My connection to Sonic Hedgehog

Some more pure science from guest blogger LisaJ:

Everyone seems to love a little Sonic Hedgehog around here. Whenever PZ discusses another function that this fascinating gene is capable of, much excitement ensues in the comment posts. So I thought I would take this opportunity to talk a bit about what I study, and how my seemingly unrelated favourite protein pathway is also connected to the Shh gene.

The main protein that I study was originally identified through studies of a pediatric eye cancer, called Retinoblastoma, as loss of function of this protein (termed the Retinoblastoma protein, or pRb for short) was found to be causative in the formation of this tumour type. What we know now is that pRb is required to control normal cell division in all cell types; its functional loss leads to the formation of many types of tumours, and is thought to be involved in the development of at least half of all human cancers. Not only is pRb essential in preventing uncontrolled cell division that can lead to cancer, but it is also essential for embryonic development, as mice deficient for the Rb gene die by about the 15th day of gestation, about 5 days before they would be born.

To explain how pRb functions in the cell, I thought that I could easily pull out a figure from a review paper that diagrams a simplified cell cycle pathway and the protein interactions that take place. But instead, I’m going to show you this beautiful piece of work that a few of my talented previous lab mates created for me one day, as a means of cheering me up after a year’s worth of protein purifications and binding assays did not give me the result I was hoping for. Although a simplistic depiction of cell cycle regulation, I think it really drives home the point well of how pRb functions.


This figure shows that pRb, or ‘RB Boy’ as is the case here, is stopped in his tracks in G1 (Gap 1) phase of the cell cycle. His watch and the stop light ahead of him are telling him to stay put for the moment until the cell’s DNA is fully repaired and ready to be duplicated, at which time he may initiate entry into the DNA synthesis phase (S phase). You’re probably wondering what that is sticking out of his pocket. Well, it’s an E2F transcription factor complex, and pRb binds E2F proteins through a domain that forms a pocket like structure. The manner by which pRb inhibits entry into S phase is by physically binding E2Fs while they are sitting on DNA; this binding interaction inhibits the activation of genes required for DNA synthesis to occur by E2Fs. Once any DNA damage has been repaired and the cell is ready to enter into S phase, a cyclin/cyclin dependent kinase (cdk) complex, shown here by the ‘cyclin/cdk bird’, phosphorylates pRb. This phosphorylation disrupts the physical interaction between pRb and E2F, and thus E2F is free to activate expression of pro S phase genes. These are the basics of how pRb regulates the cell cycle.

As I said previously, we know that pRb, by virtue of its regulation of E2F transcriptional activity, is a very important tumour suppressor. The role of these proteins in regulating tumour development has been, and still is, very heavily studied. We’re also discovering, however, that pRb has other important functions besides regulating the cell cycle and suppressing cancer. For instance, pRb/E2F proteins play an important role in regulating cellular differentiation and cell death, as well as the expansion of stem cells in a number of tissue types, including the haematopoeitic system, intestinal epithelium, and the brain (which I am most partial to). Additional functions that we’ve found for pRb/E2F in the brain include regulation of the differentiation of neurons and neuronal sub types, the migration of neurons through the brain, progenitor cell division, and the patterning of the ventral forebrain. We’re finding some very exciting functions for these proteins in the developing brain, and what’s even cooler is that the mechanisms through which pRb/E2F regulates neural development are emerging to be very similar to those that are involved in cancer development.

So of course, since the Sonic Hedgehog (Shh) gene does so many exciting things, we have also found an important functional link between E2Fs and Shh. When you knock out the E2F4 gene in mice and examine how embryonic development proceeds in the brain, you notice a very similar phenotype to what is seen when you knock out the Shh gene. The brain is quite reduced in size, and you see a large defect in the patterning of the ventral telencephalon (which is a part of the forebrain). Basically, the most prominent ventral structures, which are normally rich in neural stem and progenitor cells, are totally absent. Not surprisingly, these embryos do not survive to birth. You can see this gross defect in the figure below.

i-81292072bb7263fb08ba7897098aa236-Vlad fig 1.jpg

What you’re looking at here is an un-dissected embryonic mouse brain, at day 11 of gestation, in the first column, with the brain from a normal mouse on top and a much smaller E2F4 mutant brain on the bottom (TE in this figure stands for ‘telencephalon’). In the second column you are looking at a coronal, or frontal view cross section, of normal and mutant brains at the same embryonic stage; you can see that those bumpy structures (called the ganglionic eminences) present in the bottom half of the normal brain are almost totally absent in the E2F4 mutant.

PZ has discussed here before theimportance of Shh function in patterning the nervous system during embryonic development, by initiating its expression ventrally. What was ultimately determined about this observed E2FShh interaction is that the E2F protein is responsible for activating expression of the Shh gene within the ventral telencephalon at the correct time point during embryogenesis. Without this transcriptional activation, Shh is not adequately expressed and forebrain patterning is severely disrupted. This second figure shows really nicely how important the timing of Shh expression by E2F is.

i-b9d276aa8fa5c9b6fae3b77849519960-Vlad fig 2.jpg

You’re looking first at a side view, and then a frontal view, of a mouse embryo at day 9 of development. The wild type and E2F mutant mice in this experiment have been bred with mice that carry a transgene containing regions of the Shh promoter that are known to be essential for directing expression of Shh specifically to the brain regions that are affected in the E2F mutant. When these promoter ‘enhancer’ regions are bound by activating transcription factors, they direct expression of the beta-galactosidase gene, which is also present on the transgene downstream of the Shh enhancer sequence. Embryos are subsequently stained with a compound called X-gal, which is enzymatically cleaved by beta-galactosidase to produce a bright blue precipitate. Thus, the blue markings that you see in the figure indicate the locations that Shh is expressed in the ventral forebrain, and this expression is almost completely absent in the E2F4 mutant (lower panel). In the third column, you can see that one day later in the developmental time course, Shh expression is starting to appear in the E2F mutant mouse. It’s not an all or nothing effect here, but this seemingly short-lived dysfunction in Shh expression is what leads to the disastrous developmental effects to the animal that are described above.

This is a great example of two molecular pathways that were originally thought to do very different things converging to perform an essential, developmental function together. The number of gene products that function in our cells is huge, and it’s incredible to think of how immensely this complexity is increased when we find that many proteins are performing multiple functions on their own and in cooperation with other proteins or pathways. The wonders of biology never cease to amaze.

E2F/Shh figures from Ruzhynsky et al. Journal of Neuroscience. 2007. 27(22): 5926-35


  1. JoJo says

    Thank you. That was quite interesting. You explained it well enough that even an engineer could understand it.

  2. Beelzebub says

    Fascinating stuff. I wonder, given the combinatorics involved, how and whether we’re ever going to get complete command of these things. I realize we’ve only just begun to tease apart the intricacies, but will all of this ever seem to be old hat? For instance, will we ever be able to build computer simulations of these paths and make them coherent to ourselves or will it always be a mire of incomprehensible complexity? Perhaps this is just pessimism in the face of a daunting task.
    If nothing else it strikes me that cellular developmental biology is the very opposite to “design” in the ID sense. It seems it’s an interlocked and interdependent system that’s so subtle and complex that only evolution could have created it. Cellular structural biology has the appearance of design to it, but none of this developmental stuff.

  3. Logicel says

    Well done, Lisa. I was able to follow your description of the subtle interactions between various cellular components–I would encourage PZ to keep you on as a contributor for elucidating wondrous biology. The arty diagram made by your colleagues is fantastic (and fun).

    All I could come up with as an interesting (though probably completely offbase) sideline, is will this research lead a way into being able to get drugs through the brain barrier?

  4. Wayne Robinson says

    See wikipedia
    for the reason it’s called sonic hedgehog. In fruit flies absence of the gene causes the embryo to be covered with denticles like a hedgehog. In hedgehogs the 3 homologous genes are desert, indian (both named for species of hedgehogs) and sonic (named after a video game character, Sonic the hedgehog, which just goes to show that biologists can be just as infantile and retarded as engineers).

  5. Peter Ashby says

    In a previous lab of mine we had joint group meetings with an offshoot lab of the boss’s. They worked on pRB and pocket proteins and always gave the exact same intro to their talks to the point that our lab (embryonic development) could each have given it too. The problem was we were expected to attend when they were presenting but they didn’t bother when we did, so eventually we rebelled and joint meetings stopped.

    So it is nice to finally see pRB and pocket proteins doing something in an embryo. I can now stop blocking out the horror of endless band shift gels from tissue culture cells (our lab at least used embryo extract in some of our bandshifts). So thankyou Lisa for banishing that particular bugbear.

    Wayne Robinson perhaps worms may be more to your taste? In C. elegans genetics the rules are that gene names must follow a rigid alphanumeric naming schema. Making them very much harder to remember, which is the point of how fly and mouse people name their genes, they stick in the mind.

    My personal favourite apart from vertebrate fringes, is the cleopatra gene in flies, you see it is so called because it is only lethal when asp is present. What’s not to remember?

  6. negentropyeater says

    Thx Lisa. As a layman, I’m afraid it’s not all clear in my head and I’m a bit afraid to ask some questions that might just reveal how I’m getting everything mixed up.

    But I’ll still ask it and please tell me if I’ve really confused everything or if it’s understandable.

    If I understand, we’ve got different interactions between different gene products that come up with different functions and all this interacts within a certain type of cell. And then there’s different types of cells where some of those functions and really basic and general to most types and others more specfic. Does that mean that gradually, as we progress in understanding better all these interactions and their functions, we’re going to be able to draw some sort of functional diagram, or whatever is the correct term, for a complete cell of a certain type ? And then for more types of cells, other functional diagrams and etc…

    Is this correct or am I completely off the mark, and then if it’s still understandable, how long do you think it’ll take before we have such diagrams for a first type of cell, and then more, and then for more complex multicellular organsisms ?

    If nothing makes sense, just let me know, I really want to try to understand this stuff and it’s not clear if I’m spinning in the wrong direction and letting my mind far too clouded with all sorts of computer software analogies.

  7. Barklikeadog says

    When I went it was a %50-%50 split. Let’s do a regular pharyngulite and skew it, NOW!

  8. negentropyeater says

    personally, I prefer the term “pharyngulisation” (with s or z depending on the side of the pond you frequent most) to “Pharyngulite”.

    Agree with your point though, this is about the 6th time I see someone post this link in a thread since yesterday, PZMinions, we need blood !

  9. negentropyeater says

    to pharyngulize : verb
    pharyngulization : noun

    examples :
    – this poll has not yet been pharyngulized
    – this poll requires immediate pharyngulization

  10. says

    At the risk of making yet another redundant “Wow!” post, I’ll add to the replies here: Wow.

    I’m pretty sure that the reason the reply strings are shorter for posts like this is that most of us aren’t equipped to add more than “Wow” and “Thanks.” Myself, I have to add the few minutes’ time it takes to re-read and absorb enough of the content to know where to put it, roughly where to shelve it in my own chaotic and dusty mental library.

    Do you ever get dizzy for a second, drawing your attention back from this stuff to mere macroscopic, visible, ten-speed “real-world” processes, like turning off the lights and closing the door after work? I’d be tripping over my own feet on a daily basis, I think.

    Plus: Damn but this is fun. Thanks!

  11. Peter Ashby says

    Do you ever get dizzy for a second, drawing your attention back from this stuff to mere macroscopic, visible, ten-speed “real-world” processes, like turning off the lights and closing the door after work? I’d be tripping over my own feet on a daily basis, I think.

    It is certainly a good skill to be able to keep the forest in mind while you are peering intently at the twigs on the ends of the smalest branches of one tree, but it can be done. I was down there because I was interested in how the forest came to be and the twigs hold part of the answer. Some of my colleagues have been hard core molecular biologists for whom the twigs are the whole point. Myself and the other biologist in the lab tried hard to teach them some biology while they taught us molbio.

    The scale of the problem was indicated during my first week in the lab, I was sitting at a lab bench reading papers and getting up to speed (as you do, while gears grind) and the guys were clustered around a microscope behind me supposedly dissecting Day8.5 mouse embryos out of the packagin they come in. They were having problems and eventually I offered to look. Seems they couldn’t tell what was packaging and what was embryo. I looked down the scope and there was the head of an embryo sitting front and centre, plain as day to me surrounded by what is best described as bits of pink foam. The organisation of the embryo bits screamed out to me as being different but the molbio guys couldn’t see it.

    Eventually we all met in the middle, as we should.

  12. LisaJ says

    Thanks for your interest, and your nice comments everyone. I’m back form my sleep and my Sunday morning gym routine, so I’ll take a stab at some of your questions:

    Beelzebub @ #7: Actually, people are already building computer models of cellular signaling pathways. You are right, of course, that there is alot of work still to be done to tease apart all of the pathways and cross talk that are going on in cells, so who knows when we’ll have it sufficiently figured out. But computer models do exist for some of the information that we have now, and I’m sure will incorporate new findings in the future. As for whether or not we’ll ever have it all worked out, and be able to model every single interaction and pathway in specific cell types… well, that’s a tough one. Due to the manner in which basic research is largely conducted, with different groups studying their own isolated pathway, I think it will be tough to know if we’ve ever got it all down pat. Plus, there’s so much going on in there that it does seem unlikely that we’ll ever figure every last bit of it out. But you never know, science is always achieving what was once thought to be impossible.

    You are absolutely correct too about molecular cell biology making ID totally incomprehensible. No one would be able to think up something so complex. There has definitely got to be a much simpler way to get everything done in the cell, and any intelligent creator would have used that instead.

    Logicel @ #8: Thank you very much for your nice comment. To answer your question… no, my lab’s research really does not address the issue of getting drugs through the blood-brain barrier. However, this is an area of active investigation. My research is more directed at determining which molecules should be targeted by drugs in the future to repair lost or damaged neurons.

    Peter @ #10: I’m so glad I could help ease a little of the pain that your previous co-lab meeting members caused you. I know exactly what you mean about each member having the exact same introduction to each talk… sadly, I probably have the same starter slides as them, but I’m quickly able to move right into the brain stuff, gladly. However, I’m sorry to report that I’ve also done my fair share of cell culture band shifts in the past.

    negentropyeater @ #11: Great question. I think I’ve answered alot of it above in response to Beelzebub’s post. To answer your more specific question: yes, this can be done for specific cell types as well. But of course maps for each cell type are quite incomplete at this point, and I really have no idea when we’ll be satisfied enough to call them complete!

    Ron @ #16: Thanks alot for your post. I think you may be correct about why there are so many fewer replies to these sorts of posts. It definitely takes alot more brain power to put these posts together than most others, so it makes sense that it’s more difficult to make a comment as well. I enjoyed your last paragraph as well, very amusing :) Although the only time I really get dizzy is when I’ve been staring into the microscope for too long.

  13. Helioprogenus says

    As usual, these science posts are always so fascinating and well explained. PZ should have you minions posting regularly even when he’s back. If I were him, I’d be a little concerned about retaining my own blogs’ monopoly.

    In all seriousness though, I do have two questions.

    What is the signal or molecular pathway that allows a cell to “know” that it has repaired the damage and ready to enter the s-phase? Is it related to the cyclin dependent kinase complex?

    As far as the evolutionary history of Shh and E2F, when did the cooperation between E2F and Shh begin? Did they each start functioning in completely different systems, and then at some point, associate with each other in greater and greater ways? In other words, they must have arose independently of each other but at some point in time, started functioning cooperatively. Do we know the time frame for this, and if we don’t yet, where can we look in terms of evolutionary history and vertebrate (maybe even invertebrate) development?

  14. Peter Ashby says

    Helioprogenus the work Lisa describes does not show that shh (an intercellular signalling molecule) interacts with pRB (a protein that lives deep within the nucleus, three layers of membranes away from shh. In between are what used to be called ‘2nd messengers’ a sort of ‘here be dragons’ way of saying ‘we have no idea how signals effect cell behaviour’. Since I was taught that more than 20years ago we have mapped that blank region effetively using lots of radioactive phosphate, you find kinase No.1 (a protein that puts phosphates on other proteins) you purify it, mix it with cell extract in the presence of hot PO4, cook for a while then run the lot on a gel, cut out the hot bands and find out what proteins the hot PO4 was stuck to. If it was another kinase you repeat.

    Then you slowly link the whole thing into a meshwork of interacting proteins. I worked on analysing the phenotype of the pdk1 gene knockout mouse, a sort of node in the meshwork. The wonder was that we got any sort of embryo at all, they die much earlier than the embryos above.

    So basically the connection could have come about simply by a protein being turned on in a cell type/time/place it wasn’t normally and linking that cell surface signal to the nuclear protein by forming a bridge between two previously unlinked pathways. The relationship doesn’t happen in all the many other places shh is in operation in embryos because that bridge isn’t there. Perhaps the link enabled the cells to proliferate much faster, giving a bigger forebrain and hence an advantage. We don’t know yet.

    Since the link is only in mammalian forebrains it is unlikely to pertain in flies, or fish, or frogs. Be interesting to look in birds.

  15. LisaJ says

    Helioprogenus, thanks for your great questions.

    Your question of how the cell knows to when to re-enter the cell cycle after DNA damage has been repaired was answered very well in a previous post a couple of weeks ago by Doc. See post #103 there. Basically, p53 transcriptionally activates expression of some cyclin/cdk inhibitors (CDKIs), and once p53 has completed its DNA damage repair functions it is degraded, which leads to a loss of CDKI expression. This in turn leads to activation of cyclin/cdks and phosphorylation of pRb… then the cell starts cycling again. There is definitely more going on to initiate cell cycle re-entry, but this is believed to be the gist of what needs to happen.

    As for your second question. That’s a bit of a tough one, and we certainly don’t know all of the answers to it. What we do know is that E2F proteins are likely to be found in most eukaryotic organisms, and, as PZ has told us before, Shh is ubiquitous (except it’s not found in C. elegans). Yes, both pathways have evolved on their own, with their own distinct functions, and this interaction between them that we’ve demonstrated in the brain likely occurred sometime during the time course of mammalian evolution, or possibly even earlier. It is important to realize that these two pathways, from what we understand anyways, are still largely functioning very distinctly from one another. There may be other organs in which this interaction also occurs, but we have only demonstrated it for the forebrain. It is also important to note that this isn’t a protein-protein interaction that we’ve found, but an interaction in which the E2F protein is required to sit on the promoter site of the Shh gene and activate its expression. To really know the time frame of when this E2F4-Shh gene interaction first came about, we would have to study the brains of all organisms that have one, and determine when this first came to be.

    Thanks again!

  16. Helioprogenus says

    Thanks for the responses LisaJ and Peter Ashby. These molecular mechanisms are just so immensely fascinating, yet, sometimes really difficult to wrap one’s head across. What really grabs my attention in all this is the novel uses proteins find to direct genetic expression and initiate complex intracellular molecular cooperation.