurons (LPNs). Detailed analysis of the role of the cells revealed that s-LNvs drive the morning peak and a part of the evening activity, and the LNds regulate the evening peak [15–17]. Recently, l-LNv cells have been revealed to be involved in light-induced arousal and phase-shifting in the late night [18]. DNs, however, have some roles in locomotor rhythms, since the flies lacking LNs still show a circadian locomotor rhythm for at least a few days in constant conditions [19–21]. In larval brain, DN2 cells have the principal role in temperature entrainment and might regulate the phase of LNs, while under LD, the larval DN2 s are controlled by LNs via PDF signaling [22].

Fig. 1.

Schematic diagrams showing Drosophila (a) and cockroach (b) cerebral cells expressing PERIOD (redrawn from [23, 125]). a In Drosophila, there are seven main groups of neurons. Three groups (DN1, DN2, and DN3) are located in the dorsal region, and the remaining four (LNd, l-LNv, s-LNv, and LPN) are located laterally. The s-LNv and l-LNv express a neuropeptide, PDF. l-LNvs have their processes in the optic lobe and send their axonal projection to the contralateral optic lobe through the posterior optic tract (POT). s-LNvs have axonal projection to the dorsomedial region of the protocerebrum. b In the cockroach, Blattera germanica, three groups of neurons are located in the optic lobe (G-I, G-II, and G-III) and three groups in the dorsal protocerebrum (G-IV, G-V, and G-VI). There is another group (G-D) in the deutocerebrum. PDF is coexpressed in G-I, -II, -III, -IV and -D cells. OL Optic lobe, Pr protocerebrum, De deutocerebrum

Virtually similar clock structures are reported for other insects, including the blow flies, crickets, and cockroaches [23–25]. In the cockroaches and crickets, some of the PER-immunoreactive neurons are located in the optic lobe (Fig. 1b): These probably correspond to the LNvs in Drosophila. Their importance for overt activity rhythm generation is implicated by various surgical and electrophysiological experiments [4, 26–28]. Like in DNs of Drosophila, the neurons located in the dorsal protocerebrum may also play a role in the control of overt activity rhythms, because some residual rhythms are often observed even after optic lobe removal [29], and temperature or light cycles often induce a rhythm with circadian properties [30–32].

Molecular machinery of the circadian clock

Irrespective of the similarity in the neuronal architecture of the circadian clock system, recent studies have revealed considerable differences in the molecular mechanisms among insects.

Drosophila’s clock machinery

The molecular machinery of the circadian clock has been extensively studied in the fruit fly, Drosophila melanogaster. There are several so-called clock genes involved in the generation of the rhythmicity [9, 13]. At least three interdependent feedback loops are thought to constitute the rhythm-generating machinery (Fig. 2). One major loop is that formed by period (per), timeless (tim), Clock (Clk), and cycle (cyc) (cf. [9, 33]). The transcription of per and tim is activated by product proteins of Clk (CLK) and cyc (CYC). CLK and CYC form a heterodimer and bind to a promoter region, called E-box, of per and tim and promote their transcription during late day to early night. The product proteins, PER and TIM, increase during the night to peak at late night, when their heterodimeric complex enters into the nucleus to repress their own transcription through inhibitory action on CLK-CYC. The repression of per and tim transcription results in reduced level of PER and TIM, which eventually releases CLK-CYC from the repression to reactivate the transcription of per and tim. This reactivation leads the loop to the next round.

Fig. 2.

A scheme illustrating current knowledge about molecular machinery of the Drosophila circadian clock. The circadian clock consists of three feedback loops that are interlocked to oscillate with a period of about 24 h. See text for details

The second loop is for a circadian oscillation of Clk [34, 35]. Besides per and tim, CLK-CYC activates transcription of vrille (vri) and PAR domain protein 1ε (Pdp1ε) during late day to early night. The vri mRNA is soon translated to its product protein VRI, which enters the nucleus and inhibits transcription of CLK by binding to a promoter region, V/P-box. Thus the Clk mRNA is reduced during the night. The translation of Pdp1ε occurs in a rather delayed manner and PDP1ε protein increases during late night to early day. PDP1ε is thought to bind to V/P-box competitively with VRI and activate the transcription of Clk. Thus, the Clk transcripts increase during the day, also leading to an increase of CLK protein during the day. But a recent study revealed that PDP1ε is not essential for the operation of the loop [36].

The third loop includes clockwork orange (cwo), which is a transcriptional repressor belonging to the basic helix-loop-helix ORANGE family [37–40]. cwo is rhythmically expressed to peak under the regulation by CLK-CYC and forms its own negative feedback loop. CWO represses the expression of other clock genes, such as per and tim, through E-box elements. This negative feedback loop contributes to sustaining a high-amplitude circadian oscillation.

For precise temporal control of the molecular oscillation and for function of the molecular oscillator, posttranslational regulation of clock proteins seems crucial [9, 41, 42]. For example, the cyclic activation and inhibition of CLK/CYC is paralleled by cyclic phosphorylation levels [43]. Phosphorylation may regulate the timing of nuclear entry of PER and TIM [44, 45], as well as the degradation of TIM [46]. These posttranslational modifications are suggested to regulate the period, phase, and amplitude of the circadian clock [41].

Recent studies showed that in Drosophila, electrical silencing of clock neurons stops the circadian locomotor rhythm free-running in constant darkness (DD) [47]. Immunohistochemical studies revealed that this results from a cessation of the clock itself caused by absence of nighttime nuclear transport of TIM protein from the cytoplasm [47]. A similar stop of the clock was found in larvae [48]. Thus, the electrical activity is thought to be an essential element of the molecular clock in the cerebral pacemaker neurons. However, the molecular oscillation persists under LD in the flies with electrical silencing of clock neurons, suggesting that there is a light-dependent drive on the molecular clock that at least partially compensates electrical activity in LD [47].

Results in other insects do not fully support the Drosophila hypothesis

The progress of clock study in Drosophila has promoted research on molecular machinery of the clock in other insect species [13, 49]. The studies, however, revealed some significant discrepancies between Drosophila and others. In the silkmoth, Antheraea pernyi, and the cockroach, Blattera germanica, immunohistochemisty with anti-PER and anti-TIM revealed that these proteins always stay in the cytoplasm and never enter into the nucleus [23, 50]. Instead of negative feedback of product proteins, rhythmically expressed per antisense RNA is proposed to produce the rhythmicity in A. pernyi [51]. However, immunohistochemistry may only reveal large amounts of the proteins, and a trace amount may enter into the nucleus and repress transcription [52]. Another explanation may be that, instead of PER-TIM, CRY2, which is a mammalian type CRY [53], enters the nucleus and works as a transcriptional repressor [54, 55]. These still need to be experimentally examined.

Recently, reverse genetic approaches have been applied to those non-model insects. In many of them, cDNAs of canonical clock genes have been cloned and their expressions have been examined. In most of the tested insects, the clock genes, per, tim, and Clk, are shown to cycle in a daily or circadian fashion [50, 55–57]. Functions of some clock genes have been examined in a cultured cell system in the monarch butterfly, Danaus plexippus [55], and CRY2 seems to repress the transcriptional activity of CLK-CYC in vitro [55]. cry belongs to a photolyase family and plays a role in photic entrainment in Drosophila [58, 59]. Molecular evolutionary studies have revealed that gene duplication and loss have resulted in three modes of cry gene expression [54]; in some insects, either cry1 (Drosophila type cry) or cry2 (mammalian type cry) is expressed, and in others, both cry1 and cry2 are expressed (Table 1). In the honey bee, Apis merifella, tim does not exist in the genome, and cry2 is thought to act as a transcriptional repressor like in Danaus [60]. Thus, there seem to be considerable differences in the molecular oscillatory mechanism among insects.

Table 1.

Two types of cryptochrome in insects

Order	Species	Drosophila type (cry1)	Mammalian type (cry2)
Lepidoptera	Antheraea pernyi	AF333998	EF117812
	Danaus plexippus	AY860425	DQ184682
	Mamestra brassicae	AY947639
	Sesamia nonagrioides	DQ243704	DQ243705
	Spodoptera littoralis	EF364035	EF396286
Diptera	Anopheles gambiae	DQ219482	DQ219483
	Culex quinquefasciatus	B0WRR9	XM_001869421
	Bactrocera tryoni	AY708049
	Drosophila melanogster	NM_169852	DNE
	Sarcophaga bullata	FJ373353
	Sarcophaga crassipalpis	AB079536
Hymenoptera	Apis mellifera	DNE	EF117814
	Bombus impatiens		EF110521
	Nasonia vitripennis		XM_001606355
Coleoptera	Tribolium castaneum	DNE	EF117815
Hemiptera	Acyrthosiphon pisum	XM_001944367	XM_001950658
	Riptortus pedestris		AB379863
Orthoptera	Dianemobius nigrofasciatus	AB291231

GenBank accession number is indicated. DNE Does not exist

In crickets, the role of clock genes has been examined in a more direct way using RNA interference (RNAi) [61–63]. Like other insects, the expression of cricket’s per gene cycles in a daily fashion in both LD and DD, and peaks at early night. Double stranded RNA (dsRNA) of per knocks down per mRNA levels, and disrupts both its cycling and the locomotor activity rhythm [61]. Functional analysis with RNAi of other clock genes will clarify the molecular oscillatory mechanism in the cricket. However, so far the molecular dissection of the oscillatory mechanism has not been performed in enough species to illustrate its divergence.

Input pathways

Light entrainment

Light is the most important zeitgeber for the circadian clock to synchronize to the environmental cycles. Photoreceptors necessary for the entrainment, or resetting, of the clock have been localized in some insects. In hemimetabolous insects such as crickets and cockroaches, the compound eye is the most important photoreceptor [28, 64], but ocelli also sometimes have minor roles in the photic entrainment [65, 66]. In Drosophila, in addition to these external eyes, H–B eyelet that is a remnant of larval eye, and the blue light receptor molecule CRY1, are known for circadian photoreceptors [66].

In Drosophila, CRY1 is expressed in a circadian fashion in some of the cerebral clock neurons, and shifts the phase of the clock light-dependently [58, 67, 68]. It is known to bind TIM in a light-dependent manner and to lead to TIM’s degradation [68]. This light-dependent degradation of TIM results in a phase shift of the clock in a phase-dependent manner: A reduced level of TIM in the evening leads to a delay of the clock at certain times until TIM increases to a required level, while it causes an advance of the clock in the morning to the time at which the TIM’s level coincides [69, 70]. The phase-dependent phase shifts provide a base of non-parametric entrainment. Danaus CRY1 partially rescued light-induced phase shiftability in Drosophila cry ^b mutant flies, while CRY2 did not [55]. Experiments using RNA interference revealed that, in DpN1 cells, the light-induced decrease in TIM abundance is mediated through Danaus CRY1 [53]. Thus, CRY1 seems to mediate the light entrainment in the monarch butterfly.

Besides CRY1, rhodopsins (Rhs) are also involved in photic entrainment. In Drosophila, Rh1 and Rh6 were implicated in entrainment to red light [71], and Rh1, Rh5, and Rh6 to green and yellow light [72]. Although histamine is suggested to be an important neurotransmitter for the Rh pathway [73], other neurotransmitters and the intracellular signal pathways have yet to be elucidated. Details of what role photoreceptor molecules play in photic entrainment of other insects and their divergence among insect species deserve to be investigated.

Temperature entrainment

Temperature is also an important time cue to synchronize the clock, especially in dark places, such as eggs laid underground or insects that pupate in darkness. Temperature cycle or temperature steps are known to cause phase shifts of the clock to adjust to the time of environmental cycles [74, 75]. The molecular mechanism for the temperature synchronization is less understood. In Drosophila, the effects of temperature step-up and -down were examined at transcript levels [76]. Under constant light, temperature step-up up-regulates the Clk gene, and step-down down-regulates it. In contrast, per, tim, vri, and Pdp1ε genes are down-regulated and up-regulated by step-up and step-down, respectively. After temperature step-down, all these clock genes showed an oscillation in constant conditions. These temperature responses are virtually eliminated in Clk ^Jrk mutant flies, suggesting that the Clk gene is the principal component for temperature entrainment [76]. Although the temperature sensing system is still largely unclear, it might involve nocte and PLC, mutation of which disrupts temperature entrainment [77]. PLC was shown to have a role in temperature-dependent per splicing [78, 79]: Under low temperature conditions, splicing of an intron in the 3′ untranslated region of per is enhanced, leading to a rapid accumulation of PER protein and an advance of the phase of the evening activity [80]. In addition, tim mRNA levels are increased earlier at low temperatures through photic stimulation of tim expression [81].

Existence of temperature entrainable oscillators has been reported for other insects including crickets [75], cockroaches [31], and mosquitoes [82]. Unfortunately, no detailed investigation on molecular and cellular mechanisms of temperature entrainment is available for insects other than Drosophila.

Output pathways

The pathways through which the clock regulates overt rhythms are not well understood. Pigment-dispersing factor (PDF) is the best characterized output molecule of the circadian clock system. Its involvement was first notified when immunohistochemistry revealed that PDF colocalized with PER in some clock neurons, s-LNv and l-LNv, in the Drosophila brain (Fig.1) [83], and is now thought to be an output neurotransmitter of the circadian clock in a wide variety of insects. It is a member of the pigment dispersing hormone (PDH), an octadeca peptide that was first found in crustacea as a hormone that disperses pigments in the epithelium [84]. pdf gene expression shows no circadian rhythm in Drosophila [85], while the dorsomedial termini of PDF-expressing neurons show daily morphological changes probably controlled by the circadian clock [86]. pdf ⁰¹ mutant flies that lack PDF show a rhythm with advanced onset of the evening activity and become arrhythmic within a few days in DD [87]. PDF is required to adjust cycling amplitude, period, and phase in a variety of clock neurons in the brain [88–90] in accordance with the expression of PDF receptors in those neurons [91–93]. Also, like in Drosophila, PDF in cockroaches is both a phase regulator and a locomotor rhythm driver [94, 95]. In crickets, however, since the partial destruction of the optic lobe leads to arrhythmic locomotor activity without elimination of the PDF neurons, PDF may play only minor roles in the behavioral rhythm [96]. Besides regulation of the activity rhythms, PDF plays an important role in the sensory system of crickets. When injected into the optic lobe, PDF is revealed to increase the photoresponsiveness of the optic lobe visual interneurons during early night [97]. In fact, the optic lobe content of PDF increases during the evening [98]. The PDF-induced increase of photoresponsiveness might be associated with the morphological changes reported for flies [99], because the rate of impulse propagation depends on the axonal diameter in unmielinated neurons.

Pdp1ε and takeout (to) have been located downstream of the circadian clock in Drosophila. When the PDP1ε level was kept constantly higher or lower in clock neurons, the clock runs normally but locomotor activity rhythms are disrupted [36]. Thus, PDP1ε seems to drive behavioral rhythms when some specific genes are controlled by its binding to V/P-box. to is expressed in the inner part of cardia and the crop, under the regulation by the circadian clock, in response to starvation. The to expression is arrhythmic in arrhythmic mutants, per⁰¹, tim⁰¹, Clk^Jrk and cyc⁰¹. to mutant flies show aberrant locomotor activity and die quickly upon starvation. Based on these results, to is concluded to be a clock-controlled output gene controlling feeding behavior [100]. Recently, this link between TO and feeding is suggested to occur through juvenile hormones [101]. Involvement of Pdp1ε and to deserves to be examined in other insects. Besides these, mRNA levels of hundreds of other genes have been shown to oscillate [39, 102]. Some of these genes may be involved in the output pathways of the clock, inducing a structural daily remodeling of the output circuit [86].

Peripheral oscillators

It has been reported many times that circadian oscillators exist in various tissues of insects (see [2]). For example, physiological studies revealed that sperm release from testis to the associated vas deferens, as well as ecdysone secretion from the prothoracic gland, was controlled in a circadian manner by peripheral oscillators in moths (see [103]).

In Drosophila, per-driven luciferase activity in a transgenic reporter strain showed that per is rhythmically expressed in cells of the compound eye, antennae, proboscis, wings, and legs [104]. Basically, similar results were observed in cells of the ring gland and Malpighian tubules (Mts), which are tissues for endocrine and excretory activity in flies, respectively [105, 106]. Rhythmic expression of per in these cells continues in vitro and is able to entrain to the light–dark regime of the culture condition. Transplanted Mts maintain their original phase of oscillation even in the host flies which have been entrained to the reverse light–dark cycle to the donor [107]. Thus, these peripheral oscillators seem to function as a stand-alone pacemaker. The same is true for the olfaction rhythm as an output of the antennal clock [108], and for the circadian synaptic plasticity rhythm, which was recently found as the rhythmic change of size in synaptic boutons [86, 109].

These peripheral tissues are thought to have a similar but not identical molecular machinery of circadian oscillation shown in the central pacemaker described above. One remarkable difference is CRY1’s function. A loss of function mutant of CRY1, cry ^b, showed intact molecular oscillations of PER and TIM in s-LNv, but lost the molecular oscillation in Mts [110]. In the antennae, the rhythmic electroantennographic (EAG) responses to odorants were also abolished in the mutants [111]. These results indicate that, instead of a photoreceptor in the central pacemaker, CRY1 functions as a core component in these peripheral oscillators [112]. Unlike these, however, another peripheral rhythm, which was recently found in the cuticle deposition, persisted in cry mutant flies [113]. Thus, there are at least two kinds of circadian molecular machinery in peripheral oscillators: one that requires CRY1 as an essential component, and the other that includes it just as a photoreceptor.

The Drosophila antennal clock seems to be one of the best models to investigate how peripheral oscillators control output phenomena at the molecular and behavioral levels. Rhythmic expressions of clock gene products have been observed in the antenna of tephritid fruit flies [114] and moths [115–117]. Thus, the molecular machinery of antennal oscillator seems common principally in insects. Detailed studies in Drosophila have revealed that rhythmic expression of G-protein receptor kinase, regulated by the circadian clock, controls the EAG amplitude in antennae [108, 118, 119]. Although circadian EAG rhythms have also been reported for cockroaches [120] and moths [121, 122], the regulatory mechanism differs substantially from that of Drosophila described above. The EAG rhythms in cockroaches are driven by the central pacemaker, even though individual olfactory receptor neurons exhibit circadian rhythms independent of the central pacemaker [123]. Circadian pheromonal responses in male turnip moths are not controlled at the antennal level [121]. In hawkmoths, octopamine and tyramine rhythmically modulate the sensitivity of olfactory sensilla for pheromone-sensitivity, but have no effect on EAG amplitudes [124]. Thus, it is noteworthy that the scenario illustrated in Drosophila cannot always be applied in other insects.

Conclusions

Recent physiological and molecular studies on insect circadian rhythms considerably promoted our understanding of the underlying cellular and molecular mechanisms. The cellular network of circadian clock systems shows virtually similar structures among insects, and PDF is commonly involved as an output neuromodulator/neurotransmitter of the system. The molecular oscillatory mechanism is considered to consist of autoregulatory negative feedback loops based mainly on the study of Drosophila. However, with the recent knowledge on other insect species, the structure of the feedback loop should be reconsidered. Although most insects examined possess common canonical clock genes, some have additional genes and some lose certain genes. Perhaps the most striking finding is the involvement of cry2 as a transcriptional repressor [53, 54]. In this context, CRY2 might enter the nucleus in insects where nuclear entry of PER and TIM was not detected by immunocytochemistry, and this deserves to be addressed in future studies. In addition, CRY1 may play a major role in photic entrainment in dipteran and lepidopteran species, but rhodopsins in the external photoreceptors seems to be the major photoreceptor in orthopteran and dyctiopteran species. The molecular mechanisms for photic entrainment through the external photoreceptor pathway remain to be examined. The insect world is estimated to include nearly one million species with highly diverged morphology and physiology to adapt to various environments. Thus, it is likely that the circadian clock system also diverged from an ancestral one, in respect to the core oscillatory mechanism, as well as the photic entrainment pathways. Although information is only available for quite limited orders of the class insecta, future comparative studies should provide an insight for the general understanding of the insect clock mechanism as well as how the variety of clocks have diversified from an ancestral one.