Molecular and Neural Control of Insect Circadian Rhythms Part 5

The Contribution of Glia to Circadian Behavior

There is solid evidence for a role of glial cells in the control of circadian behavior. PER is rhythmically expressed in brain glial cells (Siwicki et al., 1988). Moreover, in gynan-dromorph studies (male/female mosaics), it was observed that PER expression restricted to glial cells could sustain weak behavioral rhythms (Ewer et al., 1992). In a more recent study, the glia-specific gene ebony (e) was found to be expressed in a rhythmic fashion in glial cells (Suh and Jackson, 2007). Interestingly, ebony mutants show arrhythmic or weakly rhythmic locomotor rhythms, but normal eclosion rhythms (Newby and Jackson, 1991). Restricted expression of EBONY in glial cells is sufficient to rescue robust rhythmic behavior in ebony mutants (Suh and Jackson, 2007). It should be noted, though, that repeated back-crossing of the e1 mutation into a wild type background also restored normal behavioral rhythms (Hall et al., 2007). This indicates that multiple genetic factors contribute to the ebony behavioral phenotype; not just the mutation in the ebony gene. Molecular rhythms in adult circadian neurons persist in e1 mutant flies; it is therefore the circadian behavioral output that is affected. How EBONY regulates circa-dian behavior is not yet known, but since it is an enzyme that can conjuguate P-Alanine to various biogenic amines, it is probably significant for the metabolism of a neurotrans-mitter important for circadian locomotor output. Studying the circadian role of glia will certainly reveal novel mechanisms that shape circadian behavior in the future.


Bridging Neural Circuitry and Genomic Studies

There has been remarkable progress in methods to profile whole-genome expression in specific cell types. This technological development has allowed the study of genomic expression in subsets of Drosophila circadian neurons, such as the small and large LNvs, as well as PDF-negative circadian neurons (Nagoshi et al., 2009; Kula-Eversole et al., 2010). In brief, specific subsets of circadian neurons can be GFP labeled with the tools we have described in section 15.4.4, and then manually sorted after dissociation. Comparing gene expression between non-circadian neurons and circadian neurons, or between different types of circadian neurons, should ultimately reveal which genes are critical for the function of the different circadian neurons. Already, the transcription factor FER2, whose expression is highly enriched in PDF-positive neurons, has been found to be essential for their proper development (Nagoshi et al., 2009). As expected, Fer2 mutant flies are arrhythmic. A specific isoform of NOCTURNIN is specifically expressed in dorsal neurons (Nagoshi et al., 2009). Flies with reduced expression of this specific iso-form are rhythmic in LL and have an attenuated PRC, indicating that it is regulating CRY circadian photores-ponses. The mechanism is not yet known, but vertebrate NOCTURNIN homologs regulate mRNA stability through polyA removal (Baggs and Green, 2003). It will be very interesting to identify the RNAs targeted by NOCTURNIN in the DNs. These two examples validate cell-specific gene profiling as a method to understand cir-cadian neural circuits both molecularly and functionally.

Interestingly, among the genes enriched in PDF positive circadian neurons, only half are shared by the large and small subtypes (Kula-Eversole et al., 2010). Studying those differentially expressed genes should reveal how the I-LNvs regulate sleep, light-mediated arousal, and phase advances (Parisky et al., 2008; Shang et al., 2008; Sheeba et al., 2008b), while s-LNvs control morning anticipation and DD behavior (Grima et al., 2004; Stoleru et al., 2005; Cusumano et al., 2009). With improved tools to target specific circadian neurons, and advances in next generation sequencing technology, much will soon be learned about the gene expression patterns that define the function of different circadian neurons.

Control of Circadian Rhythms in Non-Drosophilid Insects

Most of this topic has been dedicated to Drosophila melanogaster circadian rhythms, because it in this insect that we know by far the most about these adaptive mechanisms. However, the study of circadian clocks in non-drosophilid insects reveals an unexpected diversity in the molecular mechanisms and neural substrate controlling circadian rhythms. This section will illustrate this point through a few examples.

Neural Substrates Regulating Circadian Behavior in Non-Drosophilid Insects

To identify pacemaker neurons in the absence of genetics, a classical approach is a combination of surgical brain lesions and brain fragment transplants. Antibodies targeted against clock proteins can then be used to identify more precisely candidate pacemaker neurons. A very successful mapping of circadian pacemaker neurons using these approaches was achieved in the cockroach Leucophaea maderae. As in Drosophila, the pacemaker neurons are localized in the accessory medulla of the optic lobes, and express PDF. Moreover, PDF injection can phase-shift circadian locomotor behavior (Petri and Stengl, 1997). This suggests a significant conservation of the circadian neural mechanism between flies and cockroaches. However, in Lepidoptera for example, pacemaker neurons are not located in the accessory medulla, but most likely in the dorsal lateral protocerebrum (Truman, 1972, 1974; Sauman and Reppert, 1996; Zhu et al., 2008). In the Monarch butterfly, only four neurons in the pars later-alis show rhythmic nuclear entry of the critical repressor CRY2 (see next section), and are thus believed to be the brain pacemaker neurons (Zhu et al., 2008). There has thus been considerable evolution of the circadian neural network in insects, which is also revealed by clock gene expression studies (reviewed in Sandrelli et al., 2008; Tomioka and Matsumoto, 2010).

Surprisingly, at least in the case of the Monarch butterfly (Danausplexippus), peripheral oscillators have recently been found to regulate a behavior dependent on circadian clocks. Monarch butterflies use the sun as a compass during their annual migration to Mexico (Reppert et al., 2010). Since the sun moves during the day, the butterflies need to compensate their direction of migration as a function of time of day. This is dependent on the circadian clock (Mouritsen and Frost, 2002; Froy et al., 2003). Unexpectedly, the antenna and its local circadian pacemakers are necessary for time-compensated sun-compass navigation (Merlin et al., 2009). Whether brain pacemaker neurons are also important for this phenomenon and interact with antennal clock neurons is not yet known.

Surprising Diversity of Insect Circadian Molecular Pacemakers

As discussed above, Drosophila and mammalian pacemakers share many common principles, and homologous proteins are involved in both systems. It is worth briefly comparing the two systems before describing circadian clocks in non-drosophilid insects. First, the PER loop is very well conserved, although there are notable differences. There are two redundant functional mammalian homologs of Drosophila CLK: Clk and NPAS2 (Antoch et al, 1997; DeBruyne et al., 2007). Bmal1 is the CYC homolog (Bunger et al., 2000). Interestingly, in mammals Bmal1 bears a transcriptional activation domain, while CYC does not. Two of the three mammalian PER homo-logs are essential for circadian rhythms (Bae et al., 2001). They are degraded by the SLIMB homolog P-TRCP (Reischl et al., 2007). The mammalian TIM homolog is actually closer to Drosophila TIM2, a gene that might play a role in light responses, in addition to its role in maintenance of chromosomal integrity (Benna et al., 2010). Whether mammalian TIM might have a role in the cir-cadian pacemaker is controversial (Gotter et al., 2000; Tischkau and Gillette, 2005). It might actually connect the circadian clock with the cell cycle and DNA damage checkpoints through its interaction with CRY2 (Unsal-Kacmaz et al, 2005). In fact, CRY1 and CRY2 interact with mammalian PERs (Griffin et al., 1999; Kume et al., 1999). Unlike Drosophila TIM, which is probably not directly active in transcriptional repression of CLK/CYC, mammalian CRYs are the main repressors of Clk/Bmal1 transactivation (Kume et al., 1999). This is a major distinction between the two systems. Another important distinction is that CRYs are probably not circadian photoreceptors (Hattar et al., 2003; Panda et al., 2003). Pho-toreception is dependent on retinal opsins: rod and cone opsins, and melanopsin. Homologs of DBT and CKII are also involved in the mammalian pacemakers (Lowrey et al., 2000; Xu et al., 2005; Etchegaray et al., 2009; Maier et al., 2009).

The circadian pacemaker has been studied in several non-drosophilid species (for a detailed review, see Sandrelli et al., 2008). Interestingly, these comparative and mechanistic studies reveal different molecular organizations. For example, there is apparently no type 1 (pho-toreceptive) CRY and no Drosophila TIM homolog in the honeybee genome (Rubin et al., 2006; Yuan et al., 2007). The architecture of the PER loop is probably very similar to that of mammals, because a type II CRY is a potent repressor of CLK/CYC transcription (Yuan et al., 2007). Since the honeybee does not have type I CRY, it would be very interesting to determine whether isolated organs are nevertheless directly light sensitive, or if all circadian photoresponses are dependent on visual inputs.

Between this "mammalian" type of pacemaker and Drosophila, there are hybrid clocks. This is the case of the Monarch (Danaus plexippus) clock (Figure 8), which might be the best-understood circadian clock, at the molecular level, in a non-drosophilid insect. One of the main reasons for this deeper understanding of the Monarch clock is the existence of a Monarch cell line (Dpn1) that expresses most clock genes (Zhu et al., 2008). PER, TIM,and CRY2 cycle with phases and amplitudes reminiscent of those observed in vivo under LD conditions.

Comparison between the Drosophila and Monarch butterfly circadian pacemaker. Simplified models of the circadian pacemaker and CRY input pathway of Drosophila (top) and Monarch butterflies (bottom). The major difference is the presence of a type II CRY in Monarch, which functions as the main element of the negative limb of the PER feedback loop.

Figure 8 Comparison between the Drosophila and Monarch butterfly circadian pacemaker. Simplified models of the circadian pacemaker and CRY input pathway of Drosophila (top) and Monarch butterflies (bottom). The major difference is the presence of a type II CRY in Monarch, which functions as the main element of the negative limb of the PER feedback loop.

However, Dpn1 rhythms are not self-sustained, which means that an important element is missing in these cells for the clock to keep running under constant conditions. Nevertheless, results obtained in this cell line, in which RNAi can easily be performed, combined with in vivo data and Drosophila S2 cell transactivation assays, give us a detailed view of the functioning of the Monarch circadian clock (Zhu et al., 2005, 2008). The clock truly works like a mammalian-Drosophila hybrid. Light is perceived by CRY1 (Type I CRY), and triggers a cascade of degradation: first TIM, then PER, and ultimately CRY2. All three proteins (TIM, PER, and CRY2) are part of a complex, with CRY2 being, as in mammals, the primary repressor of CLK/CYC activity (Zhu et al., 2005, 2008). As in mammals, CYC has an activation domain, a characteristic found in most insects (Chang et al., 2003; Zhu et al., 2005; Rubin et al., 2006; Sandrelli et al., 2008). The Tribolium castaneum clock is probably located somewhere between the Monarch and the bee clock, since the Tribolium genome does not contain a type I CRY, but does bear both a type II CRY and a Drosophila-type TIM (Yuan et al, 2007).

Of course, functional analysis of non-drosophilid cir-cadian genes has mostly been performed in cell culture, using transcriptional assays to determine which molecules are activators or repressors in the circadian pacemaker. What is missing here is a genetic demonstration of the role of these molecules for circadian time-keeping. RNAi techniques such as the injection of dsRNAs have had limited success in insects, but appear to be working remarkably well in the cricket Grillus bimaculatus. In this insect, dsRNA-mediated knockdown of PER expression to trough level is sufficient to disrupt circadian behavior (Moriyama et al., 2008, 2009). TIM knockdown did not abolish rhythms, but period was short, suggesting that TIM is also involved in the cricket clock (Danbara et al., 2010). Whether it is essential for rhythmicity in this insect remains to be determined. These recent developments strongly suggest that it will be possible in the close future to test circadian gene function in vivo in many more species.

Conclusions

Our understanding of circadian clocks in insects, and in the animal kingdom in general, owes a lot to the studies in Drosophila melanogaster. There is no doubt that much still has to be learned in this organism. The details of the regulation of PER, TIM, and CLK phosphorylation are still far from being understood, and there are undoubtedly more genes regulating the circadian pacemaker that will have to be identified. For example, little is known about the role of Histone modifications in Drosophila. Also, the role of miRNAs in the control of circadian rhythms has just begun to be investigated. We have only a very vague understanding of how circadian neurons talk to each other in the circadian neural network, and the function of at least two-thirds of adult circa-dian neurons has yet to be found. Finding a function for these neurons will probably require the design of novel behavioral paradigms. Indeed, adult circadian behavior has mostly been synonymous with locomotor rhythms measured under specific experimental conditions. What flies are trying to do in the small test tubes in which they are confined with a little bit of sugar food is unclear: are they trying to escape to find mates, or a more nutritious source of food? The finding that the contribution of a subset of DN1s is dependent on ambient temperature and light intensity, while the contribution of other circadian neurons does not appear to be modulated by these inputs, further raises these questions (Zhang et al., 2010b). Which activity needs to be restricted to favorable environmental conditions, and which should be carried over anyway? Interestingly, the very same DN1 neurons that integrate light and temperature inputs play an important role in the circadian timing of male courtship (Fujii and Amrein, 2010).

There are, of course, very important questions on the role of insect circadian clocks that cannot be addressed with Drosophila – or can be addressed only with much greater difficulty. We are, for example, thinking of pho-toperiodism. Flies show a shallow photoperiodic ovarian diapause (Saunders et al., 1989; Saunders, 1990). Thus, assessing the role of the circadian clock in this process is difficult. Many insects demonstrate much more robust photoperiodic behavior. For example, the cricket Modi-cogrillus siamensis’ nymphal development is accelerated under a long photoperiod. Interestingly, injection of per dsRNAs entirely disrupts the circadian clock and affects nymphal development, which now occurs under the same kinetics as in animals kept in constant darkness (Sakamoto et al., 2009). PER therefore appears to be critical for the timing of photoperiodic nymphal development, although this does not yet demonstrate that the circadian clock itself is indeed the timer required for photoperiodism in M. siamensis. (Emerson et al., 2009; Sakamoto et al., 2009). PER function in photoperiodism could be independent of its circadian role. This example illustrates clearly that we are at a point where gene function analysis can be performed in non-drosophilid insects. Transgenesis has been developed for many insects, RNA interference can be obtained through different approaches (virus, dsRNA injection, transgenesis), and even site-directed mutagen-esis is now a possibility with the development of zinc-finger nucleases (Remy et al., 2010; Takasu et al., 2010). This means that behaviors such as time-compensated sun-compass navigation or circadian foraging behavior might soon be accessible to genetic dissection. We might thus be at the dawn of an explosion of knowledge in the field of insect circadian rhythms.

Next post:

Previous post: