Category Archives: Clock News

Evolutionary Medicine: Does reindeer have a circadian stop-watch instead of a clock?

ResearchBlogging.orgWhenever I read a paper from Karl-Arne Stokkan’s lab, and I have read every one of them, no matter how dense the scientese language I always start imagining them running around the cold, dark Arctic, wielding enormous butterfly nets, looking for and catching reindeer (or ptarmigans, whichever animal the paper is about) to do their research.

Reindeer_bw.GIFlepidopterist.gif

If I was not so averse to cold, I’d think this would be the best career in science ever!
It is no surprise that their latest paper – A Circadian Clock Is Not Required in an Arctic Mammal (press release) – was widely covered by the media, both traditional and blogs, See, for example, The Scientist, BBC, Scientific American podcast and Wired Science.
Relevant, or just cool?
It is hard to find a science story that is more obviously in the “that’s cool” category, as opposed to the “that’s relevant” category. For the background on this debate, please read Ed Yong, David Dobbs, DeLene Beeland, Colin Schultz, and the series of Colin’s interviews with Carl Zimmer, Nicola Jones, David Dobbs, Jay Ingram, Ferris Jabr, Ed Yong and Ed Yong again.
I agree, it is a cool story. It is an attention-grabbing, nifty story about charismatic megafauna living in a strange wilderness. I first saw the work from the lab in a poster session at a conference many years ago, and of all the posters I saw that day, it is the reindeer one that I still remember after all these years.
Yet, the coolness of the story should not hide the fact that this research is also very relevant – both to the understanding of evolution and to human medicine. Let me try to explain what they did and why that is much more important than what a quick glance at the headlines may suggest. I did it only part-way a few years ago when I blogged about one of their earlier papers. But let me start with that earlier paper as background, for context.
Rhythms of Behavior
In their 2005 Nature paper (which was really just a tiny subset of a much longer, detailed paper they published elsewhere a couple of years later), Stokkan and colleagues used radiotelemetry to continuously monitor activity of reindeer – when they sleep and when they roam around foraging.
You should remember that up in the Arctic the summer is essentially one single day that lasts several months, while the winter is a continuous night that lasts several months. During these long periods of constant illumination, reindeer did not show rhythms in activity – they moved around and rested in bouts and bursts, at almost unpredictable times of “day”. Their circadian rhythms of behavior were gone.
But, during brief periods of spring and fall, during which there are 24-hour light-dark cycles of day and night, the reindeer (on the northern end of the mainland Norway, but not the population living even further north on Svaldbard which remained arrhythmic throughout), showed daily rhythms of activity, suggesting that this species may possess a circadian clock.
Rhythms of Physiology
In a couple of studies, including the latest one, the lab also looked into a physiological rhythm – that of melatonin synthesis and secretion by the pineal gland. Just as in activity rhythms, melatonin concentrations in the blood showed a daily (24-hour) rhythm only during the brief periods of spring and fall. Furthermore, in the latest paper, they kept three reindeer indoors for a couple of days, in light-tight stalls, and exposed them to 2.5-hour-long periods of darkness during the normal light phase of the day. Each such ‘dark pulse’ resulted in a sharp rise of blood melatonin, followed by just as abrupt elimination of melatonin as soon as the lights went back on.
reindeer melatonin.jpg
Rhythms of gene expression
Finally, in this latest paper, they also looked at the expression of two of the core clock genes in fibroblasts kept in vitro (in a dish). Fibroblasts are connective tissue cells found all around the body, probably taken out of reindeer by biopsy. In other mammals, e.g., in rodents, clock genes continue to cycle with a circadian period for a very long time in a dish. Yet, the reindeer fibroblasts, after a couple of very weak oscillations that were roughly in the circadian range, decayed into complete arrhytmicity – the cells were healthy, but their clocks were not ticking any more.
reindeer fibroblasts.jpg
What do these results suggest?
There is something fishy about the reindeer clock. It is not working the same way it does in other mammals studied to date. For example, seals and humans living in the Arctic have normal circadian rhythms of melatonin. Some other animals show daily rhythms in behavior. But in reindeer, rhythms in behavior and melatonin can be seen only if the environment is rhythmic as well. In constant light conditions, it appears that the clock is not working. But, is it? How do we know?
During the long winter night and the long summer day, the behavior of reindeer is not completely random. It is in bouts which show some regularity – these are ultradian rhythms with the period much shorter than 24 hours. If the clock is not working in reindeer, i.e., if there is no clock in this species, then the ultradian rhythms would persist during spring and fall as well. Yet we see circadian rhythms during these seasons – there is an underlying clock there which can be entrained to a 24-hour light-dark cycle.
This argues for the notion that the deer’s circadian clock, unless forced into synchrony by a 24 external cycle, undergoes something called frequency demultiplication. The idea is that the underlying cellular clock runs with a 24-hour period but that is sends signals downstream of the clock, triggering phenotypic (observable) events, several times during each cycle. The events happen always at the same phases of the cycle, and are usually happening every 12 or 8 or 6 or 4 or 3 or 2 or 1 hours – the divisors of 24. Likewise, the clock can trigger the event only every other cycle, resulting in a 48-hour period of the observable behavior.
If we forget for a moment the metaphor of the clock and think instead of a Player Piano, it is like the contraption plays the note G several times per cycle, always at the same moments during each cycle, but there is no need to limit each note to appear only once per cycle.
On the other hand, both the activity and melatonin rhythms appear to be driven directly by light and dark – like a stop-watch. In circadian parlance this is called an “hourglass clock” – an environmental trigger is needed to turn it over so it can start measuring time all over again. Dawn and dusk appear to directly stop and start the behavioral activity, and onset of dark stimulates while onset of light inhibits secretion of melatonin. An “hourglass clock” is an extreme example of a circadian clock with a very low amplitude.
While we mostly pay attention to period and phase, we should not forget that amplitude is important. Yes, amplitude is important. It determines how easy it is for the environmental cue to reset the clock to a new phase – lower the amplitude of the clock, easier it is to shift. In a very low-amplitude oscillator, onset of light (or dark) can instantly reset the clock to Phase Zero and start timing all over again – an “hourglass” behavior.
The molecular study of the reindeer fibroblasts also suggests a low-amplitude clock – there are a couple of weak oscillations to be seen before the rhythm goes away completely.
There may be other explanations for the observed data, e.g., masking (direct effect of light on behavior bypassing the clock) or relative coordination (weak and transient entrainment) but let’s not get too bogged down in arcane circadiana right now. For now, let’s say that the reindeer clock exists, that it is a very low-amplitude clock which entrains readily and immediately to light-dark cycles, while it fragments or demultiplies in long periods of constant conditions.
Why is this important to the reindeer?
During long night of the winter and the long day of the summer it does not make sense for the reindeer to behave in 24-hour cycles. Their internal drive to do so, driven by the clock, should be overpowered by the need to be flexible – in such a harsh environment, behavior needs to be opportunistic – if there’s a predator in sight: move away. If there is food in sight – go get it. If you are full and there is no danger, this is a good time to take a nap. One way to accomplish this is to de-couple the behavior from the clock. The other strategy is to have a clock that is very permissive to such opportunistic behavior – a very low-amplitude clock.
But why have clock at all?
Stokkan and colleagues stress that the day-night cycles in spring help reindeer time seasonal events, most importantly breeding. The calves/fawns should be born when the weather is the nicest and the food most plentiful. The reindeer use those few weeks of spring (and fall) to measure daylength (photoperiod) and thus time their seasonality – or in other words, to reset their internal calendar: the circannual clock.
But, what does it all mean?
All of the above deals only with one of the two hypotheses for the adaptive function (and thus evolution) of the circadian clock. This is the External Synchronization hypothesis. This means that it is adaptive for an organism to be synchronized (in its biochemistry, physiology and behavior) with the external environment – to sleep when it is safe to do so, to eat at times when it will be undisturbed, etc. In the case of reindeer, since there are no daily cycles in the environment for the most of the year, there is no adaptive value in keeping a 24-hour rhythm in behavior, so none is observed. But since Arctic is highly seasonal, and since the circadian clock, through daylength measurement (photoperiodism) times seasonal events, the clock is retained as an adaptive structure.
This is not so new – such things have been observed in cave animals, as well as in social insects.
What the paper does not address is the other hypothesis – the Internal Synchronization hypothesis for the existence of the circadian clock – to synchronize internal events. So a target cell does not need to keep producing (and wasting energy) to produce a hormone receptor except at the time when the endocrine gland is secreting the hormone. It is a way for the body to temporally divide potentially conflicting physiological functions so those that need to coincide do so, and those that conflict with each other are separated in time – do not occur simultaneously. In this hypothesis, the clock is the Coordination Center of all the physiological processes. Even if there is no cycle in the environment to adapt to, the clock is a necessity and will be retained no matter what for this internal function, though the period now need not be close to 24 hours any more.
What can be done next?
Unfortunately, reindeer are not fruitflies or mice or rats. They are not endangered (as far as I know), but they are not easy to keep in the laboratory in large numbers in ideal, controlled conditions, for long periods of time.
Out in the field, one is limited as to what one can do. The only output of the clock that can be monitored long-term in the field is gross locomotor activity. Yet, while easiest to do, this is probably the least reliable indicator of the workings of the clock. Behavior is too flexible and malleable, too susceptible to “masking” by direct effects of the environment (e.g., weather, predators, etc,). And measurement of just gross locomotor activity does not tell us which specific behaviors the animals are engaged in.
It would be so nice if a bunch of reindeer could be brought into a lab and placed under controlled lighting conditions for a year at a time. One could, first, monitor several different specific behaviors. For example, if feeding, drinking and defecation are rhythmic, that would suggest that the entire digestive system is under circadian control: the stomach, liver, pancreas, intestine and all of their enzymes. Likewise with drinking and urination – they can be indirect indicators of the rhytmicity of the kidneys and the rest of the excretory system.
In a lab, one could also continuously monitor some physiological parameters with simple, non-invasive techniques. One could, for example monitor body temperature, blood pressure and heart-rate, much more reliable markers of circadian output. One could also take more frequent blood samples (these are large animals, they can take it) and measure a whole plethora of hormones along with melatonin, e.g., cortisol, thyroid hormones, progesterone, estrogen, testosterone, etc (also useful for measuring seasonal responses). One could measure metabolites in urine and feces and also gain some insight into rhythms of the internal biochemistry and physiology. All of that with no surgery and no discomfort to the animals.
Then one can place reindeer in constant darkness and see if all these rhythms persist or decay over time. Then one can make a PhaseResponse Curve and thus test the amplitude of the underlying oscillator (or do that with entrainment to T-cycles, if you have been clicking on links all along, you’ll know what I’m talking about). One can test their reproductive response to photoperiod this way as well.
Finally, fibroblasts are peripheral cells. One cannot expect the group to dissect suprachiasmatic nuclei out of reindeer to check the state of the master pacemaker itself. And in a case of such a damped circadian system, testing a peripheral clock may not be very informative. Better fibroblasts than nothing, but there are big caveats about using them.
Remember that the circadian system is distributed all around the body, with each cell containing a molecular clock, but only the pacemaker cells in the suprachiasmatic nucleus are acting as a network. In a circadian system like the one in reindeer, where the system is low-amplitude to begin with, it is almost expected that peripheral clocks taken out of the body and isolated in a dish will not be able to sustain rhythms for very long. Yet those same cells, while inside of the body, may be perfectly rhythmic as a part of the ensemble of all the body cells, each sending entraining signals to the others every day, thus the entire system as a whole working quite well as a body-wide circadian clock. This can be monitored in real-time in transgenic mice, but the technology to do that in reindeer is still some years away.
Finally, one could test a hypothesis that the reindeer clock undergoes seasonal changes in its organization at the molecular level by comparing the performance of fibroblasts (and perhaps some other peripheral cells) taken out of animals at different times of year.
What’s up with this being medically relevant?
But why is all this important? Why is work on mice not sufficient and one needs to pay attention to a strange laboratory animal model like reindeer?
First, unlike rodents, reindeer is a large, mostly diurnal animal. Just like us.
a1 reindeer.jpg
Second, reindeer normally live in conditions that make people sick, yet they remain just fine, thank you. How do they do that?
Even humans who don’t live above the Arctic Circle (or in the Antarctica), tend to live in a 24-hour society with both light and social cues messing up with our internal rhythms.
We have complex circadian systems that are easy to get out of whack. We work night-shifts and rotating shifts and fly around the globe getting jet-lagged. Jet-lag is not desynchronization between the clock and the environment, it is internal desynchronization between all the cellular clocks in our bodies.
In the state of almost permanent jet-lag that many of us live in, a lot of things go wrong. We get sleeping disorders, eating disorders, obesity, compromised immunity leading to cancer, problems with reproduction, increase in psychiatric problems, the Seasonal Affective Disorder, prevalence of stomach ulcers and breast cancer in night-shift nurses, etc.
Why do we get all that and reindeer don’t? What is the trick they evolved to stay healthy in conditions that drive us insane and sick? Can we learn their trick, adopt it for our own medical practice, and use it? Those are kinds of things that a mouse and a rat cannot provide answers to, but reindeer can. I can’t think of another animal species that can do that for us. Which is why I am glad that Stokkan and friends are chasing reindeer with enormous butterfly nets across Arctic wasteland in the darkness of winter 😉
Lu, W., Meng, Q., Tyler, N., Stokkan, K., & Loudon, A. (2010). A Circadian Clock Is Not Required in an Arctic Mammal Current Biology, 20 (6), 533-537 DOI: 10.1016/j.cub.2010.01.042

My latest scientific paper: Extended Laying Interval of Ultimate Eggs of the Eastern Bluebird

ResearchBlogging.orgYes, years after I left the lab, I published a scientific paper. How did that happen?
Back in 2000, I published a paper on the way circadian clock controls the time of day when the eggs are laid in Japanese quail. Several years later, I wrote a blog post about that paper, trying to explain in lay terms what I did, why I did it, what I found, and how it fits into the broader context of this line of research. The paper was a physiology paper, and my blog post also focused on the physiological aspects of it.
But then, I wrote (back in March 2006 – eons ago in Web-time) an additional blog post on one of my old blogs (reposted on this one here, here and here) in which I followed further, thinking about the data in more ecological and evolutionary terms, and proposing hypotheses following from the data that can only be tested in other species out in the wild. As you can see if you click on the links, this post did not receive much commentary.
Then, about a year ago, I received an e-mail out of the blue, from a researcher at the Cornell Ornithology Lab, essentially offering to test one of the hypotheses I outlined in that post. My first reaction was “sure, go ahead, I am happy someone wants to do this, but please cite the blog post as the origin of the hypothesis”… The response was along the lines of “no, no, no – we are thinking about working WITH you on testing this hypothesis”. Wow! Sure, of course, I’m game!
They already had preliminary data which they sent to me to take a look. They are coming from an ecological tradition and are very familiar with the ecological literature, some of which they sent to me to read. On the other hand, I am coming from a physiological tradition and am very familiar with that literature, some of which I sent to them to read.
A month or so later, one of them, Caren Cooper, came down to Chapel Hill. We met and, over coffee, spent a couple of hours staring at the data and discussed what it all means. Then we got started at writing the paper.
And now, the paper is out: Caren B. Cooper, Margaret A. Voss, and Bora Zivkovic, Extended Laying Interval of Ultimate Eggs of the Eastern Bluebird, The Condor Nov 2009: Vol. 111, Issue 4, pg(s) 752-755 doi: 10.1525/cond.2009.090061
In this paper – which is really a preliminary pilot study (who knows, we may yet get a grant to do more) – Caren and Margaret set up video cameras on a bunch of nests of Eastern Bluebirds (Sialia sialis). From the tapes they got times when the eggs were laid. The times were approximate. But the analysis gave us exactly the same result when we used the times when the nest was obviously empty before the bird sat on it to lay the egg, the times when the bird first got up to reveal the egg to the camera, and the mid-point between those two times.
I am not aware of anyone ever looking at timing of egg-laying in wild birds out in the field. There is a huge literature on timing of laying in quail and chicken (and some in turkeys) in the laboratory, but none I am aware of in wild birds. Most researchers, when asked when their species lays eggs are surprised at the question and answer something along the lines of “no idea, but we find the eggs when we come to check the nests in the morning, so perhaps over night, or at dawn?” So, this paper is a first in this domain.
What we have shown is that bluebirds, just like chicken and quail, have an S-shaped pattern of egg-laying patterns (see my older post for theory and graphic visualization).
The question is: how does a bird “know” when to stop laying? When is enough enough? When is the clutch (all of the eggs laid in one breeding attempt) complete? Most of ecological literature is focused on energetics: are birds getting hungry, have they depleted some important source of energy, etc.
But the circadian field looks for internal mechanisms. Running a circadian clock takes very little energy. Even when the animals are extremely hungry, the clock keeps ticking with no changes in frequency (if anything, the amplitude gets bigger, implying even more work!). Even when an animal gets very sick and is dying, at the time when many bodily functions start ceasing, the clock works until the very end. Being produced by a molecular feedback loop in which some reactions use and others release energy, and all of this happening in just a small number of brain cells, the clock is very energy efficient and does not require the organism to be healthy and well fed.
What is important in regard to circadian regulation of egg-laying is to understand that female birds have not one, but two circadian clocks. Let’s call one of them A and the other one B. Clock A is located in the brain (or retinae or pineal or some combination, depending on the species) and is sensitive to light: it readily entrains to a light-dark cycle. No matter what the intrinsic frequency of the clock may be (as uncovered in constant darkness conditions), it is forced to a frequency of exactly 24 hours by the entraining power of the day/night cycle.
Clock B, on the other hand, is intimately tied to reproduction. It is a result of an interplay between the clock in the brain and neuro-endocrine signals between the brain and the ovary (which may itself house its own part of the clock). Brain clock sends hormonal signals to the ovary. Those signals entrain the ovarian rhythms AND result in ovulation. Ovulation itself produces hormones that signal to the brain clock and entrain it. This feedback loop is in itself The Clock. This clock is light-blind and its intrinsic frequency is not 24 hours – it is around 26-27 hours in both quail and chicken, and almost two days long in turkeys.
These two clocks, A and B, interact with each other. Let’s imagine a hypothetical scenario in which clocks A and B are very tightly coupled. The external light-dark cycles that all the birds in the wild are constantly exposed to entrain the clock A to the exactly 24 hours period. Clock B, being tightly coupled to Clock A is then also forced to oscillate with a period of exactly 24 hours. What would that mean to the bird? She would be laying one egg per day, always at exactly the same time of day, every single day of her life: in spring, summer, fall and winter. She’d spend all her resources on making big yolky eggs every day. She would be sitting on a huge pile of eggs throughout her life. She would not be able even to move short-distance to a better nesting ground, let alone prepare and undergo a long-distance migration. Her eggs would be also hatching at the rate of one per day. Thus, she would have progeny of a variety of ages at all times, each age having different requirements for care or abilities to follow the mother around. Some hatchlings would freeze to death in winter, or starve to death at time when the food is scarce. Others would die from predation at times when they are highly visible (in the snow) or just because there are so many of them they cannot all hide under a bush.
An opposite scenario: clocks A and B do not interact with each other at all. In this case, A would be entrained to the 24 hour cycle of night and day. Clock B, being light-blind, would freerun with its own endogenous frequency, i.e., with a period of roughly 26-27 hours. Again, the poor bird would be laying one egg per day all of her life. The only difference is that the eggs would not be laid always at the same time of day, but scattered all over the 24-hour cycle. Both scenarios are obviously maladaptive to the bird.
But, oscillator theory provides a third scenario in which clocks A and B are only loosely coupled. There are phase-relationships between the two clocks when they are coupled: A entrains B. There are phase-relationships when the two are at odds: A inhibits B (and thus no ovulation happens). The phase-relationships are dependent on daylength: when the days are short in winter A inhibits B and no eggs are laid. When the days are very long in the middle of the summer (or in constant light) all phases are permissive to ovulation and the clock B can freerun with its own period of 26-27 hours.
But the interesting phenomenon happens in-between, once the length of the day gets just a little bit longer in spring, in normal breeding season. There is only a narrow zone of phase-relationships in which the two clocks are coupled – outside of that zone, ovulation is inhibited. Thus the clock A starts ticking at the beginning of that zone (e.g., at dawn in some species, at around noon in quail) and starts freerunning through it until it “phase-locks” with the clock A and, for a while, appears to be running with the period of 24 hours. But underneath, the pulses of hormones are gradually shifting later and later, just a little bit each day. Finally, these hormonal influences allow the clock B to again break free from the clock A, freerun some more until it gets out of the permissive phase – the feedback loop is broken and the ovulations stops. The clutch is over.
a3%20OVI%20-%20medium%20PP.jpg
The resulting pattern is S-shaped: early in the clutch eggs are laid a little bit later each day, the middle of the clutch appears entrained to the 24-hour cycle, and the last egg or two again are laid later until the egg-laying stops completely. In quail, which was bred for centuries for egg-production, the selection affected the strength of coupling between the two clocks. Thus, in photoperiods (daylengths) that are just barely longer than the ‘critical photoperiod’ (the minimal daylength needed to provide any permissive phases at all, thus the first daylength in spring at which the bird can start laying), quail will have S-shaped patterns but the middle portion, the “straight one” that is entrained, is artificially long – I have seen clutches lasting for two months and consisting of 60 eggs!
Birds out in the wild, where natural selection is likely to produce an optimal clutch-size (not a maximal one that humans prefer), may or may not use the same mechanism to determine how and when the clutch starts and ends. So, what we did was see if Bluebirds also show the S-shaped pattern that would suggest they do. And they do:
Condor image.JPG
The first egg in the clutch is laid earlier than the subsequent eggs. All the eggs in the middle (1-6 of them, not 30 – we collapsed them all into one “time-point” in the graph) are laid at about the same time, indicating entrainment of B by A (i.e., to the light-dark cycle). The second-to-last egg may be laid a little later, and the very last egg is laid much later. These results suggest that quail is not a weird unique animal, or that Galliformes (chicken-like birds) are different from other kinds, e.g.., Passeriformes (songbirds). The mechanism is likely the same – not dependent on external factors like food and energy, but a result of a fine-honed system of interactions between two circadian clocks.
Of course, this is just a first observational study, but the results are encouraging. Next steps would be to: a) improve the temporal precision of measurements by, perhaps, installing thermo-couples in the nests (there is a huge but short-lasting body temperature spike exactly at the time of lay), b) increase the sample size, c) compare the bluebirds living in three very different latitudes where both the weather conditions and photoperiodic changes are different to see how the natural selection shaped their responses, and d) do a comparative study of a few more species belonging to other groups. We’ll see if we’ll try to submit a grant proposal in the future.
Unfortunately, this paper is not Open Access. I wanted to send it to PLoS ONE, which I think is the best journal in the world and IS the future of publishing. But it was important for Caren and Margaret to publish in a journal that their peers consider important, and Condor is a fine little journal for this. So I agreed to go along with it.
Also, the listing of the original blog post in the List Of References, to my dismay, disappeared between the Provisional PDF and Final PDF versions. It is now linked to inline in the text, placing it down to the level of the dreaded “personal communication”, once again foiling our attempts to give serious science blogging some respect. Ah well….
Interestingly, I did not know when the paper came out. Apparently, it was published back in November. I learned about it a couple of days ago when I got a first reprint request from a researcher in Russia!
But hey, I am happy. I got a paper published. And now I am using my blog and social networks to promote it… 😉
Cooper, C., Voss, M., & Zivkovic, B. (2009). Extended Laying Interval of Ultimate Eggs of the Eastern Bluebird The Condor, 111 (4), 752-755 DOI: 10.1525/cond.2009.090061

Three Circadian Articles in PLoS ONE today

That is, among 20 new articles in PLoS ONE today. As always, you should rate the articles, post notes and comments and send trackbacks when you blog about the papers. You can now also easily place articles on various social services (CiteULike, Mendeley, Connotea, Stumbleupon, Facebook and Digg) with just one click. Here are my own picks for the week – you go and look for your own favourites:
Distinct Functions of Period2 and Period3 in the Mouse Circadian System Revealed by In Vitro Analysis:

The mammalian circadian system, which is composed of a master pacemaker in the suprachiasmatic nuclei (SCN) as well as other oscillators in the brain and peripheral tissues, controls daily rhythms of behavior and physiology. Lesions of the SCN abolish circadian rhythms of locomotor activity and transplants of fetal SCN tissue restore rhythmic behavior with the periodicity of the donor’s genotype, suggesting that the SCN determines the period of the circadian behavioral rhythm. According to the model of timekeeping in the SCN, the Period (Per) genes are important elements of the transcriptional/translational feedback loops that generate the endogenous circadian rhythm. Previous studies have investigated the functions of the Per genes by examining locomotor activity in mice lacking functional PERIOD proteins. Variable behavioral phenotypes were observed depending on the line and genetic background of the mice. In the current study we assessed both wheel-running activity and Per1-promoter-driven luciferase expression (Per1-luc) in cultured SCN, pituitary, and lung explants from Per2−/− and Per3−/− mice congenic with the C57BL/6J strain. We found that the Per2−/− phenotype is enhanced in vitro compared to in vivo, such that the period of Per1-luc expression in Per2−/− SCN explants is 1.5 hours shorter than in Per2+/+ SCN, while the free-running period of wheel-running activity is only 11 minutes shorter in Per2−/− compared to Per2+/+ mice. In contrast, circadian rhythms in SCN explants from Per3−/− mice do not differ from Per3+/+ mice. Instead, the period and phase of Per1-luc expression are significantly altered in Per3−/− pituitary and lung explants compared to Per3+/+ mice. Taken together these data suggest that the function of each Per gene may differ between tissues. Per2 appears to be important for period determination in the SCN, while Per3 participates in timekeeping in the pituitary and lung.

Sleep Deprivation Influences Diurnal Variation of Human Time Perception with Prefrontal Activity Change: A Functional Near-Infrared Spectroscopy Study:

Human short-time perception shows diurnal variation. In general, short-time perception fluctuates in parallel with circadian clock parameters, while diurnal variation seems to be modulated by sleep deprivation per se. Functional imaging studies have reported that short-time perception recruits a neural network that includes subcortical structures, as well as cortical areas involving the prefrontal cortex (PFC). It has also been reported that the PFC is vulnerable to sleep deprivation, which has an influence on various cognitive functions. The present study is aimed at elucidating the influence of PFC vulnerability to sleep deprivation on short-time perception, using the optical imaging technique of functional near-infrared spectroscopy. Eighteen participants performed 10-s time production tasks before (at 21:00) and after (at 09:00) experimental nights both in sleep-controlled and sleep-deprived conditions in a 4-day laboratory-based crossover study. Compared to the sleep-controlled condition, one-night sleep deprivation induced a significant reduction in the produced time simultaneous with an increased hemodynamic response in the left PFC at 09:00. These results suggest that activation of the left PFC, which possibly reflects functional compensation under a sleep-deprived condition, is associated with alteration of short-time perception.

Regulation of BMAL1 Protein Stability and Circadian Function by GSK3β-Mediated Phosphorylation:

Circadian rhythms govern a large array of physiological and metabolic functions. To achieve plasticity in circadian regulation, proteins constituting the molecular clock machinery undergo various post-translational modifications (PTMs), which influence their activity and intracellular localization. The core clock protein BMAL1 undergoes several PTMs. Here we report that the Akt-GSK3β signaling pathway regulates BMAL1 protein stability and activity. GSK3β phosphorylates BMAL1 specifically on Ser 17 and Thr 21 and primes it for ubiquitylation. In the absence of GSK3β-mediated phosphorylation, BMAL1 becomes stabilized and BMAL1 dependent circadian gene expression is dampened. Dopamine D2 receptor mediated signaling, known to control the Akt-GSK3β pathway, influences BMAL1 stability and in vivo circadian gene expression in striatal neurons. These findings uncover a previously unknown mechanism of circadian clock control. The GSK3β kinase phosphorylates BMAL1, an event that controls the stability of the protein and the amplitude of circadian oscillation. BMAL1 phosphorylation appears to be an important regulatory step in maintaining the robustness of the circadian clock.

Clocks, sleep and non-visual photoreception on Scienceblogs.com

I am not the only one on ScienceBlogs.com to write about circadian rhythms, sleep and (non-visual) photoreception. Over the years, my SciBlings have written about these and related topics as well. Here is a sampler – go and dig for more on their blogs.
Stimulant Improves Sleep
Locked-In Syndrome
Opioids and Sleep Disorders
Home Testing for Sleep Apnea?
Pure Hypomanics: Living Zippedy Doo Dah Lives?
SFN Update: Sleep Deprivation Impacts Memory, Reduces Hippocampal Activity
Data Faker Turns Himself In
Agomelatine: A New Approach For Depression
Casual Fridays: Dave FINALLY finishes analyzing the procrastination data
Just Give ’em Some Nyquil While You’re At It
The Long, Long Sleep
Scientist to Men: Don’t sleep over
Two Stories on Sleep
Why do I feel like I’m falling when I go to sleep?
Power naps work in improving memory performance
Daylight Savings Time Affects Heart Attack Incidence
Don’t Trust an Insomniac
The Perverse Imp
Dolphins stay alert after five straight days of round-the-clock vigilance
The Night-Shift and Naps
Intranasal Orexin/Hypocretin: The Ultimate Uptime Drug?
Sleep behaviour and sleep postures
Sleep loss & false memories
Thinking During Sleep
The value of a good night’s sleep
Sleep, Sex, and Drosophila
Chloral Hydrate (Alkyl halide sleep aid)
Sleep and Heat
Make sure you get some sleep — or at least some caffeine — before that test
The point of sleep, or, Do fruit flies dream of six-legged sheep?
Portable brain activity-recorder shows that sloths aren’t all that sleepy
Lonely people have less efficient sleep
Friday Sprog Blogging: cross-country travel and kid circadian rhythms.
Ask a ScienceBlogger: A Sun Ray a Day….
Drink Your Milk! Go Outside and Play! You Just Might Live Longer
Day Two At SICB
Bears are in on the Hoax, Too
Living Clocks of Arctic Animals
How to evolve a watch
Circadian Clock Neurons
Time-Blind?
Evolution of vertebrate eyes
The eye as a contingent, diverse, complex product of evolutionary processes
Rhabdomeric and ciliary eyes
Medicine and Evolution, part 6: Ivan Schwab on eye evolution
Eyes, Part One: Opening Up the Russian Doll
Eyes, Part Two: Fleas, Fish, and the Careful Art of Deconstruction
Have fun for the rest of the week…

Yes, Archaea also have circadian clocks!

ResearchBlogging.orgIf you ever glanced at the circadian literature, you have probably encountered the statement that “circadian rhythms are ubiquitous in living systems”. In all of my formal and informal writing I qualified that statement somewhat, stating something along the lines of “most organisms living on or near the Earth’s surface have circadian rhythms”. Why?
In the earliest days of chronobiology, it made sense to do most of the work on readily available organisms: plants, insects, mammals and birds. During the 20th century, thousands of species of animals, fungi, protists and plants – all living on the planet’s surface – were tested for the possession of the circadian clock, and one was always found. Hence the “ubiquitous” statement seen in so many papers.
But, as it was later discovered, for some marine organisms moon cycles are more important than day-night cycles so they have evolved lunar clocks in addition or instead of circadian clocks (see sponges and cnidaria, for some examples). In the intertidal zone, the tides are more important for survival than the daily rhythms, so the organisms living there have evolved tidal clocks. Animals that live in caves have lost circadian rhythms, at least in behavioral output (a clock may still be operating underneath, driving metabolic or developmental rhythms). In the polar regions, rhythmicity may be seasonal. In subterranean animals, like Blind Naked mole-rats, most individuals are without rhythms, but young males that leave the colony in order to join another one develop rhythmicity during their adventurous journey. In social insects, only the individuals that go outside the hive to forage exhibit daily rhythms.
How does one figure out if an organism has a clock? You need to pick a good output and a way to continuously monitor it. Then you put the organism in constant conditions of light, temperature, air pressure, sound etc., and monitor the output for many days. If you do the statistics on the data at the end of the experiment and see that there is a periodicity in the data (for at least the first 2-3 days)that is reasonably close to 24 hours (between 16 and 32 hours is usually thought to be the limits), you know that your organism of choice has a circadian clock.
In a related experiment, you expose the organism to an environmental periodicity – usually a light-dark cycle, as it is usually the strongest cue, as the evolution of circadian clocks and light-detecting mechanisms is closely intertwined – to see if the rhythmicity of the organism can be synchronized (entrained) to the environmental cycle, indicating that it is a biological function and not the chance quirk in your data. Without these two experiments providing positive data it does not make sense to do any further investigations into mechanisms of entrainment, anatomical location of the clock or the cellular mechanism of the clock.
The trick is to find a good output to monitor. It is easy for rodents – they will run in running wheels (so will cockroaches). Songbirds will jump from perch to perch. Lizards will walk around the cage and tilt the floor from one side to another. And while behavioral output – the general locomotor activity – is not the most reliable (it is very prone to masking effects, so for instance mice will generally not run in wheels in bright light, while rats will), it is usually the easiest and cheapest to monitor and, in most cases (see an example where it failed, while monitoring body temperature worked) will be sufficient.
But what do you do when the organism does not have a measurable behavioral output, especially one that can be continuously monitored by machines? You start thinking very, very hard. And you come up with an alternative. You may be able to implant radiotransmitters and monitor body temperature. Or you may record vocalizations. Or you may take small blood samples several times per day and assay for something like melatonin.
The technological constrains limited our ability to discover circadian clocks in bacteria until the 1990s. Until then, the existence of such clocks was a mystery (one that everyone in the field was eager to see solved). I have written several posts about the discoveries of clocks in bacteria: Circadian Clocks in Microorganisms, Clocks in Bacteria I: Synechococcus elongatus, Clocks in Bacteria II: Adaptive Function of Clocks in Cyanobacteria, Clocks in Bacteria III: Evolution of Clocks in Cyanobacteria, Clocks in Bacteria IV: Clocks in other bacteria, Clocks in Bacteria V: How about E.coli? The understanding of the way bacterial clocks work (more like a relay or a switch than a clock) made us rethink the clock metaphor we have been using for almost a century.
So it appears that most Eukaryotes have clocks and at least some bacteria have them as well. But the other large group – the Third Domain: Archaea – eluded us thus far. After all, Archaea are notoriously difficult to culture in the laboratory and it took some time to figure out how to keep them alive outside of their natural extreme environments.
Do Archaea have clocks? We did not know. Until now. A couple of weeks ago, PLoS ONE published a paper that is the first to demonstrate the daily rhythms in an Archaeon: Diurnally Entrained Anticipatory Behavior in Archaea by Kenia Whitehead, Min Pan, Ken-ichi Masumura, Richard Bonneau and Nitin S. Baliga. Here is the text of the Abstract:

By sensing changes in one or few environmental factors biological systems can anticipate future changes in multiple factors over a wide range of time scales (daily to seasonal). This anticipatory behavior is important to the fitness of diverse species, and in context of the diurnal cycle it is overall typical of eukaryotes and some photoautotrophic bacteria but is yet to be observed in archaea. Here, we report the first observation of light-dark (LD)-entrained diurnal oscillatory transcription in up to 12% of all genes of a halophilic archaeon Halobacterium salinarum NRC-1. Significantly, the diurnally entrained transcription was observed under constant darkness after removal of the LD stimulus (free-running rhythms). The memory of diurnal entrainment was also associated with the synchronization of oxic and anoxic physiologies to the LD cycle. Our results suggest that under nutrient limited conditions halophilic archaea take advantage of the causal influence of sunlight (via temperature) on O2 diffusivity in a closed hypersaline environment to streamline their physiology and operate oxically during nighttime and anoxically during daytime.

What does that mean? What did they do?
First, they picked a good candidate species – Halobacterium salinarum. Why is it a good candidate? Because it lives near the Earth’s surface, in salty lakes and ponds (like this one, in Africa):
salinarium in a lake.gif
Many Archaea live in places where no light ever penetrates: deep inside the rock or ice or the oceanic floor. Some Archaea are exposed to light in cyclical fashion but not a 24-hour cycle – I have written somewhere before that I expect the Archaea living in the waters of the Old Faithful geiser in Yellowstone National Park to have a 45-minute clock instead. But Halobacterium salinarum is exposed to the natural periodicity of the day-night cycle on the surface and is thus a good candidate for an Archaeon that may have evolved a circadian clock. This is how the Halobacterium salinarum look like under the microscope:
salinarium micrograph.gif
There is another reason this is a good candidate. The light-dark cycle has a potential adaptive consequence to the critter. Water that is saturated with salt will have a high variation of its oxygen content and this variation is dependent on the environmental temperature: when it is colder outside, oxygen can more readily disolve in the salty water. When it is warm, it cannot.
The environment where Halobacterium salinarum lives is cold during the night and hot during the day. But the temperature changes are much more gradual and slow than changes in illumination (as well as less dependable: there are colder and warmer days), so being in tune with the light is a better way to synchronize one’s activities than measuring temperature (or oxygen content) directly. By entraining to a ligh-dark cycle, these organisms can make switches in their oxygen-dependent metabolism in a more timely (and thus more energy-efficient) fashion: by predicting instead of reacting to the changes in temperature over the course of 24 hours.
So, Whitehead et al placed some Halobacterium salinarum in light-dark cycles and subsequently released them into constant darkness. But what did they measure? Archaea are known to be lousy wheel-runners!
In bacteria, much of the work is done by measuring bioluminescence coming from the expression of the luciferase gene inserted next to one of the clock gene promoters. But here, we don’t know which if any gene is a clock gene and we do not have the technology ready yet. But, these days microarrays are cheaper and easier to use then some years ago when I started grad school. And remember that Everything Important Cycles!
So they took samples of the organism six times per day and ran them on microarrays, comparing the expression of all the genes between the sampling times, both during entrainment to LD cycles and in the subsequent DD (constant dark) environment:
archaea microarrays.JPG
What they discovered is that about 12% of the genes cycle with the period of 24 hours in LD cycles and continue to cycle in DD with a circadian period of around 21.6 hours:
archaea periods.JPG
What is most interesting is that the genes that cycle are the genes that are involved in oxygen (or oxygen-dependent) metabolism – exactly the kinds of genes that are expected to cycle in this organism. Some of these genes are also know to be directly regulated by oxygen. Now we know they can also be regulated – directly and/or indirectly through a clock – by light, inducing expression in preparation for the changes in oxygen concentration, not just in direct response. In this way, the cell is ready to use oxygen a little bit ahead of time. No time wasted.
I am very excited about this finding. This opens up a whole avenue of future research, something that the authors also realize:

Indeed, further detailed experimentation is necessary to ascertain precise phasing, temperature compensation, adaptability to different periods of entrainment etc. to ascertain the mechanistic underpinnings of this diurnal entrainment and its physiological implications.

Once we know there is a clock in Archaea – and now we do due to this paper – we can start studying it in detail.
Furthermore, this finding has big implications for the study of the evolutionary origins of the circadian clock (and light-reception associated with it). The molecular mechanism of the clock is very different between Bacteria and Eukaryotes, leading the field to conclude that the clock evolved independently in these two groups (and perhaps more – some people think that protist, plant, fungal and animal clocks evolved independently of each other as well). Now we can try to figure out how Archaea measure time. Is their mechanism similar to that in Bacteria? Or in Eukaryotes? Or something completely different, indicating another independent evolutionary origin? Or something in-between Bacteria and Eukaryotes, containing some elements of both, suggesting that perhaps there was only one evolutionary origin for clocks in all the life on Earth. The authors note that this last scenario is a strong contender:

Finally, the discovery of diurnal entrainment of gene expression in an archaeon also raises important questions regarding the origin of light-responsive clock mechanisms. This is because archaeal information processing machinery is assembled from components that share ancestry with eukaryotic (general transcription factors and RNA polymerase) and bacterial (sequence-specific transcription regulators) counterparts [44]. Furthermore, components of both bacterial [45,46] and eukaryotic [47] clocks are encoded in its genome [6,32].

Of course, since this is an Open Access article, you can and should read it yourself to get more details. And post ratings, notes and comments while you are there.
Whitehead, K., Pan, M., Masumura, K., Bonneau, R., & Baliga, N. (2009). Diurnally Entrained Anticipatory Behavior in Archaea PLoS ONE, 4 (5) DOI: 10.1371/journal.pone.0005485
Update: see comment thread for more. Unfortunately, scientists still at this day and age do not report everything and keep data secret. Apparently, this was the case in the question posed by this study. I hear from trusted sources that there is still not evidence for a clock in Archaea beyond the direct effects of light on gene expression and O2 metabolism.

Why social insects do not suffer from ill effects of rotating and night shift work?

ResearchBlogging.orgMost people are aware that social insects, like honeybees, have three “sexes”: queens, drones and workers.
Drones are males. Their only job is to fly out and mate with the queen after which they drop dead.
Female larvae fed ‘royal jelly’ emerge as queens. After mating, the young queen takes a bunch of workers with her and sets up a new colony. She lives much longer than other bees and spends her life laying gazillions of eggs continuously around the clock, while being fed by workers.
Female larvae not fed the ‘royal jelly’ emerge as workers.
Workers perform a variety of jobs in the hive. Some are hive-cleaners, some are ‘nurses’ (they feed the larvae), some are queen’s chaperones (they feed the queen), some are guards (they defend the hive and attack potential enemies) and some are foragers (they collect nectar and pollen from flowers and bring it back to the hive).
What most people are not aware of, though, is that there is a regular progression of ‘jobs’ that each worker bee goes through. The workers rotate through the jobs in an orderly fashion. They all start out doing generalized jobs, e.g., cleaning the hive. Then they move up to doing a more specialized job, for instance being a nurse or taking care of the queen. Later, they become guards, and in the end, when they are older, they become foragers – the terminal phase.
This pattern of behavioral development is called “age polyethism” (poly = many, ethism = expression of behavior), or sometimes “temporal polyethism” (image from BeeSpotter):
Age polyethism.jpg
This developmental progression in behavior is accompanied by changes in brain structure, patterns of neurotransmitter and hormone synthesis and secretion, and patterns of gene expression in the central nervous system.
Some years ago (as in “more than ten years ago”) Gene Robinson and his students started looking at daily patterns of activity in honeybees. The workers in their early stages are doing jobs inside the hive, where it is always dark. They clean the hive, take care of the eggs and pupae, and feed the larvae and the queen around the clock. Each individual bee sometimes works and sometimes sleeps, without any semblance of a 24-hour pattern. Different individuals work and sleep at different, apparently random times. The hive as a whole is thus constantly busy – there is always a large subset of workers performing their duties, day and night.
As they get older, they start doing the jobs, like being guards, that expose them to the outside of the hive, thus to the light-dark and temperature cycles of the outside world.
Finally, the foragers only go out during the daytime and have clear and distinct daily rhythms. Furthermore, the foragers have to consult an internal clock in order to orient towards the Sun in their travels, as well as to be able to communicate the distance and location of flowers to their mates in the hive using the ‘waggle dance’. As bees are social insects, it is difficult to keep individuals in isolation for longer periods of time, but it has been done successfully and, in such studies, foragers exhibit both freerunning (in constant darkness) and entrained (in light-dark cycles) circadian rhythms, while younger workers do not.
In the Robinson lab, then PhD student Dan Toma and postdoc Guy Bloch did much of the early and exciting work on figuring out how the rhythmicity develops in individual worker bees as they pass through the procession of ‘jobs’.
In an early study, they measured levels of expression of mRNA of the core clock gene Period (Per). The gene was expressed at low levels and no visible daily rhythm in early-stage workers, but at much higher levels and in a circadian fashion in foragers.
As the levels of expression were measured crudely – in entire bee brains – it was impossible at the time to be sure which of the two potential mechanisms were operating: 1) the celluular clock did not work until the bee became a forager, or 2) the cellular clocks were working, but different cells were not synchronized with each other, producing a collectively arrhythmic output: both as measured by gene expression of the entire brain and as measured by behavior of the live bee.
Either way, the study showed correlation: the appearance of the functional circadian clock coincided with other changes in the brain structure, brain chemistry and bee behavior. They could not say at the time what causes what, or even if the syncronicity of changes was purely coincidental. They needed to go beyond correlation and for that they needed to experimentally change the timing to see if various processes can be dissociated or if they are tightly bound to each other.
And there is a clever way to do this! First, they took some hives and removed all the foragers from it. This disrupted the harmony of the division of labor in the hive – too many cleaners and nurses, but nobody is bring the food home. When that happens, the behavioral development of other workers speeds up dramatically – in no time, some nurses and guards develop into foragers. And, lo and behold, the moment they became foragers, they developed rhythms in behavior and rhythms of the Per gene expression in the brain. So, as the development is accelerated, everything about it is accelerated at the same rate: gene expression, brain structure, neurochemistry, and behavioral rhythmicity.
Nice, but then they did something even better. They removed most of the cleaners and nurses from some hives. Again, the balance of the division of labor was disrupted – plenty of food is arriving into the hive but there is not enough bees inside to take care of that food, process it, feed the larvae, etc. What happened then? Well, some of the foragers went back into the hive and started performing the house-keeping duties instead of flying out and about. And, interestingly, their brain structure and chemistry reverted its development to resemble that of cleaners and nurses. They lost behavioral rhythmicity and started working randomly around the clock. And the rhythm of clock-gene expression disappeared as well.
So, genetic, neural, endocrine, circadian and behavioral changes all go together at all times. Social structure of the colony, through the patterns of pheromones present in the hive, affects the gene expression, brain development and function, and behavior of individual bees. Just like the gene expression and behavioral patterns, the patterns of melatonin synthesis and secretion in honeybee brains is low and arrhythmic in young workers and becomes greater and rhythmic in foragers. With the recent sequencing of the honeybee genome, the potential for future research in honeybee chronobiology looks promising and exciting.
But are these findings generalizable or are they specific to honeybees? How about other species of bees or other social insects, like wasps, ants and termites? Are they the same?
Other species of socials insects have been studied in terms of age polyethism as well. The earliest study I am aware of (let me know if there is an older one) studying behavioral rhytmicity in relation to behavioral development was a 2004 Naturwissenschaften paper by Sharma et al. on harvester ants. In that study, different castes of worker ants exhibited different patterns – some were strongly diurnal, some nocturnal, some had strange shifts in period, and some were arrhythmic. Those with rhythms could entrain to light-dark cycles as well as display freerunning rhythms in constant darkness.
Just last month, a new paper on harvester ants came out in BMC Ecology (Open Access). In it, Ingram et al. show that foragers have circadian rhythms (both in constant darkness and entrained to LD cycles) in expression of Period gene (as well as behavioral rhythms), while ants working on tasks inside the hive do not exhibit any rhythms either in clock-gene expression or in behavior, suggesting that the connection between age polyethism and the development of the circadian clock may be a universal property of all social insects.
We know that in humans, night-shift and rotating-shift schedules are bad for health as the body is in the perpetual state of jet-lag: the numerous clocks in our bodies are not synchronized with each other. We have evolved to be diurnal animals, entrained to environmental light cycles and not traveling over many time zones within hours, or working around the clock. Social insects have evolved a different strategy to deal with the potentially ill effects of shift-work: switch off the clock entirely until one develops far enough that time-keeping becomes a requirement.
Yang, L., Qin, Y., Li, X., Song, D., & Qi, M. (2007). Brain melatonin content and polyethism in adult workers of Apis mellifera and Apis cerana (Hym., Apidae) Journal of Applied Entomology, 131 (9-10), 734-739 DOI: 10.1111/j.1439-0418.2007.01229.x
Sharma, V., Lone, S., Goel, A., & Chandrashekaran, M. (2004). Circadian consequences of social organization in the ant species Camponotus compressus Naturwissenschaften, 91 (8) DOI: 10.1007/s00114-004-0544-6
Ingram, K., Krummey, S., & LeRoux, M. (2009). Expression patterns of a circadian clock gene are associated with age-related polyethism in harvester ants, Pogonomyrmex occidentalis BMC Ecology, 9 (1) DOI: 10.1186/1472-6785-9-7

Clock News

Two interesting new papers in PLoS Biology today:
A Role for the PERIOD:PERIOD Homodimer in the Drosophila Circadian Clock:

The current models of circadian clocks in flies and mammals involve the formation of complexes between clock proteins in the cytoplasm. These complexes are usually heterodimers (that is, made up of two different clock proteins) and appear to enter the nucleus at certain times of the circadian day in order to shut down their own gene expression by deactivating specific transcription factors. After progressive phosphorylation the repressor proteins eventually are degraded so that a new cycle of transcription can begin. Here we present evidence that in addition to heterodimeric complexes, the clock protein PERIOD (PER) also forms homodimers (pairs of identical proteins). Based on a structural model a PER mutant was designed, which is not able to form homodimers but can still bind to its partner TIMELESS (TIM). Flies expressing this mutant PER protein show abnormal clock function in regard to PER nuclear translocation, repressor activity, and behavioral rhythms. The circadian clock model in flies therefore needs to be extended by adding the PER:PER homodimer as a functional unit. Recent structural studies with mammalian PER proteins suggest that homodimers between clock proteins are an important general feature of eukaryotic clocks.

Structural and Functional Analyses of PAS Domain Interactions of the Clock Proteins Drosophila PERIOD and Mouse PERIOD2:

Most organisms have daily activity cycles (circadian rhythms), which are generated by circadian clocks. Circadian periodicity is produced by specific clock protein interactions and posttranslational modifications as well as changes in their cellular localization, expression, and stability. To learn more about the molecular processes underlying circadian clock operation in fruit flies and mouse, we analysed the homo- and heterodimeric interactions of the clock proteins Drosophila PERIOD (dPER) and mouse PERIOD2 (mPER2). We show that dPER and mPER2 use different interaction surfaces for homodimer formation, which are associated with different dimerization affinities. In addition, we present a structure-based biochemical analysis of the heterodimeric interaction of dPER with its partner Drosophila TIMELESS (dTIM). We identify a versatile molecular surface of the PERIOD proteins, which mediates homodimer formation of mPER2 but is used for dPER-dTIM heterodimer formation in Drosophila. Our results reveal quantitative and qualitative differences in the molecular interactions of PERIOD clock proteins in flies and mammals, allowing them to adjust to their different binding partners and regulatory functions in these different organisms.