Molecular Mechanism of the Circadian Clock

Abstract

Nearly all organisms possess a circadian clock, a genetically determined device that generates endogenous oscillations with a period of approximately 24 h. From a molecular perspective, the circadian clock relies on negative transcription-translation feedback loops. In insects, the molecular and genetic basis of the circadian clock machinery has been revealed by the remarkable genetic tools available to the fruit fly Drosophila melanogaster. However, the dawn of reverse genetics methods applicable to nonmodel species has led to recent significant advances in our understanding of the circadian clock beyond Drosophila. To illustrate the molecular mechanism behind the insect circadian clock, the first section focuses primarily on Drosophila melanogaster as the best established and most detailed insect model. Conserved components of the insect clocks are then identified at the genetic level, and lineage-specific idiosyncrasies and variations in setup are highlighted and further discussed. Functional evidence from non-Drosophila insects is reviewed, and the main descriptive data from molecular biology are presented in an evolutionary context and briefly summarized.

1 Introduction

Before we dive into the details and setup of molecular machinery driving the circadian clock in various insect species, the general background should be presented. The purpose of the biological clock ticking with a period of approximately 24 h lies on the time scale quite far from the time needed for typical biochemical processes, such as transcription, translation, protein phosphorylation, dephosphorylation, and degradation. Thus, a set of multiple interlocked loops, including some or even all of the above-listed regulatory processes, are found in the circadian clocks of various systems. It is unclear whether the circadian clock has evolved independently in kingdoms several times or if one ancestral clock has been heavily modified in different lineages of organisms during their evolution. At least, it is safe to say that we can find rather differently built clocks in cyanobacteria, plants, fungi, protozoa, and bilaterian animals (reviewed in Dunlap 1999). In cyanobacteria, the elegant phosphorylation cycle among just a few kinases can stably run even in vitro. Plant, fungal, and bilaterian clocks utilize a combination of interlocked feedback loops with transcription factors (also known as positive components) in the center that drive the expression of multiple genes, including the negative elements.

In Bilateria, two excellent models shaped our knowledge when research on one synergistically supported and motivated research on the other. First, Drosophila genetic tools lead to seminal discoveries of the period (per) and timeless (tim) genes. Then, new clock genes were identified in flies and mice, gradually building up the picture of the conserved animal clock setup with several specific feature characteristic of the fly and feature and component characteristic of the mouse clock. However, with the growing genomic sequencing, it became clear that the picture is not that simple and even some insect species remarkably differ from Drosophila. As functional genetic research further expanded, we can now see that insect clocks are more colorful than might have been expected at first. After all, with more than 400 million years of evolution and with millions of species known today, I would consider insect evolution as a remarkable collection of various solutions to the same or slightly different problems, a beautiful selection experiment for which the notes are not available. To illustrate the molecular mechanism behind the insect circadian clock, I will first focus mostly on the Drosophila melanogaster circadian clock as the best established and most detailed insect model (although, in some cases, the comparison to other insects will be provided immediately). Then, the conserved components of the insect clocks are identified, and lineage-specific idiosyncrasies are highlighted and further discussed.

2 Clock Setup in Drosophila

Unprecedented genetic tools predispose D. melanogaster to be a powerful model for gene discovery and elucidation of underlying molecular mechanisms. There are several excellent and detailed reviews on the Drosophila circadian clock and the mechanism involved in its regulation (Ozkaya and Rosato 2012; Hardin 2011; Peschel and Helfrich-Förster 2011; Stanewsky 2003; Hall 2003; Tataroglu and Emery 2015; Lim and Allada 2013a). A seminal screen by Ronald Konopka led to the identification of the per gene at the beginning of the 1970s (Konopka and Benzer 1971), whereas the gene was mapped by positional cloning more than a decade later (Zehring et al. 1984; Bargiello et al. 1984). The key feature of per regulation was a cyclical abundance of its transcript oscillating with a period of 24 h and, as discovered by Siwicki et al. (1988), the cyclical abundance of the PER protein, which peaked with a few-hour delay after the per mRNA reached its maximum. See Fig. 4.1a for an illustration of the expression pattern in the whole head extracts. The cyclical transcription implied a feedback loop (Hardin et al. 1990); however, a simple transcription-translation feedback loop (TTFL) should produce oscillations on a scale of minutes. Thus, additional regulatory steps should be involved in TTFL participating in the circadian clock.

Fig. 4.1

Drosophila-negative transcription-translation feedback loop (TTFL). (a) Relative abundance of per and tim-d mRNAs compared to PER and TIMd (Drosophila-type TIM) proteins illustrates oscillation in abundance and a delay between the accumulation of mRNA and protein in whole head extracts (redrawn from Dunlap 1999). (b) Schematic depiction of TTFL with major steps and regulatory mechanisms, illustrating the activity and inhibition of a hypothetical transcription factor (TF). (c) In the early evening, CLK-CYC drives transcription from per and tim-d genes. Activation domain (AD) is located on CLK protein in Drosophila. The resulting mRNAs are translated, but the proteins are (mostly) degraded. (d) As the night progresses, PER and TIMd proteins are stabilized by heterodimerization, although PER-PER homodimers are also observed. (e) During the late night, PER and TIMd enter the cell nucleus, where PER molecules inhibit the transcription factor CLK. (f) As PER and TIMd are eventually degraded, the inhibition fades away, and a new round of transcription starts in the morning

Figure 4.1b illustrates the general principle of the negative TTFL with major conceivable regulatory mechanisms adjusting the period of its oscillation. First, a transcription factor(s) drives the expression of mRNA coding for the negative element, which, before translation, needs to be processed, and its stability might be further regulated. To prevent degradation, the negative element protein(s) might be stabilized by modification steps such as phosphorylation, by interaction with partner protein(s) forming either the heterodimer or homodimer, and even by additional modifications (glycosylations, SUMOylation, etc.). The last step of the feedback is the interaction of the negative element with the transcription factor(s), preventing its activity. As a consequence, no mRNA coding for the negative element is transcribed. Transcription can only start once the nucleus-localized negative element protein is removed from the system. Thus, protein stability and subcellular localization are key regulatory steps of the whole process leading to 24-h oscillations.

2.1 Period and Drosophila-Type Timeless

In Drosophila, the first well-described TTFL comprises proteins PER and TIM, here referred to as Drosophila-type tim-d (TIMd for the protein) to distinguish it from its paralogous mammalian-type (tim-m, TIMm). Similar to per/PER expression, tim-d and TIMd oscillate in abundance in a 24-h cycle, when mRNA peaks circa 4 h before TIMd (Fig. 4.1a). Furthermore, the subcellular localization pattern of PER and TIMd changes over time, all of which led to the establishment of the following model. Let us focus on the most important clock neurons in the fly brain responsible for behavioral rhythmicity (the expression might differ in various cell types across the body; see Chap. 5). During the early evening (Fig. 4.1c), per and tim-d mRNAs are transcribed; however, the corresponding proteins are not detectable, as they are degraded. Then, during the early night, PER and TIMd are stabilized by dimerization, and both proteins are detected in the cytoplasm (Fig. 4.1d). Gradually, PER and TIMd accumulate, and then, during late night (Fig. 4.1e), both proteins are detected in the nucleus. At the same time, transcription of per and tim-d mRNAs stops because PER interacts with the transcription factor CLOCK (CLK). CLK, a basic helix-loop-helix (bHLH) Per-ARNT-Sim (PAS) protein, forms a dimer with another bHLH-PAS protein CYCLE (CYC), and together they drive expression from the so-called E-box (cis-regulatory DNA elements with motif CACGTG). E-boxes are localized in the promoters of multiple genes, including per and tim-d, and are key for their cyclical expression. The inhibition of CLK-CYC continues as long as PER is present in the nucleus. After its depletion (Fig. 4.1f), the transcription of per and tim-d mRNAs resumes again.

The above-described model has been synthesized from data gathered on mutant and wild-type flies (Hardin et al. 1990; Siwicki et al. 1988; Zerr et al. 1990; Sehgal et al. 1994; Sehgal et al. 1995; Allada et al. 1998; Rutila et al. 1998; Glossop et al. 1999). The feedback mechanism was further reconstructed in Drosophila Schneider 2 (S2) cell cultures (Darlington et al. 1998), and the role of the E-box was addressed both in S2 cells and in flies using elegant in vivo luciferase reporters (Brandes et al. 1996; McDonald et al. 2001). The protein domains important for PER-TIMd interaction and nuclear localization were systematically explored in S2 cells (Saez and Young 1996; Saez et al. 2011), even though real-time monitoring in S2 cells using FRET (fluorescence resonance energy transfer) revealed that the spatial organization of the dimer differs in the cytoplasm and nucleus (Meyer et al. 2006). Furthermore, functional PER-PER homodimers exist in flies, and if their mutual dimerization is impaired, molecular and behavioral rhythmicity is severely disrupted (Landskron et al. 2009). In addition to nuclear localization signals (NLS), PER and TIMd also contain nuclear export signals (NES) (Ashmore et al. 2003; Hara et al. 2011; Jang et al. 2015; Singh et al. 2019; Cai et al. 2021). For details on TIMd and TIMm biology, see a thorough review by Cai and Chiu (2021).

A great deal of data was revealed thanks to a beautiful collection of genetic mutants created in Drosophila, which, after all, provide the most accurate understanding of the impact on the ultimate output, the fly’s behavior. A fundamental feature of the circadian clock is temperature compensation, the phenomenon when the free-running period remains constant over a physiologically acceptable range of temperatures (the meaningful temperature range for recording fly locomotor activity is approximately 15–29°C). Some of the per and tim-d mutations result in a good temperature-compensated clock of either a long or short free-running period (Rothenfluh et al. 2000a, 2000b; Hamblen et al. 1998). However, per^L and multiple tim-d mutants produce a slower clock at high temperatures (Konopka et al. 1989; Matsumoto et al. 1999; Singh et al. 2019), which can be understood as a temperature-overcompensated clock, although one might argue that the mutated proteins are simply unstable at high temperatures. Interestingly, the mutated protein TIM^SL shortens the free-running period when expressed in the per^L mutant and completely restores the temperature compensation deficiency (Rutila et al. 1996). The opposite trend, temperature undercompensated clock, is produced by per^SLIH, per^S, and per^T mutations (Konopka et al. 1989; Konopka et al. 1994; Hamblen et al. 1998). The last three per mutants were particularly instrumental in elucidating how phosphorylation influences PER stability (Chiu et al. 2008; Chiu et al. 2011).

2.2 Posttranslational Modifications

The key part of the negative PER-TIMd TTFL is the delay between mRNA and protein peaks. The regulation of PER and TIMd stability is tightly connected with phosphorylation and dephosphorylation performed by several kinases and phosphatases. The phosphorylation of PER can also serve as a fascinating tale illustrating the synergistic research on the fruit fly and mammalian clocks. In 1998, two seminal papers described DOUBLETIME, a casein kinase 1 (CK1) isoform¹ type, as a key clock component in Drosophila (Price et al. 1998; Kloss et al. 1998). In mammals, a homologous casein kinase 1ε (CK1ε) was identified as a component of the hamster clock when the tau mutation was mapped by positional cloning as an arginine-to-cysteine amino acid substitution at a highly conserved position within the kinase domain (Lowrey et al. 2000). Only a year later, Toh et al. (2001) described in humans the molecular basis for familial advanced sleep phase syndrome (FASPS) as a serine-to-glycine mutation within the CK1ε binding region of human PER2², which results in hypophosphorylated PER2.

The physical interaction between DBT kinase and PER is remarkably stable (Kloss et al. 2001). Interestingly and surprisingly, in vitro studies on DBT^L and DBT^S using nonphysiological substrates suggested that both mutant proteins have reduced kinase activity (Kivimaë et al. 2008; Venkatesan et al. 2019), even though the resulting behavioral phenotype of these mutants is the opposite (Price et al. 1998). The molecular mechanism behind this conundrum was first revealed for mammalian PER2 in the context of multi-kinase hierarchical activities. The current phosphoswitch model involves two competing phosphorylation sites on mouse PER2, the FASP site and the phosphodegron sites, which regulate PER2 stability in opposing ways (Zhou et al. 2015; Masuda et al. 2020). Importantly, different splicing isoforms of mammalian CK1δ, a paralog of mammalian CK1ε, have different priming phosphorylation capacities – that is, the ability to phosphorylate residues in a protein region where no other phosphorylation mark is attached. CK1δ1 and CK1ε are more active in priming phosphorylation of the FASP site, whereas CK1δ2 is more potent in priming the degron site (see the excellent review by Narasimamurthy and Virshup 2021).

In D. melanogaster, the intronless dbt gene encodes only one protein version; thus, the priming phosphorylation of PER must be regulated by a different mechanism. Similar to mammalian PER2, Drosophila PER contains multiple phospho-clusters (Chiu et al. 2008; Garbe et al. 2013; Kivimaë et al. 2008; Top et al. 2018). The phosphorylation of the N-terminal region (serine 47 and nearby amino acids) by DBT generates a high-affinity binding site for supernumerary limbs (SLIMB) (Chiu et al. 2008), an F-box protein targeting PER for ubiquitination and subsequent proteasomal degradation (Grima et al. 2002; Ko et al. 2002; see also Sect. 4.2.4). On the other hand, phosphorylation by NEMO/NLK kinase at the per-short domain, a region located in the center of PER protein (positions 585–600 in 1224-amino acid PER), stimulates additional phosphorylation of several nearby sites by DBT. This multisite phosphorylation of the central part of PER prevents N-terminal phosphorylation, which is key for SLIMB binding, resulting in a time-delay phosphorylation circuit (Chiu et al. 2011). The complex phosphorylation pattern of PER seems to be one of the mechanisms behind the temperature compensation of the circadian clock in Drosophila (Joshi et al. 2022).

Several additional kinases were discovered as important components of the Drosophila clock affecting the negative TTFL with PER and TIMd. Some of them, such as NEMO (NMO), CK1α, and p38 phosphorylate PER (Chiu et al. 2011; Lam et al. 2018; Dusik et al. 2014), whereas CASEIN KINASE 2 (CK2) phosphorylates both PER and TIMd in Drosophila (Cai et al. 2021). SHAGGY (SGG), a kinase identified as a TIMd-phosphorylating enzyme, might also phosphorylate PER (Martinek et al. 2001; Top et al. 2016).

In addition to kinases, enzymes with the opposite role, phosphatases, participate in the regulatory loops of the circadian clock and counterbalance the impact of kinases on various TTFL components. PROTEIN PHOSPHATASE 2 (PP2), a multiunit enzyme, specifically dephosphorylates PER and regulates its abundance, PROTEIN PHOSPHATASE 1 (PP1) targets both PER and TIMd, and phosphatase of regenerating liver-1 (PRL1) selectively dephosphorylates TIMd in darkness (Sathyanarayanan et al. 2004; Fang et al. 2007; Kula-Eversole et al. 2021; Agrawal and Hardin 2016).

However, since many of the abovementioned kinases and phosphatases tend to recognize and (de)phosphorylate multiple substrates, their role in the clock has gradually become increasingly complex. For example, the transcription factor CLK is cyclically phosphorylated, with its minimum phosphorylation detected in the early night and its maximal phosphorylation in the morning (Lee et al. 1998). Its partner CYC is constitutively expressed (at least in Drosophila and related flies; see CLK-CYC idiosyncrasies in the lineage-specific description later in this chapter); thus, the cyclical phosphorylation of CLK seems to be the key rhythmic component of the circadian activator (Yu et al. 2006). The phase-specific hyperphosphorylation is DBT-dependent and leads to maximal repression of CLK activity. Remarkably, PP2, introduced in a previous paragraph as PER-regulating phosphatase, stabilizes CLK by counteracting the activity of DBT (Kim and Edery 2006). Furthermore, DBT-mediated degradation of CLK is further counterbalanced by CK2, which inhibits CLK degradation and reduces its activity (Szabo et al. 2013).

Not only is the phosphorylation state important to define the stability of the proteins and determine their interaction with partner proteins, but it might also specifically impact their subcellular localization, for example, the nuclear export of TIMd (Cai et al. 2021; Fang et al. 2007). An additional posttranslational modification, such as O-GlcNAcylation of PER, reduces PER interaction with CLK, a mechanism that might link the clock with the metabolic signals stemming from feeding activity (Li et al. 2019; Liu et al. 2021).

2.3 Entrainment by Light

The above-described machinery illustrates the molecular mechanism of the Drosophila clock “ticking” in constant darkness. However, the clock must be synchronized (entrained) with the external time of the surrounding environment. The most powerful cues are light and temperature. The major light-mediated synchronization involves the PER-TIMd feedback loop, although opsin-based receptors contribute as well (see Chap. 3). As was noted early with tim-d discovery and characterization, TIMd stabilizes PER, as there is no PER detected in the tim-d⁰¹ mutant. At the same time, TIMd is degraded upon light illumination, which led to a model where entrainment is achieved via light-mediated degradation of TIMd, which is further relayed to PER depletion, resulting in either phase advance or phase delay of the clock (Hunter-Ensor et al. 1996; Myers et al. 1996; Zeng et al. 1996).

However, TIMd is not sensitive to light by itself; instead, the signal must be transduced from a photoreceptor, which then leads to tyrosine phosphorylation of TIMd and its subsequent degradation through a ubiquitin-proteasome mechanism (Naidoo et al. 1999). The actual photoreceptor of the light-input pathway turned out to be flavoprotein CRYPTOCHROME (Stanewsky et al. 1998; Emery et al. 1998), a member of a large protein family present in all kingdoms and including different types of photolyases (Mei and Dvornyk 2015; Xu et al. 2021). Here, we will refer to this protein as Drosophila-type (CRYd), although CRYd is found in the majority of Protostomia and even some basal Deuterostomia (Kotwica-Rolinska et al. 2022a). CRYd functions as a photoreceptor within clock neurons located deep in the fly brain (Emery et al. 2000a), and its mutation or complete depletion results in behavioral rhythmicity in constant light (Emery et al. 2000b; Dolezelova et al. 2007), a condition under which wild-type flies become completely arrhythmic. Simplified systems of cell cultures and yeast helped to shed light on the mechanism: S2 cell transfection experiments revealed that CRYd blocks TIMd+PER-dependent inhibition of CLK-mediated transcription under light but has no impact in darkness, and the yeast two-hybrid system identified a light-dependent interaction between CRYd and TIMd (Ceriani et al. 1999). Furthermore, Rosato et al. (2001) discovered CRYd interaction with PER. The most variable region of various CRY proteins lies in their C-terminus (C-tail), a key part necessary in Drosophila CRYd for regulating its interaction with TIMd upon light illumination. If the C-tail is removed from CRYd, either by a stop codon mutation or engineered in a synthetic construct, the resulting “C-tailless CRYd” is constitutively active and interacts with TIMd even in constant darkness, resulting in a long free-running period (Busza et al. 2004; Dissel et al. 2004).

The interaction between TIMd and CRYd is further affected by the N-terminal region in TIMd. The s-tim-d allele encodes a 1398 amino acid-long S-TIMd protein, whereas the ls-tim-d allele contains two alternative start codons, resulting in transcripts encoding 1398 (S-TIMd) and 1421 (L-TIMd) amino acid-long proteins, respectively (Rosato et al. 1997). The ls-tim-d allele originated in D. melanogaster in southern Italy circa 10,000 years ago and has spread in all directions (Tauber et al. 2007). L-TIMd shows a diminished interaction with CRYd (Sandrelli et al. 2007), ls-tim-d flies are significantly more rhythmic under continuous light than s-tim-d flies (Deppisch et al. 2022), and only ls-tim-d flies can synchronize to seminatural conditions with short “white nights” typical of high latitudes (Lamaze et al. 2022).

Constant light rhythmicity was observed in flies with mutated JETLAG (JET), an F-box protein with leucine-rich repeat (LRR) (Koh et al. 2006; Peschel et al. 2006). The jet^c mutation results in a single amino acid change in the LRR region, the part of the protein important for substrate binding. Yeast two-hybrid experiments identified an interaction between JET and CRYd and further revealed that this interaction is reduced between JET^c and CRYd. Interestingly, JET and TIMd do not bind each other in the yeast system, but their interaction was detected in S2 cells (one of several examples illustrating the limitations of yeast-based experiments for reconstructing the Drosophila system). The current (working) model assumes the light-dependent interaction between CRYd and JET, as well as the interaction of TIMd with JET, and both CRYd and TIMd are then degraded by the proteasome (Peschel et al. 2009).

TIMm, a protein also known as TIMEOUT or TIM2, was first identified in mice (Zylka et al. 1998) and subsequently in fruit flies (Benna et al. 2000). This protein is essential for development in both mammals and flies, which heavily complicates its functional reverse genetics research. Flies with only one tim-m functional allele possess reduced sensitivity to light (Benna et al. 2010). RNA-mediated silencing (RNAi) in crickets and linden bugs affects behavioral rhythmicity (Nose et al. 2017; Kotwica-Rolinska et al. 2022a), which, together with the involvement of tim-m in the neuronal activity rhythm of the suprachiasmatic nucleus in rats (Barnes et al. 2003) and the connection of human tim-m with familial advanced sleep phase syndrome (Kurien et al. 2019), suggests TIMm as a conceivable and conserved clock component. For a more detailed description of TIMd/m proteins in Drosophila, see the review by Cai and Chiu (2021). For a general overview of the light input into various circadian clocks, see the review by Johnsson et al. (2014).

2.4 Protein Degradation

The oscillation of the clock system, either in constant darkness or in a light/dark regime, depends largely on the well-regulated protein turnover. The SKP/CULLIN/F-box-containing complexes (SCF complexes) function as E3 ubiquitin ligases targeting proteins for 26S proteasomal degradation. SLIMB is an F-box/WD40-repeat protein participating in the CULLIN-1-based E3 ubiquitin ligase complex that binds phosphorylated residues in the N-terminal region of PER (Chiu et al. 2008), ubiquitinates PER, and thus stimulates its degradation. Flies with mutated slimb are arrhythmic, and the same phenotype is obtained when a dominant-negative form of SLIMB is expressed in clock cells (Ko et al. 2002; Grima et al. 2002). Furthermore, highly phosphorylated PER and TIMd are constitutively present in the constant darkness of slimb mutants. However, the cyclical oscillation of PER and TIMd abundance is maintained even in slimb mutants under the light/dark regime due to the degradation pathway utilizing CRYd and JET proteins (see Sect. 4.2.3). Another ubiquitin ligase, the CULLIN-3 (CUL-3)-based complex, interacts with low-phosphorylated species of TIMd in the absence of PER. SLIMB, on the other hand, binds more phosphorylated TIMd, including PER-bound TIMd (Grima et al. 2012). CIRCADIAN TRIP (CTRIP) is an E3 ubiquitin ligase that seems to regulate the stability of both PER and CLK when ctrip downregulation results in a long free-running period, high CLK levels, and persistence of phosphorylated PER during the subjective day when PER is normally degraded (Lamaze et al. 2011).

A study aiming at the identification of cyclically ubiquitinated proteins, the cycling ubiquitylome, revealed a 2–2.5-fold oscillation in abundance for 52 proteins (15 % of all identified ubiquitinated proteins) (Szabo et al. 2018). A remarkable 29-fold oscillation was found for transcription of Megator (MTOR), a nuclear pore complex component. Additional cyclic ubiquitylation affects MTOR, which then feeds back to the pacemaker when it regulates the subcellular localization of the core clock proteins (Szabo et al. 2018).

2.5 bZIP Proteins PDP1, VRI, and KAY

The next feedback loop involves two basic leucine zipper (bZIP) transcription factors (Fig. 4.2a, b). First, vrille (vri) mRNA is cyclically transcribed by the CLK-CYC complex in a phase similar to tim-d expression (Blau and Young 1999; McDonald and Rosbash 2001). The second bZIP protein, PAR (proline and acidic rich) DOMAIN PROTEIN 1 (PDP1), particularly its isoform epsilon (PDP1ε), is cyclically translated from Pdp1ε mRNA, whose expression is also driven by the CLK-CYC complex from an upstream promoter in the Pdpd1 gene. However, Pdp1ε peaks several hours after vri reached its maximum, and a similar delay was reported for the peaks of Pdp1ε and VRI proteins (Cyran et al. 2003). Even though PDP1ε and VRI remarkably differ in the organization of protein domains, both proteins contain a highly conserved DNA-binding domain. Indeed, PDP1ε and VRI bind identical cis-regulatory elements in DNA, the so-called V/P motif (also known as the D-box), but with the opposite impact (Fig. 4.2b). The early peaking VRI serves as a transcriptional repressor on the Clk promoter. A few hours later, PDP1ε reaches its maximum, replaces VRI, and activates Clk transcription. The resulting two-loop model with the negative (VRI) and positive (PDP1ε) components then explained the cyclical expression pattern of Clk mRNA, which runs in antiphase to the expression of per, tim-d, and vri mRNAs (Cyran et al. 2003). A third bZIP protein, specifically its α isoform produced from the upstream promoter in the kayak gene, further contributes to VRI/PDP1ε feedback. Downregulation of KAYAKα (KAYα), a homolog of mammalian FOS, in circadian pacemaker neurons prolongs period length. KAYα binds to VRI and thus inhibits its suppression of the Clk promoter. Surprisingly, KAYα also represses CLK activity. These opposite roles of KAYα in the two-loop model were interpreted as a mechanism bringing stability and precision to the system (Ling et al. 2012).

Fig. 4.2

(a) Major transcription-translation feedback loops (TTFL) identified in insects (note that some components might be missing in certain insect lineages). The end style and color indicate whether the loop activates the transcription factors CLK-CYC/BMAL (positive feedback) or inhibits their activity (negative feedback). Loops and key modulatory processes are numbered, and the same numbering is used in Fig. 4.3. For simplicity, promoters are not shown in panel a, and transcription with subsequent mRNA processing is depicted as arrows running from the nucleus to the cytoplasm. The core of system ① consists of bHLH PAS proteins CLK and CYC/BMAL that drive the expression of the majority of components, including PERIOD (PER) ②, Drosophila-type TIMELESS (TIMd) ③, mammalian-type CRYPTOCHROME (CRYm) ④, bZIP transcription factors VRILLE (VRI) and PAR DOMAIN PROTEIN 1 (PDP1) ⑩, bHLH protein CLOCKWORK ORANGE (CWO) ⑪, and nuclear receptors E75, HR3, and UNF ⑫. Additional components include Drosophila-type CRY (CRYd) ⑤, which interacts with TIMd upon light illumination and, with the involvement of F-box protein JETLAG (JET), mediates the major light entrainment of the system (see Chap. 3 for details on photic entrainment). Mammalian-type TIM (TIMm) ⑥ contributes to rhythmicity in some insects and is involved in light entrainment in Drosophila; however, the mechanisms behind both roles are unclear. Key regulatory steps involve posttranslational modification, such as phosphorylation by multiple kinases ⑦, dephosphorylation by phosphatases ⑧, and degradation machinery ⑨. In Drosophila, Clk is cyclically expressed ⑬, and splicing of tim-d mRNA is regulated by PSI ⑮ in a temperature-dependent manner. KAYAK ⑭, a bZIP protein related to mammalian FOS, contributes to rhythmicity in Drosophila. (b) Scheme illustrating the role of cis-regulatory elements in Drosophila promoters. “AD” in CLK and PDP1 proteins refers to the activation domain, ccgs stands for clock-controled genes. See the text for a detailed explanation

2.6 Clockwork Orange (CWO)

Somewhat similar two-component feedback with opposite roles utilizes CWO, a protein belonging to the bHLH-ORANGE family related to mammalian DEC1 and DEC2, both of which are important for rhythmicity in mice (Honma et al. 2002). CWO was identified as a direct target of CLK (Kadener et al. 2007) and in parallel by genome-wide functional screening using an RNA interference (RNAi) system in flies (Matsumoto et al. 2007). CWO competes with the CLK-CYC dimer in binding to tandem E-boxes when the DNA-binding capacity is the highest for the CLK-CYC dimer, intermediate for CWO, and the weakest for the CLK-CYC-PER complex (Fig. 4.2b). Consequently, CWO binds E-boxes of core clock genes (per, tim-d, vri, Pdp1) in antiphase to CLK-CYC (Zhou et al. 2016). A recent study by Rivas et al. (2021) uncovered CLOCK INTERACTING PROTEIN CIRCADIAN (CIPC), an ortholog of mouse CIPC, as an additional component of the loop or more precisely a subloop within the CWO feedback (Fig. 4.2a). Drosophila CIPC decreases CLK-CYC-mediated transcription; however, the degree of repression is variable: the strongest for per, intermediate for tim-d, and minimal for vri. Flies with Cipc silenced by RNAi and Cipc null mutant flies produce a short free-running period. At the same time, the expression of Cipc is suppressed by CWO. Therefore, in addition to displacing CLK-CYC from tandem E-boxes and suppressing CLK-CYC transcription, CWO also indirectly activates CLK-CYC by removing CIPC repression (Rivas et al. 2021).

2.7 Nuclear Receptors E75, HR3, and UNF

The mammalian clock contains a feedback loop with RORα and REV-ERBα, two orphan nuclear receptors cyclically expressed by CLK-BMAL1 (brain and muscle aRNT-like 1, a mammalian homolog of Drosophila CYC). Both RORα and REV-ERBα recognize and compete for the same DNA motif RORE, which is localized in the promoter of mammalian Bmal1 and Clk genes. RORα serves as a transcription activator, whereas REV-ERBα suppresses transcription. The role of homologous nuclear receptors was addressed by RNAi in the basal insect Thermobia domestica when Kamae et al. (2014) identified E75 (homolog of REV-ERBα) and HR3 (homolog of RORα) as clock components. Both E75 and HR3 are cyclically expressed, and silencing either of them by RNAi influenced the phase of tim-d, cyc, and Clk expression in Thermobia.

Jaumouille et al. (2015) systematically analyzed all 18 nuclear receptor genes in Drosophila by RNAi in clock neurons. While E75 was identified as an important component of the fruit fly clock, silencing HR3 did not influence the rhythmicity. Interestingly, the silencing of another nuclear receptor, DHR51 (unfulfilled, UNF), was significant for the free-running period. S2 cell and Drosophila in vivo experiments then confirmed that E75 together with UNF coregulate CLK-CYC-mediated transcription of per when they bind to per regulatory sequences.

2.8 Regulation of the Drosophila Clock at the RNA Level

Circadian rhythmicity is further regulated at the level of posttranscriptional RNA processing, which includes alternative splicing, polyadenylation, mRNA stability, and regulation by microRNAs (for a detailed review, see Lim and Allada 2013a). The first of them, alternative splicing, was first documented for the per gene, where the last exon is retained at high temperatures, while splicing is enhanced at low temperatures. Interestingly, this exon is positioned in the 3’ untranslated region (3’ UTR) of mRNA, and its splicing leads to an advanced phase in per mRNA accumulation, which in turn results in advanced evening activity at low temperatures (Majercak et al. 1999).

Alternative splicing was also reported for tim-d, where two cold-specific isoforms are upregulated at low temperatures. At high temperatures, another isoform, tim-d-medium, is produced. This isoform is characterized by retention of the intron between exons 13 and 14, resulting in a premature stop codon and an unstable, and probably nonfunctional, TIMd-medium protein (Martin Anduaga et al. 2019). Thermosensitive splicing of this tim-d intron requires the alternative splicing regulator P-element somatic inhibitor (PSI) (Foley et al. 2019) and additional spliceosome factors (Shakhmantsir et al. 2019).

The length of the poly adenine (A) tail heavily influences mRNA stability. The deadenylase POP2 specifically shortens the tim-d mRNA poly(A) tail, thus destabilizing the transcript and leading to lower TIMd levels. Interestingly, POP2 activity is inhibited by PER, the partner of TIMd (Grima et al. 2019).

However, another level in RNA metabolism, its translation, is subject to circadian clock regulation. ATAXIN-2 (ATX2), an RNA-binding protein, is necessary for PER accumulation in clock neurons. ATX2 is crucial for the functions of twenty-four (TYF), a key activator of PER translation, which is associated with a 5’-cap-binding complex (Lim et al. 2011). TYF and ATX2 interact with polyadenylate-binding protein (PABP) (Lim and Allada 2013b; Zhang et al. 2013b).

2.9 Network Properties of the Clock

Although the clock mechanism was first accepted as a cell-autonomous oscillatory system, circadian rhythmicity also requires intercellular communication and is, therefore, a result of the circadian network. The first evidence was indicated by the discovery of a neuropeptide pigment-dispersing factor (pdf) mutant (Renn et al. 1999). Then, specific neurons regulating the morning and evening activity were identified and genetically manipulated (Grima et al. 2004; Stoleru et al. 2004).

The study by Dissel et al. (2014) further emphasized the role of particular groups of different clock cells and their mutual communication. The interaction among different neuronal oscillators (distinct groups of neurons), each characterized by a specific neuropeptide and coupled to other oscillator groups (neurons) in the network, drives the rhythmic activity of flies (Yao and Shafer 2014). The importance of clock groups was shown in calcium imaging experiments when the Ca²⁺ rhythms displayed by a particular group of cells corresponded to the morning or evening locomotor activity peaks (Liang et al. 2016), and the key role of neuropeptide signaling in these intracellular communications was further supported by Ca²⁺ imaging (Liang et al. 2017).

2.10 Temperature and the Clock

Temperature is an important and variable factor in the environment that interferes with the functioning of the circadian clock. The clock is only useful for the organism if it keeps ticking at a constant (or near-constant) speed within a physiologically acceptable range of temperatures. This temperature independence, known as temperature compensation, is particularly important for poikilotherms, which are organisms without regulation of body temperature, such as insects. On the other hand, chemical reactions are strongly influenced by temperature (Arrhenius 1889). An explanation of this logical “conflict” could be a combination of reactions, where some reactions have opposite effects on the free-running period than others (Hastings and Sweeney 1957). Thus, the period-extending reactions would, for instance, prolong the free-running period even more at high temperatures. At the same time, this lengthening would be compensated by other reactions that more strongly shorten the free-running period at high temperatures. These opposing reactions could impact the stability of some negative or positive components of the feedback loops and might influence their subcellular localization (nuclear export and import), their interaction with partners, their activity, or perhaps a combination of (all) the abovementioned mechanisms. We cannot exclude the possibility that temperature compensation may be partially affected at the level of interneuronal communication. In this case, temperature compensation would also be a network property of the clock.

Although the molecular mechanism behind temperature compensation remains largely elusive, genetic and biochemical data suggest that phosphorylation of distinct regions of PER protein could be consistent with the model proposed above, even though the details of the mechanism will most likely differ between mammals and Drosophila (see Sect. 4.2.2.). Components of the negative feedback loop are important for temperature compensation, as illustrated by mutations specifically interfering with PER-PER homodimerization in Drosophila (Landskron et al. 2009) and a subset of TIMd mutations (Matsumoto et al. 1999; Singh et al. 2019; Rutila et al. 1996; see Sect. 4.2.1.).

In addition to single amino acid mutant variants isolated in genetic screens, even certain Drosophila per alleles occurring abundantly in nature show altered temperature compensation. Drosophila PER contains a series of threonine-glycine (Thr-Gly) repeats that are polymorphic in length. The most common variants consist of 14, 17, 20, or 23 Thr-Gly pairs, and their frequencies in naturally occurring European populations display a clear latitudinal cline (Costa et al. 1992). Furthermore, a detailed analysis involving transgenic flies in which the Thr-Gly repeats and flanking regions were modified revealed the role of this PER in temperature compensation (Sawyer et al. 1997).

Although the free-running period remains constant (compensated) in a certain temperature range, the daily distribution of activity is often affected by temperature. Furthermore, circadian clock genes impact daily activity patterns under light-dark regimes. In D. melanogaster, morning and evening activity peaks are separated by midday siesta. At high temperatures, the morning peak is earlier, and the evening peak is delayed, whereas, at low temperatures, the trend is reversed. The temporal regulation of the morning and evening peaks includes alternative splicing of per mRNA in its 3’ UTR and complex alternative splicing of tim-d transcripts (see Sect. 4.2.8; Majercak et al. 1999; Martin Anduaga et al. 2019; Foley et al. 2019).

3 Lineage-Specific Variations in the Clock Setup

The remarkable progress in genomics and transcriptomics during the last decade resolved the phylogeny of insects with reasonable precision and dated the separation of all major insect lineages (Misof et al. 2014; Johnson et al. 2018; Kawahara et al. 2019; McKenna et al. 2019; Wipfler et al. 2019). At the same time, these phylogenomics-oriented studies produced a remarkable wealth of data in which clock genes can be identified. In the case of nonmodel insects, the transcriptomes are particularly valuable, as the complete coding sequences are often retrieved and gene paralogs can be reliably defined. Thus, the presence of different circadian clock genes in the insect phylogeny could be usefully mapped and important gene changes and losses identified. Pinpointing gene loss is a nontrivial endeavor. If the gene in question is well conserved across insects (let us say at the level of a protein it encodes) yet is absent in some species, we may claim that we are not able to identify it in a genome of certain insects³. However, genome assemblies of nonmodel insects are often fragmented; moreover, some genomes are remarkably large. For example, the genome of the migratory locust is more than twice the size of a human genome (Wang et al. 2014). Therefore, the absence of a gene in an individual species must be interpreted with extreme caution. Nevertheless, if the genomes and transcriptomes of all species within a specific monophyletic group (i.e., a group of organisms with a common ancestor) do not contain a particular gene, and if the quality of these genomes/transcriptomes is reasonably good, then the gene loss becomes the most parsimonious explanation.

This part of the chapter aspires to compare the circadian clock across insects using D. melanogaster as the reference model. Here, it is important to keep in mind that the depth of our knowledge remarkably varies among different insect groups. For some, only genomic/transcriptomic evidence is available, whereas, in others, the circadian clock was functionally studied by gene silencing using RNAi and, in some species, stable genetic mutations were introduced. Another set of insights includes temporal expression analyses and immunohistochemistry, valuable data that might support mechanistic explanation, even though functional reverse genetic evidence is not available.

We will start with the closest relatives of D. melanogaster and gradually focus on the major changes in the circadian clock setup. The key TTFL loops and modification enzymes, such as kinases and phosphatases, depicted in a generalized insect clock model (Fig. 4.2a) are first approached from a phylogenetic perspective in Fig. 4.3, where panel “a” defines major gene losses and changes during insect evolution, whereas panel “b” summarizes the experimental evidence for each loop (using the numbers in circles) in a particular lineage. The description of the circadian clock at the RNA and protein levels is summarized separately in Table 4.1.

Fig. 4.3

Lineage-specific changes in the insect clock setup depicted from a genetic perspective. Colors correspond to Figs. 4.1 and 4.2. (a) A simplified phylogenetic tree with highlighted gene losses and the transition of BMAL to CYC. The genetic setup seems to be conserved in entire insect orders (Lepidoptera, Hymenoptera, Orthoptera) or lower taxonomic groups (aphids, mosquitoes), but in several cases, representative species are depicted. (b) Experimental evidence for the role of specific loops and components in particular insect lineages (numbers correspond to the scheme in Fig. 4.2a: CLK and CYC/BMAL ①; PER ②; TIMd ③; CRYm ④; CRYd ⑤; TIMm⑥; kinases ⑦; phosphatases ⑧; degradation machinery ⑨; bZIP transcription factors VRI and PDP1 ⑩; CWO ⑪; nuclear receptors E75, HR3, and UNF ⑫; cyclical expression of CLK ⑬; KAYAK ⑭; and splicing of tim-d mRNA regulated by PSI ⑮. If experiments were performed only in an artificial system of cell culture, the number is depicted without the circle. CRYd has been lost in Periplaneta, whereas its presence/absence is not fully clear in Rhyparobia; thus, X is depicted in parentheses. Deviation from the Drosophila model is highlighted by green color. As a reference, a mouse clock setup is depicted with homologous components shown in brown. (c) Depiction of major combinations of TTFL components in insects

Table 4.1 Overview of descriptive analyses characterizing circadian clock factors

The D. melanogaster clock setup seems to be, at least according to available descriptive data, conserved in cyclorrhaphan flies, including the housefly Musca domestica (Codd et al. 2007; Bazalova and Dolezel 2017), the olive fly (Bertolini et al. 2018), and the Medfly Ceratitis capitata (Kotwica-Rolinska et al. 2022a). The tim-d mutant of Chymomyza costata, a drosophilid fly living in temperate regions, not only confirmed the role of tim-d in its molecular oscillator (Kobelkova et al. 2010) but also supported the involvement of the circadian machinery in seasonality (Pavelka et al. 2003; Stehlik et al. 2008).

3.1 Unique Features of the Drosophila (Cyclorrhaphan) Clock

Two or perhaps three major and probably connected features are unique to Drosophila and related fly species. First, Drosophila CLK contains a transactivation domain at its C-terminus, and a similar pattern is found in cyclorrhaphan Diptera (including Musca, the olive fly, and the medfly). Perhaps connected to that, Clk mRNA is cyclically expressed in these species, whereas the cyc level is constant (Cyran et al. 2003; Codd et al. 2007; Bertolini et al. 2018). BMAL1, the mammalian homolog of Drosophila CYC, is characterized by cyclical bmal1 expression in mice, and the transactivation domain, the so-called the BMAL1 C-terminal region (BCTR), lies in the C-terminus⁴. The same protein architecture was identified in the BMAL of the silk moth (Chang et al. 2003) and even sand flies (Meireles-Filho et al. 2006), indicating the unique and relatively recent rearrangement of the positive components in the Drosophila clock (Fig. 4.3a, c). The role of CLK-BMAL as positive regulators of the circadian machinery was experimentally confirmed by RNAi in Heteroptera, Orthoptera, and Thermobia (Kotwica-Rolinska et al. 2022a; Uryu et al. 2013; Moriyama et al. 2012; Kamae et al. 2014) (Fig. 4.3b). In Lepidoptera, stable genetic mutants were created for both CLK and BMAL (Markert et al. 2016; Zhang et al. 2017) in addition to the reconstruction of the feedback in cell culture (Chang et al. 2003). Furthermore, the bmal transcript is cyclically expressed in numerous (and phylogenetically distant) insects, including sand flies (Meireles-Filho et al. 2006), crickets (Uryu et al. 2013), and Thermobia (Kamae et al. 2014).

The second unique feature of the Drosophila setup is the absence of mammalian-type CRYPTOCHROME (CRYm), also known as CRY2, CRYII, or mCRY (Kotwica-Rolinska et al. 2022a). In mammals, two paralogous mammalian-type CRYs are key components of the negative TTFL (Kume et al. 1999; van der Horst et al. 1999). A similar transcriptional repressive function was confirmed in cell cultures for CRYm from mosquitoes, Lepidoptera, Coleoptera, and Hymenoptera (Yuan et al. 2007). Functionally, CRYm was identified as a clock component by RNAi in mosquitoes, Heteroptera, crickets, and cockroaches (Meuti et al. 2015; Ikeno et al. 2011a; Tokuoka et al. 2017; Bazalova et al. 2016; Werckenthin et al. 2020). Stable genetic mutants were created in the linden bug Pyrrhocoris apterus and the monarch butterfly Danaus plexippus (Kotwica-Rolinska et al. 2022a; Zhang et al. 2017). The latter model was instrumental in explaining that monarch CRYm regulates circadian repression of BMAL through the activation domain (AD) via two independent mechanisms (Zhang et al. 2017). The perfect correlation of CRYm loss with the absence of AD in BMAL/CYC of certain dipteran insects implies that once CRYm was lost, probably in an ancestor of Cyclorrhapha, repression of AD on BMAL was no longer used. The activation domain is localized on CLK in insects that have lost CRYm (Thakkar et al. 2022).

The third specificity of Drosophila is the involvement of the nuclear receptor UNF (Jaumouille et al. 2015), whereas basal insect Thermobia utilizes HR3, a homolog of the well-established mammalian component RORα (Kamae et al. 2014) (see Sect. 4.2.7 also). It is tempting to speculate that UNF recruitment to the circadian machinery might be connected to the shift of the activation domain from BMAL to CLK because the oscillating expression of Clk and noncyclical expression of cyc, characteristic only for cyclorrhaphan flies, imply evolutionary changes in the promoters of both genes. It will be interesting to see if HR3 and/or UNF participate in the circadian clock of other insect groups.

3.2 Losses of CRYd and TIMd

The mammalian clock setup is characterized by the absence of CRYd and TIMd. Therefore, it was a remarkable surprise to see a comparable situation in Hymenoptera when the genome of the honey bee Apis mellifera was analyzed (Rubin et al. 2006). A recent detailed inspection confirmed that TIMd and CRYd were lost in all Hymenoptera. Furthermore, TIMd has been lost in the animal kingdom at least three times, in Hymenoptera, vertebrates, and the majority of termites, and these three losses always correlate with the loss of CRYd (Kotwica-Rolinska et al. 2022a; Fig. 4.3a).

However, in several lineages, CRYd has disappeared and TIMd is still present. We can see a particularly nice gradient of gene losses in cockroaches and termites (which are from a phylogenetic point of view a subset of cockroaches): the basal cockroaches (such as Blattella germanica) possess the complete toolkit, Periplaneta lost CRYd, and termites also lost TIMd (except for just one basal termite species Porotermes, where a portion of tim-d transcript was found). It will be extremely interesting to see how the light entrainment pathway functions in these three types of clocks detected in cockroaches/termites.

In Orthoptera, a sister group to cockroaches, the role of both cryptochrome types was tested by RNAi applied to the cricket Gryllus bimaculatus, leading to a new model with a CRYm-CRYd oscillatory loop independent of PER-TIMd (Tokuoka et al. 2017). Furthermore, light input seems to require c-FOS, a bZIP protein participating in the mammalian entrainment pathway (Kornhauser et al. 1990). In the proposed cricket model, the light signal is perceived by the compound eyes, from which the information is transmitted to the clock neurons to stimulate c-fos mRNA expression, which is finally relayed on the CRYm-CRYd loop that feeds back on CLK-CYC/BMAL (Kutaragi et al. 2018).

Interestingly, RNAi silencing of Gryllus tim-d did not result in behavioral arrhythmicity; instead, the free-running period was significantly shorter (Danbara et al. 2010). A similar short free-running period was observed after tim-d silencing in the firebrat T. domestica (Kamae and Tomioka 2012). These findings contrast with the fundamental role of tim-d in the circadian clock of Drosophila but also in Lepidoptera (Nartey et al. 2021). The role of tim-d was rigorously addressed in the linden bug P. apterus (Heteroptera), an insect that does not have cry-d and instead possesses cry-m (Bajgar et al. 2013). Genetically created null mutant in tim-d showed a free-running period of their locomotor activity shortened by more than 1 h, but the rhythmicity was robust. These tim-d phenotypes provided a possible explanation for evolutionary changes in the clock setup within animals in general and insects in particular (Kotwica-Rolinska et al. 2022a): “The dispensability of TIMd in P. apterus suggests a scenario of transition between clock architectures relying on distinct components of their negative feedback loops. The proposed model implies that the clock would be functional in each step of the transition from the ancestral state to the PER & CRYm system known today in vertebrates. A similar clock gene combination in Hymenoptera (PER & CRYm) indicates a convergent evolution of the circadian system, although functional evidence from this insect group is not yet available. The circadian clock observed in P. apterus could then correspond to an early clock setup that facilitates tim-d loss without a complete collapse of circadian cycling. However, the timing and causality of the proposed events might have been lineage-specific, where either the loss of cry-d triggered the transition of TIMd to its modulatory role, or alternatively, the loss of JET or change in TIMd properties compromised its interaction with CRYd, which in turn was subsequently lost.”

3.3 JET and FBXL3/21 Proteins

JET, a protein essential for the interaction between CRYd and TIMd in Drosophila, which results in their degradation, is found in numerous protostomian lineages but is also independently lost at least six times in insects (Kotwica-Rolinska et al. 2022a). Thus, various combinations of JET with TIMd and CRYd exist in insects (Fig. 4.3c): JET, TIMd, and CRYd (Diptera, Lepidoptera, some Coleoptera, Blatella); JET, TIMd, and no CRYd (Periplaneta); and JET, no TIMd, and no CRYd (Cryptotermes). Similarly, JET was lost, but TIMd and CRYd are present in aphids (Cortés et al. 2010). In Pyrrhocoris, JET was lost and only TIMd is present, whereas all these three genes were lost in Hymenoptera.

In mammals, CRYm interacts with FBXL3 and FBXL21, two closely related paralogs also known as overtime and after-hours, respectively (Godinho et al. 2007; Siepka et al. 2007; Hirano et al. 2013). In protostomes and basal deuterostomes, only the ancestral protein FBXL3 is found. Similar to JET, FBXL3 has been lost multiple times in insects and is present in four lineages: Ephemeroptera, Blattodea (cockroaches + termites), Thysanoptera, and Hymenoptera (Kotwica-Rolinska et al. 2022a). Thus, FBXL3 is not necessary for CRYm function in some systems, such as the circadian clock of Lepidoptera or Pyrrhocoris. Nonetheless, it will be very interesting to see whether FBXL3 participates in CRYm regulation in Hymenoptera, a group of insects with a clock setup remarkably similar to that of mammals.

3.4 Conserved Components of Insect Clocks

PER, an iconic circadian clock protein found in all insects, seems to participate in the rhythmicity in all tested species, albeit to a different extent. It is absolutely necessary for Drosophila, but in Pyrrhocoris, approximately one-third of genetic mutants are still robustly rhythmic (Kotwica-Rolinska et al. 2022a). Similarly, RNAi silencing of per in the Madeira cockroach resulted in only a partial phenotype (Werckenthin et al. 2020). It is possible that in the abovementioned (and many other) insect species, CRYm serves as the most important negative element, whereas the PER contribution to the negative feedback is smaller than that in Drosophila. However, depletion of PER in the silk moth Bombyx mori leads to arrhythmicity (Ikeda et al. 2019), even though the lepidopteran clock relies on CRYm (Zhang et al. 2017).

The role of bZIP proteins VRI and PDP1 was tested in Pyrrhocoris, where vri silencing nonsignificantly shortened the free-running period, while Pdp1 knockdown resulted in arrhythmicity (Kotwica-Rolinska et al. 2022a). In crickets, however, cosilencing of cyc was necessary to obtain vri- and Pdp1-dependent changes in the free-running period (Narasaki-Funo et al. 2020). In the northern house mosquito Culex pipiens, both vri and Pdp1 were tested for their role in diapause, and their cyclical expression was confirmed (Chang and Meuti 2020).

CWO, a bHLH protein whose depletion slows down the clock in Drosophila, was silenced with a comparable 1-h extension of periodicity in Pyrrhocoris (Kotwica-Rolinska et al. 2022a) and crickets (Tomiyama et al. 2020).

The role of components working at the protein level, i.e., kinases and F-box proteins, was minimally tested outside Drosophila. Silencing dbt robustly extended the free-running period in Pyrrhocoris, and a less extreme yet remarkable phenotype was observed after slibm knockdown, whereas nmo knockdown marginally sped up the clock (Kotwica-Rolinska et al. 2022a).

3.5 Descriptive Studies

The majority of the mechanisms described in this chapter were elaborated in D. melanogaster and further challenged and expanded in several insect models. Although an occasional spontaneous mutation or a variant is mapped in a nonmodel organism, these examples are relatively rare for chronobiology outside of D. melanogaster (Pavelka et al. 2003; Kozak et al. 2019). Therefore, reverse genetic tools have been key for the analysis of nonmodel organisms. One of the most powerful tools is RNAi, which, if spread in the organism systemically, can be a cost-effective and fast approach for chronobiological studies. However, the target transcript might only be partially reduced; thus, the data interpretation becomes, at least in some specific cases, nontrivial. Therefore, stable genetic modifications are more attractive from an interpretation point of view, but these techniques are also quite laborious and time demanding. Nevertheless, some insect models have become accessible to stable modification and are successfully used in chronobiology (Markert et al. 2016; Zhang et al. 2017; Kotwica-Rolinska et al. 2019; Ikeda et al. 2019).

In some groups, however, no reverse genetics data are available. However, many of these insect species are extremely interesting for their biology and even for their role as agricultural pests or for their ability to transmit pathogens. Thus, even descriptive evidence might provide important and valuable hints about the molecular mechanism behind their circadian clocks. Therefore, Table 4.1 briefly summarizes the evidence obtained at the RNA and protein levels across selected insect groups. For the anatomy of the insect circadian clock, see Chap. 5.

4 Conclusion and Future Perspective

The genetic architecture of the above-described circadian clocks observed in various insect lineages invites us to speculate why and how this diversity has originated. Insects are a large group with more than 400 million years of evolution (Misof et al. 2014). Some lineages, such as aphids, underwent noteworthy gene expansion, resulting in >35 thousand genes, which is remarkably more than the gene count in D. melanogaster (~13.5 thousand) or even in the body louse Pediculus humanus (~11 thousand) (Thomas et al. 2020).

A possible mechanistic explanation for the various clock setups might be provided by mammalian research when Pett et al. (2018) showed that the importance of individual feedback loops differs in a tissue-specific manner. In mammals, this mechanistic flexibility may account for organ-specific differences in clock gene expression and allow for the hierarchical organization of the mammalian clock. A somewhat similar tissue-specific clock architecture exists in D. melanogaster, where CRYd serves as a light photoreceptor in clock neurons (Emery et al. 2000a), whereas CRYd is a CLK/CYC repressor in the periphery (Collins et al. 2006). Indeed, the role in the peripheral clock allowed the identification of cry-d^b mutation in a luciferase reporter-based screen (Stanewsky et al. 1998).

In insects, we can assume that similar flexibility in clock setup allows for gene loss, which has some impact on rhythmicity, albeit not detrimental. Indeed, this is the case for the tim-d gene in P. apterus and probably other insect species, where TIMd participates only as a modulator of the free-running period (Kotwica-Rolinska et al. 2022a; Danbara et al. 2010). Thus, modifications of a certain loop may, for example, impact the light entrainment capacities of the system, whereas the clock remains either fully functional or with only a mild impact on its properties. Whether the combination of these changes is beneficial or disadvantageous for the particular organism depends on the life strategy in a specific environment. For example, organisms living in high geographical latitudes face multiple challenges, including extreme photoperiods (day-to-night ratio). Under such conditions, the less light-sensitive clock might be an advantage.

Another conceivable selection pressure might be the role of circadian clock genes in different time-measuring systems than the circadian clock itself. The possible role of (at least a subset of) clock genes was suggested in photoperiodic timing (Ikeno et al. 2010, 2011b, 2013; Kotwica-Rolinska et al. 2017, 2022b; Sakamoto et al. 2009) but not for the tidal and lunar rhythms (Takekata et al. 2014; Zhang et al. 2013a). Therefore, components recruited by a different time-measuring system(s) might be, in addition to their role in the circadian clock, under selection pressure connected to the properties of the noncircadian system.

Taken together, during the last two decades, we could see remarkable progress in understanding the circadian clock mechanism at the molecular level. Research on D. melanogaster has been dominating insect molecular chronobiology for decades, and the fruit fly will always be an excellent model with unparalleled tools and opportunities. However, emerging reverse genetic tools available to nonmodel organisms will facilitate research on genetically interesting clock setups and chronobiological, ecological, and physiological phenomena unavailable, or weakly expressed, in D. melanogaster.