What I've described so far is entirely transcriptional regulatory network. Just genes and proteins and how much or how little of they're expressed. What should be obvious is that there's a lot more going on in a cell, besides at the level of transcription. Transcription is just a way of getting proteins that can do the job of metabolism, and sensing, and responding to a given situation, responding to the outside environment. We can start to appreciate this, with this simple scheme. The genome gives rise to the transcriptome. The transcriptome is all RNA transcripts expressed under a certain condition at a different time. The transcriptome gives rise to a proteome, that would be all the proteins that are expressed at a different time, and the proteome gives rise to the metabolome which all metabolites that are expressed at a given time. You can probably continue on with this sort of concept naming almost all processes in a the cell, the phosphorylome, the acetylome and so on. You can see that there's a lot going on at these outer levels. One more point before going on, namely, that this is not a unidirectional process from genome to phosphorylome. Already between proteome, transciptome, and genome, the feedback loop shows interactions between levels. There are more of these to come. So what is known about rhythms at these different levels? In the liver for instance about 10% of RNAs will be rhythmic. They'll be high and low about once a day. The proteins from the same tissue will also show a pattern of ten percent circadenly rhythmic expression, but interestingly, these rhythmic proteins are not all derived from rhythmic RNAs. Rhythmic RNA can lead to constitutively expressed protein, if the protein is very stable, for instance, or to rhythmic protein. I suppose it can lead to no protein expressed at all, but this is quite difficult to demonstrate with experiments. The other situation, also occurs, namely non-rhythmic RNAs can lead to rhythmic protein expression or more commonly, to non-rhythmic expression. Similar statistics can be described for the metabolome. Rhythmic or constitutive metabolites derive from rhythmic proteins, and so on. We'll talk more about how this might occur in the next lecture on clock regulated outputs. Here, I want to discuss feedbacks from some of these higher levels back onto the molecular clock mechanism. The first example is phosporylation. Phosphorylation is the addition of big, negatively charged phosphate group to another molecule. It turns out almost 30% of proteins are phosphorylated. Maybe that seems like nothing special since it's so widespread, but it turns out that the state of phosphorylation determines how most proteins function, if they're active or inactive and that's obviously important. How does this figure into the molecular mechanism of the clock? The profile of clock protein that I showed you earlier in this lecture, and here it is again, is a polyacrylamide gel that separates the proteins based on molecular weight. The proteins are run from the top to the bottom, and the bigger proteins get hung up and moved more slowly. So, in addition to seeing a big change in the amount of protein over the course of the day, there's a big change in the size of the protein shown here. How can the size of a protein change? It can do so if modifications like phosphate groups are sequentially added on, over the course of the life of the protein. It turns out that this pattern is quite common for many clock proteins. They show massive phosphorylation, as part of their daily expression pattern. Why is phosphorylation an interesting mechanism for the circadian clock? Phosphorylation has a number of demonstrated functions on various molecules. It determines whether a transcription factor is on or off, obviously a central mechanism of the clock. It also can impact sub cellular localization. If a protein is not in the nucleus, it can act very effectively regulate transcription. Phosphorylation can determine how quickly the protein is degraded. Its half life, also important characteristics of a clock protein. In the case of clock proteins in particular this last aspect has been the focus of most work, with altered kinetics in the stability or half life of the protein resulting from abnormal phosphorylation. How do we know that phosphorylation is important for the clock? Mutations in the <i>casein kinase 1</i> gene lead to some of the most impressive clock mutant phenotypes that we've ever seen. The <i>casein kinase 1</i> tau mutant has a short free running period, and extremely early entrained phases. This has been described in mice and hamsters, and there's even a pedigree of extremely early chronotypes in humans, that have a mutation in their <i>casein kinase 1</i> gene. The take home message here, is that there's a lot of clock-regulated phosphorylation going on. Another way to think about it, is that this regulatory system of the cell, one that regulates massive amounts of proteins, is also a key aspect of the clock mechanism. This phosphorylation that we've talked about here, comes from <i>casein kinase 1</i>, but also from other kinases as well. Although much less studied the enzymes that remove phosphates called phosphatases, are also involved in regulating the circadian clock. Now I want to tell you about another mechanism of clock regulation, coming from one of the non transcriptional levels of the cell. To set this up, I have to talk about the clock molecule and its function as a transcription factor. Transcription factors are an interesting class of molecules. They all have not one job, but at least two. They have to bind a specific promoter DNA sequence, and they have to organize the transcriptional machinery to make the RNA. Two distinct tasks. And as a result, transcription factors are built of modules. As though they had been stuck together in the course of evolution. These modules are so robustly constructed that you can switch them by annealing the DNA binding sequence of one factor, to another's activation domain, and vice versa, to make hybrid proteins, and in so doing, change the DNA binding choices and the activation functions. This has been an important experimental tool for instance, either to map the structure of a particular protein, into discrete functions, or to borrow specific functions to use as reporters, in so-called two-hybrid screens, a method where interacting proteins can be fished out of a given cell. Where were we, what about CLOCK, and how it works as a transcription factor. We talked about CLOCK binding to specific DNA sequences, E boxes, and the promoters in rev erb and ror genes. CLOCK binds DNA by virtue of its bHLH domain. It binds to a collection of other proteins also, so this is partly the activator function, using its PAS domain. In addition to these domains, CLOCK has at least one more function, as a histone acetyltransferase. This means that CLOCK is an enzyme that labels histones, the proteins that are used in packaging DNA, with acetyl motifs. This is extremely common to find that transcription also involves local changes in the packaging of the DNA. It's thought that access to the DNA will generally result in more transcription whereas a lack of access will shut transcription down. In the case of CLOCK the system modification function is self contained. CLOCK also acetylates its binding partner BMAL1. If the acetylation by the CLOCK protein is involved in circadian rhythms then it might be expected that there would be a CLOCK regulated deacetylase, also involved in the molecular clock mechanism. It turns out that this function is contributed by SIRTUIN 1 or SIRT1. SIRT1 has been associated with longevity in yeast and it's been analyzed extensively as nutrient sensor within cells. It regulates many proteins. Its own function like that of many enzymes, depends on the cofactor NAD, or nicotinamide adenine dinucleotide, and NAD reflects the redox state of the cell. It turns out that NAD levels are rhythmic, and also the enzyme that controls their production, NAMPT, is clock controlled. So, by regulating production of NAD the activity of SIRT1 is regulated, and its activity on CLOCK, BMAL1, and PER are regulated. This is thought to be a feedback onto the clock mechanism for metabolic state. As we say, hindsight is perfect, and in hindsight, this is highly intuitive since an animal's, nutritional state is time of day dependent, and thus, it's at once highly predictable signal for the clock as a zeitgeber, and it's also something that the clock has to anticipate, cycles of feeding or feasting, and fasting. Where does this leave us with the molecular mechanism of the clock? Well, there was the first transcriptional feedback loop then there was a second one. And now one can be added that has the clock activators regulating NAD it's enzymes and SIRT1, with SIRT1 feeding back on the transcriptional feedback loop. Note that this last loop is bringing both metabolism of the cell and epigenetics, the modification of chromatin to the clock mechanism. The story that I just mapped out was worked out for mice. You're probably wondering what's the story with the mechanism of the human clock. Although there will certainly be some differences, there are mostly similarities so far. Of course the types of experimental methods to figure out the human clock mechanism are entirely different and they thus take very different forms. The first indications of what these genes would be came from pedigrees of families, that had inherited forms of extremely early sleep onset. There were at least a couple of families that showed half of the family tree going to bed at 8 in the evening, and getting up at around 4:00 in the morning, and the other half going to bed at around midnight, and getting up at 8:00. Very clear inheritance patterns. When the geneticists mapped where in the genome this misbehavior was encoded, the first two genes that came up were <i>period 2</i> and <i>casein kinase 1</i>. This provided a solid starting point to go into rhythmic cells, and manipulate candidate clock gene expression, and ask what effect that would have on other molecular cellular rhythms. In this way it was shown that the clock network that's been worked out in mice is largely similar in humans. As I indicated there will be many differences in the details and this is an active area of research. In human genetic studies on the general population, inheritance of sleep timing is also seen, but it's a much weaker trait than what was found in those extraordinary pedigrees. In studies on the general population, various naturally occurring polymorphisms in human clock genes, have been associated with traits called morningness and eveningness. But there's no good consensus on which genes, and which changes in them, might be regulating our sleep timing. [SOUND]
