The Lark-Mouse and the Prometheus-Mouse

Two interesting papers came out last week [from the Archives - click on the clock logo to see the original post], both using transgenic mice to ask important questions about circadian organization in mammals. Interestingly, in both cases the gene inserted into the mouse was a human gene, though the method was different and the question was different:

Turning a Mouse Into A Lark

The first paper (Y. Xu, K.L. Toh, C.R. Jones, J.-Y. Shin, Y.-H. Fu, and L.J. PtáÄek
Modeling of a Human Circadian Mutation Yields Insights into Clock Regulation by PER2. Cell, Vol 128, 59-70, 12 January 2007) is concerned with the human clock mutation that is responsible for FASBS (familial advanced sleep-phase syndrome), an extreme 'lark' phenotype that runs in families. People in these families wake up before dawn, around 4am or so, and are ready for bed as early as 7pm (I am wondering if their adolescents also experience a temporary phase delay? If not, that would be quite informative about the mechanism of that type of developmentaly-mediated delay).

The paper is humongously long with a whole series of complicated experiments. Even press releases and media reports are tough slogging! Let me try to point out just the key findings here.

First finding is that inserting the mutant human Per2 gene into the mouse results in a similar phenotype - a phase-advance of the rhythm in light-dark cycles and a shorter period of the rhythm in constant darkness. Mice are nocturnal and transgenic mice wake up before dusk and go to sleep before dawn - much earlier than their wild-type conspecifics.

This finding to some extent alleviates the Number One concern: why should we expect the human gene to behave the same in the mouse genetic background? In the human, Per2 gene interacts with all the human versions of other genes, including the core clock genes. All mouse genes are somewhat different, so the interactions are expected to be different. The similarity in phenotype suggests that the differences between the sequences of clock genes in these two species are not crucial for their function. All the core clock genes appear to readily react with each other in this kind of a genetic chimaera.

Examining the mutant mouse allowed Ptacek and colleagues to uncover both the cause of hPer2's hypophosphorylation and its consequences. They found that the single amino acid substitution in the altered protein blocks a cascade of multiple phosphorylations that would normally take place. The effect of this is to dampen the machinery that feeds back to the hPer2 gene, decreasing production of the hPer2 protein. It is this decreased production of hPer2 that shifts the sleep-wake cycle.

The finding was unexpected, said Ptacek. "Based on the common model of hPer2 function, we had a completely different prediction about what the consequences would be," he said. "We were confident that the hypophosphorylation would lead to slower degradation of the protein. Instead, we show in these experiments that normal phosphorylation of the protein is required for increased transcription of the gene."

The finding was actually consistent with another study of regulation of Per function that I blogged about in great detail here (and for even more background and the connection of this mechanim to bipolar disorder and lithium, read this).

Briefly, mutation in Per2 results in the loss of a phosphorilation site on its protein. The site is crucial, because only when it is phosphorilated, can additional sites also be phosophorilated.

Based on those studies, "everybody had thought a short or long period depended on a change in protein stability," she said. "That's how we thought the system should work. But this paper shows that is not the case. It comes back instead to the transcription level as the most important step."

-----------------------

Their studies in mice revealed that the amino acid change associated with FASPS, which alters the charge of the residue, alters the ability of PER2 to regulate its own transcription. PER2 presumably manages such regulation through interaction with other proteins since it doesn't bind DNA itself, they said.

The findings led the researchers to suggest a model of clock function in which cells sense changing PER2 levels over time, beginning a new daily cycle when a certain threshold is crossed.

The Prometheus Mouse, losing its liver every day and starting over at night

The second paper last week (Kornmann B, Schaad O, Bujard H, Takahashi JS, Schibler U (2007) System-Driven and Oscillator-Dependent Circadian Transcription in Mice with a Conditionally Active Liver Clock. PLoS Biol 5(2): e34 doi:10.1371/journal.pbio.0050034) used a really cool technique called TET-OFF. A gene and a tetracyclin-response element are placed into a viral vector which is injected locally.

What I really like about this technique is that the transgenesis is localized, i.e., one does not make an entire animal transgenic, only one little part of it. Second nifty thing about the technique is that the transgenesis is inducible. Feeding the animals doxacyclin (a version of tetracyclin), shuts down the gene - making this really into a genetic knock-out (TET-ON does the opposite, i.e,. induces the expression of the gene, something less desirable as the embryonic development occurs in the knock-out condition). The mouse develops normally and the gene is temporarily shut off in the adult during the experiment.

In this study, a TET-OFF construct for the gene Rev-ERB was inserted into the liver. Check the two links to my posts above (in the section about the FASBS paper) on the mechanism of action of this gene. In short, when Rev-ERB is cycling, the clock runs normally. When it is not expressed, the clock stops. If it is continuously expressed in large quantities, the clock 'jams' and stops as well.

Again, the paper is very long and complicated, and I'll try to extract only what is the most important take-home message from the study.

What they tested was the dependence of a peripheral clock in the liver on the systemic signals emanating from the central pacemaker in the brain (the SCN). By applying doxacyclin in various genetic backgrounds in whole animals and in liver slices, they discovered that the loss of the liver clock eliminates rhythmicity in almost 90% of the genes in whole animals, and 100% in slices. This means that there are about 10% of the genes that respond to systemic circadian signals. Interestingly, one of these is Per2, a core clock gene. Others are heat-shock proteins and some genes involved in the immune function.

So, a peripheral clock drives rhythms locally. In normal healthy animals, in each tissue it is the genes carrying the core functions of that tissue that exhibit circadian oscillations, while all the other genes are not-expressed or expressed at steady low levels. The central brain pacemaker entrains the peripheral clocks via neural and hormonal signals. This study shows that the central pacemaker can do more - directly entrain a subset of the genes even when the peripheral clock is not functioning.

So, the definitions of pacemakers and peripheral clocks change once again. Decades ago, it was thought that central pacemakers ran everything - the cell at the periphery were passive recipients of the temporal information. Work form the last ten years or so led to discoveries of greater and greater independence of the peripheral clocks, so the thinking changed - all the pacemaker does is entrain the other clocks. This study points out that the truth is somewhere in between - the pacmekaer entrains the peripheral cloks AND is also capable of direclty driving some of the functions in the periphery.

figure from here