Martin Chalfie (2015) - Tickling Worms: Suprises from Basic Research

So, I'm going to give a slightly different talk. If you came here thinking I was going to talk about green fluorescent protein, it's on the web. Go, please listen to it there. I'm going to take the advice of my colleague and friend Stuart Firestein who wrote this nice book published a couple years ago by Oxford University Press, called Ignorance. And his contention is that when scientists get together, what they really want to talk about is not all the stuff that they've done, but all the stuff that they don't understand. And so this is basically going to be the idea behind this talk. What are all the problems we have not solved and I'm interested in? It's also a shameless way to try to convince somebody in the audience that they might want to work in the lab. In any case, what I work on, what I worked on before GFP, after GFP and still are trying to figure out is this small animal caenorhabditis elegans. And particularly I've been interested in the sense of touch in this animal. Scientists have known for over 130 years how it is we detect the world around us as long as the signal outside of us is light. We know about the molecule rhodopsin. And for the last 30 or so years we've known of how we can detect signals outside of us, actually also inside of us, if those signals are chemical signals. So we know that receptors for sense of smell, taste, neurotransmitters, hormones, and we have a pretty good idea of the molecular mechanism. But the sense of touch and all the other mechanical senses that we have, hearing, balance, stretch of our muscles and tendons, detection of blood pressure, five or six different types of senses of touch in our skin, this is an unknown, we really don't know how that is. And so when I actually started my post doc, working at the LMB, I collaborated with John Sulston, shown here on this picture, to try to understand something about mechanosensation and we used C. elegans and these six blue cells here are the cells that sense touch. And you might be asking yourself, what sophisticated piece of equipment do we use to test touch in a one millimetre long animal? And the sophisticated piece of equipment is an eyebrow hair glued to a toothpick. You then take, you take this hair and you tickle the animal, you stroke it across the tail and it goes forward. You do it in the head, it goes backwards. You look for mutants that don't do that. You then have to distinguish mutants that don't respond to the hair from dead animals, but after you get over that it's pretty nice, and we've been able to get lots of mutants. And along the way, we know that these are the six cells that do sense touch in the animal. And we take a cross section of them and they have unusual microtubules so they have unusual structures, some extra-cellular matrix material. And all in all, over the years, we've collected well over a hundred mutant lines defining 17 genes that are needed for touch sensitivity. These are not all the genes that are needed, these are the ones we got. And these have been divided into two classes, five of them are genes that are needed to make the cells because you're going to get a touch insensitive animal if you don't have the cells. So you need to make the cells. But the others, the cells are made, they just don't work. And it was among them that we hoped to find the sensor for touch. So let's talk about development. There's a number of questions we want to ask and I'm going to give a little bit of the background, then to all the things that we don't know. So we want to know how cell fate is determined, how do we get these particular types of cells? The problem of differentiation. And it's not enough to just make a particular type of cell, you have to maintain that differentiation, you have to keep it in a differentiated state. So how does that cell fate maintain? And finally, there are some subtle differences in these cells. Take a look at these cells, and these cells. This pair, one on each side, arises when the animal is an embryo, this pair does the same thing. These have cell bodies and processes that go forward, same with these, but these have an extra process, and there's a couple of other subtle differences between them. How do the differences come about? What can we say? And another thing is why are there six of these cells, why aren't there 12 or 100? I'm not going to talk about that problem, we've worked on some genes that affect that. But let me talk about cell fate and how it's determined. Now I think this says something about maybe how we do experiments in my lab, slowly and intermittently. So in 1981, we collected all these mutants and among them were mutations in two genes, unc-86 and mec-3 that we thought were interesting. Unc-86 was interesting because when that was defective, there were no cells. So this was a gene needed to generate the cells that were going to become the touch sensing cells. But once those cells were made, if they didn't have mec-3, they didn't develop. They developed as other types of nerve cells. So they didn't have that special characterization. So without it, the cells aren't made. Mec-3 specifies the cell fate. We usually call this a selector gene, something that is the controller directing the cell to become the particular thing it is. And this was, turned out these were completely novel genes, they defined new types of proteins that were needed for this, which was fine. And we had a hypothesis, that I hope will start here. Whoops, there we go. That one turned on the other, and then those turned on all those subsequent genes that were needed for the cell. Well, took us another seven years before we started cloning things, and we did, we found that mec-3 was a transcription factor, which is a good for a selector. That it would turn on other genes. That would be good. And a year later we found it had a wonderful property. Not only did it turn, we think it turns on the other genes, it turns itself on. So this gave us an explanation for maintenance because a gene, once it was turned on, could continue to keep itself on by working on its own regulation. And we thought we had answered the problem. This was, now we were done. Not quite. So it didn't work. Then about 4 years later we found it's not A turns on B turns on the rest of it. It turns out that A turns on B, unc-86 turns on mec-3, and then they make a dimer and it's actually both of them together that turns on. Later, we did some other experiments that also answered this same sort of question, and then we thought we were done. We had explained everything, we had all the transcription factors, we were happy. And then 12 years later we did another experiment. And we found that there was one other transcription factor that was made by the cells. In fact, it was the target of all, it was one of the targets of all of this. And this transcription factor was really strange because when you didn't have it, you didn't, you got some animals that were perfectly normal, some that were a little bit bad, some a little bit worse than that, and some that were completely gone. Incredible amount of variability. So where does that variability come from? That was the question we wanted to ask. And we found that the variability came from the fact that this wonderful auto-regulation that we thought was perfect, is really pretty crummy. And the cell needs an insurance policy, and that's the insurance policy. And we called that refinement. And it keeps things on, and I'll say a little bit more about that because I'm sort of fascinated by that problem. And then, we then thought we were done. And then, this year, we found that this same action of refinement actually starts at the beginning of the process, where other genes help unc-86 refine its action so it can, with high fidelity, turn on mec-3. So a series of things. Let me just tell you, because it's a cute experiment and we didn't invent the method. Let me, whoops, there we go. Alexander van Oudenaarden, at the time was at MIT, he's now in the Netherlands, had developed a wonderful method that allows us to count individual messenger RNA molecules in cells. And Irini Topalidou, a post-doc at the lab at the time, decided to look at our cells and now we could actually count each one of the messenger RNA molecules. And what you get is a distribution. And without this refinement gene, you see that there's a big scatter. This is something that's known throughout biology. These are stochastic processes, so sometimes they work really well, sometimes they don't work really well. That could be a problem. Especially, it looks like a problem in this case because with alr-1, we reduce the stochastic nature here. So it is serving, this is why we said refinement, it changes it. It results in actually only a two fold increase in the transcription. And at first we though, oh. We didn't expect this result, this was really lunatic. What we expected was this result, that the distribution would shift to the right, we'd just get more molecules, but the distribution variability would stay the same. Instead, we got this refinement. How do we think about this? It gets back to that idea of a selector gene. A selector gene is one that somehow turns on a batterie of genes. This is a different type of transcription factor. So we have this binding. The processes that are going to be important here are that you have our dimer, there's going to be an on-rate, but that's going to be determined by diffusion. What's really going to be important is how well this binds and stays bound. Sometimes it's going to stay on forever and it's always going to work because it can activate, but there's a probability that it'll activate. And so the net effect is sometimes it'll be on all the time and work all the time and you'll get the maximum, and sometimes it's going to fall off and you're going to get very little, or it's not going to be very efficient. And we thought that what happened, and there is a site right next to where this binds for our other component, and that will change, it could change the off-rate, it could also increase the probability that this will be effective. But it can't go any better than the gene itself, the UM selector itself. And so we get this refinement. We've subsequently found that that same sort of refinement seems to be taking place here in the initial thing. Unc-86 is not good enough to turn on the gene we want, in fact a lot of times it fails. Other times it works, so we get two populations, the failures, and the non-failures. But when this is here, everything works. And now, subsequently, when this is on and we're having the maintenance phase, we have another thing that ensure this activity. So that is where we are at the moment. I'll tell you some problems we have with this in a second, but let me point to one other thing. How do we get these different types of cells? And one of the things that people have known for quite a while is that there's this wonderful series of transcription factors called the Hox transcription factors. First really described in Drosophila, and it's a whole linear chain. And each one seems to affect a different part of the body. And so we thought, oh that's wonderful, we'll have a different Hox gene for here, and one for here. Well, I've just told you, the Hox genes turn on the ability to be touch sensing neurons in both sets of cells. It turns out to be different genes, but they have two functions. One is to make sure these cells are going to differentiate into touch cells. They have a second function, the second function is they make this cells different from this cell. Actually it's more than that. They have very little effect on making this different, they change this cell. And we know the pathway of that, there's a gene that works back there, and one that works out here. We thought, ah this is great, we know which one it is. And then of course we did the other experiment to find that this gene, called C-13, works in the back, too. But it needs a co-factor and the co-factor's only in the front. So on the one hand you can turn on things, the other one you have another gene that's now turning on a battery of genes. What we don't know is: Is the gene working in the tail that causes all the difference is it working as a selector gene to turn on specific things, or is it a refiner gene that's helping some other selector gene that we don't know about yet? So one of the questions. So let me just talk about some of the mysteries that we, whoops. Let me see if that goes back, it doesn't. Ah, this is supposed to work. There we go. Come on. Hm, there we go. Some of the problems. I've told you about the control system. What we would really like to know is what does the control system turn on. So we know this in terms of the selector gene mec-3, but we don't know it in terms of these subtle differences. And we're started to try to go after that by looking for mutants that have shortened processes, and other problems that are only, other characteristics only in one cell or not. We also found another thing in these cells. When the animal hatches, the touch sensing neuron runs right next to the muscle. That's that green dot next to the pink muscle. But when the animal becomes adult, a funny thing happens: The skin, this blue stuff, starts to surround the cell and pushes it away from the muscle. And in a sense, it just completely ensheaths that nerve cell. Now in lots of biological systems, cells are ensheathed by other cells. And we realized, we could study this phenomenon by putting red fluorescent protein in the muscle and green fluorescent protein in the nerve cell and simply take the adults, or take animals like this and mutate them. And see if we could find mutants in which there's no separation. And we have now about 50 of them. We're still trying to sort out what this all means. So this is something that we're doing now, but are looking at. Other developmental problems. As I've sort of mentioned, how general is this refinement idea? Is this just something we happened to have gotten two examples of or are there many examples of this? Is this a general phenomenon to reduce stochastic? Now you can also realize that that's a wonderful way to play with development. Because if you have a stochastic system, one that you could switch from one to the other, one might be able to have the equivalent of our refining genes that would push things in one direction. Or sort of an anti-refining one that would push it in the other thing so it wouldn't work. So one could have all sorts of interesting controls of how cells differentiate by manipulating this system. And we'd really like to know if that's the case. How is it that some of the things act as selectors and some of them as refiners, how does that come about? We haven't a clue. What actually makes it so that that mec-3 selector gene is actually expressed in those particular cells? We have some ideas of what restricts it, but we have no idea, why it's actually turned on in the particular cells that it is. And as I said: What is all the stuff downstream that actually puts this into effect? We'd really love to know that. And there are other little questions like: What the hell are the microtubules doing in those cells? And how do they get organized the way they do? And finally, we really starting to embark on trying to answer the question of how are the synapses made in the cells. How do they connect to other cells, how do they recognize that as a problem? Now, I hope you've gotten an idea of, or I have the idea at least in this system, that what I really have stumbled into in this work that's really taken now decades and decades of my life, is I have found myself in Hershey heaven. Now most people don't know about Hershey heaven. It has nothing to do with chocolate, I'm sorry to say, but it has to do with this man Alfred Hershey who is reported to have said, that his idea of scientific heaven was I don't think this is what he meant. I think this is part of the quote. Because you do the same experiment over and over again, that's boring, right. What he really meant was, you do the same experiment over and over again, and you learn something new every time you do that. And that would in fact be scientific heaven. And as I hope I've indicated so far, these cells have given us many avenues to work on. This is a figure I made a while ago on our papers and different topics that we've studied. What I want you to get out of this is, this has taken us in many different directions that we don't do anything really in a linear manner. We'll be studying, for example, genetics and genomics and there will be big gaps in time. Or we'll study microtubules and there will be big gaps in what we do. It depends on who's excited or what problem and when a technology is developed and how we think about things, all of a sudden we go back and there's that gnawing problem in the back of our head. For those of you that are interested, that's GFP. I really want to talk a little bit now about transduction because that was one of the main problems we wanted to address. And among all those function genes that we were able to study, we have been able to find the transducer. This is a representation of all of the various function genes, but the transducer, the molecule that's sensing touch is a channel made up, it's a trimer, two of a protein called mec-4 and one of mec-10. They are very similar proteins and they made this channel that mechanical force opens. And we were able to demonstrate this a couple of years ago and so it gave us the first animal neuronal sensor of mechanical force. And so we were very proud that we were able to actually demonstrate this directly. But along the way, other finding this unusual type of sodium channel, we found a cholesterol binding protein in the membrane. That may also be a steroid binding protein. We found a protein that modifies microtubules and may be involved in their structure. And we found a new type of chaperone in the endoplasmid reticulum, and is actually needed to form and bring to the surface the channel for touch. And so, many different avenues that have come out of this. I don't want to talk about the transduction as much as I want to talk about what we were currently stumbling over at the time, and that is that we, all of our senses, you start off trying to find out what is the thing that does the sensing, but then very rapidly after that a different question comes up because all our senses are modified. You go from a dark room to bright sunlight, from sunlight in here, your eyes adapt. All of you, until I say the word will not be feeling your clothes that you're wearing because you have habituated to that. But once I say that you start to squirm and of course you feel that you are wearing clothes and you're relieved. In any case, our senses are modified. And so if they're being modified, how does that come about? At what level does this modification take place? And here again, we used a very sophisticated piece of equipment to study this, this was done by a graduate student. I had a, I should tell you, I had a wonderful idea. I mean it was brilliant. And that idea was, we've always gotten mutants that were insensitive to touch, but we've never gotten mutants that were super-sensitive to touch. And that was because our limits of detection, it was just really hard to do that. That's another one of the problems we'd really like to answer, getting super-sensitives. And I had an idea: Let's take the petri dishes and put them on car speakers and give them a low amount of vibration. And we'll vibrate them, we'll find a sub-threshold vibration and then find animals that respond to that sub-thresh. It was a brilliant idea, it did not work at all. But that meant we had a lot of car speakers, you know, car radio speakers in the lab and a much smarter person than me, graduate student Xiaoyin Chen, decided he was going to do something different. And that was, and I have to tell you for you graduate students out here, that please don't do this, you are not going to make friends this way: He set these things up, he gave them a, like a 50 hertz buzz and then left the lab to let people hear them for hours on end. He did not endear himself to them, they eventually banished him to a room that he could do this in peace. But what he found, and what we expected in the example I gave you about your clothes, is the example is you expect that there be insensitivity, that you would habituate. What we found however, is that the animals did habituate at first, but after a while they became sensitive again so they blocked the habituation. But they only did it for the cells in the front of the animal, not in the back. We've worked through the mechani. He also found four different ways that also just affected the front and not the back, but they had the opposite effect. Instead of turning on the system, they turned it down. All of these things took at least three hours, so they're not very quick. All of these are bad things, high salt, low oxygen, dauer is a dormancy state of the animal, things are bad. And he worked out the molecular mechanisms of these, and I'm not going to go through all the grizzly details, but he found that vibration worked through a completely different set of sensors on the cell membrane, activated this kinase and we go through this pathway. The end result is we get more channels. More channels means more sensitivity. But about all the negative things, the negative things worked differently. They didn't work on the cell, they interfered with a set of neurons that made one of the 40 insulins. He tested all 40, he found that this was the only one that was important, and these turn off this production. When this is turned off, this system gets turned on, we get ubiquitination of the channel, and it gets taken off the surface. Less channels, less sensitivity. And we also found with high salt, there was a different one of the 40 insulins. Now, this is interesting in many different ways. The first thing is that this is looking, when you look at this, it's actually saying that the nervous system, which we thought we knew everything about, all the connectivity, actually works through these neural hormones and so I call that the shadow nervous system. We actually don't know what's going on. But this part of the system actually integrates two different defects to give this. All three of these are integrated at this point in the cell. And all four of these inputs are integrated at that point. So there's a lot of signal comparison here to wind up with the net result of having the animals forward or back. I don't have time to actually go into why this is. I have slides to show that which I'll zip through. Why would one, would these worms want to turn up or down the nervous system? We think they're, because after all, it's unlikely that the worms in the wild are going to encounter car speakers. But they do encounter a situation where they get repetitive signals to them and that is rain. Because usually they go to rotting fruit, there's bacteria in there, and so rainstorms would give the long-term stimulation that should turn off the, should turn on the head sensitivity. Well why is that important? It turns out that the big predator of these worms is a fungus that has lasso. And the worms go in the lasso and it clamps on them. They never back up. So after a rainstorm, a protective mechanism would be, be really sensitive and get yourself out of this. And that's what it looks like it is, and there's been experiments that have shown that. There are other things: down regulating the front changes the complete balance of the nervous system. So that if you tap a plate when both the front and back are working, the animal goes backwards. But if you tap a plate when the head is reduced, it only goes forward. So these modulations actually change what we thought was a static nervous system into one that can respond to different behaviours. So let me do a lot of pressing here. And go to this and talk about some of the problems that we have. One of them is, we know of hundreds of neuropeptides in this animal. What are they doing? Do others, other than these two insulins affect touch? And we have suspicions that they do. We also know that these sensing cells express 19 different neuropeptides genes themselves, so the neuropeptides might be signals. We haven't a clue as to what they're doing. So what is this shadow nervous system? How does this map on to the circuitry we know from the chemical and electrical synapses? Don't have a clue. Are there other types of modulation? We've actually found a whole series of different types of modulation of touch. Some work on seconds, some work on minutes, some work more long-term like I've told you. We'd like to understand the molecular mechanism of all of those and how they integrate together. One funny thing about habituation. You notice how quickly when I said clothes, you knew that you had clothes on. So habituation works over a long time, but then is (snapping), can be almost instantly reversed. What's the mechanism that's going on there? And then the fundamental problem is, how in the world does that touch sensor work? So I'm hoping some of the physicists and engineers in the audience will be attracted to this problem and come up, because we would really like to know how a mechanical signal is sensed by these channels and how that works. And then extending that, how does touch work in humans? Is it the same sort of proteins, are they different, are there principles that are the same? I want to add one other thing. We're not really sure if what we found is connected, but you noticed I talked about the neuropeptides and I said that these were insulins. The animal has a lot of insulins. Well it turns out that one of the very first symptoms that people have of type two diabetes, it's the reason they go to their doctor, is they say, I'm numb in my fingers and my toes. So there's a thing called diabetic neuropathy. I used to think it was after the fact, after you've had the defect for a while. It turns out to be very early. Why is this important? Well it turns out it's the number one cause of amputation worldwide. So understanding some of that mechanism, I'm not saying that what we have is the same thing as in people, but it starts you thinking about how modulation might be important in those systems as well. And since I have no time, I'll just let you read this, my favourite quote about basic research and the importance of basic research, which was Robert R. Wilson's quote when he testified in Congress. Let me just read it, "It has only to do with the respect." They wanted to know what was good, the Fermilab Accelerator going to do for the national defence, it was in the middle of the Vietnam War. And he said: Thank you.

Martin Chalfie (2015)

Tickling Worms: Suprises from Basic Research

Martin Chalfie (2015)

Tickling Worms: Suprises from Basic Research

Abstract

Research, at least my research, has never been linear. I have found that my lab and I often double back on problems after years of inactivity or go off in entirely new directions as dictated by the work and people’s interests This lack of direction results, at least in part, from the fact that I am a geneticist and mutants have an annoying, yet wonderful, habit of leading one into new areas of study. I will describe how a simple assay to look for mutants in the nematode Caenorhabditis elegans that are insensitive to touch (stroking animals with an eyebrow hair glued to a toothpick) led me and my lab to investigate problems in cell determination, cell differentiation, mechanosensory transduction and modulation, and neural circuitry and the integration of sensory signals. Along the way, these studies resulted in the introduction of Green Fluorescent Protein (GFP) as a biological marker, several other methods, and maybe even some insights into a few human diseases. Although we actually have answered some of the questions we set out to study, the excursions far from what I thought I was studying have often been the most exciting.

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT


Related Content