Gradient Tracking: How Cells Drive Without Eyes

Top

00:00:12.17 Hi, my name is Jayme Dyer and I'm going to tell you about the work
00:00:16.17 I did as a PhD student at Duke University. I'm a basic cell biologist,
00:00:21.10 which means I'm interested in how cells perform complex behaviors
00:00:24.19 even though they don't have brains. And specifically what I did
00:00:27.23 for my PhD is study how cells are able to drive without eyes.
00:00:32.13 And I don't mean that they get into a car and they drive away,
00:00:35.04 I mean that they go in a particular direction to find a target.
00:00:38.13 Now a really good example of a cell type that can do this is
00:00:41.25 sperm. Sperm are really good at finding eggs, and yet, obviously,
00:00:46.15 they don't have a brain and they don't have eyes to see where they're
00:00:49.13 going. But they're really good at finding eggs. So, it turns out
00:00:53.13 they use a mechanism to find eggs that's very similar to what many
00:00:57.14 different cell types do. A process called gradient tracking, and I'll tell you
00:01:00.28 what is in a second. Another type of cell that can do this is white blood
00:01:05.16 cells. So, when I cut my finger this fall cutting sweet potatoes, I
00:01:10.12 opened the skin -- unintentionally, and introduced bacteria from the environment
00:01:15.01 into my body. Now the reason why I didn't get an infection from those
00:01:19.00 bacteria is because I have white blood cells whose job it is to go around my
00:01:23.08 body and engulf bacteria to remove them. Now, not all of the white
00:01:28.02 blood cells in my body live in my fingertip, they live all over.
00:01:31.03 And their job is to go to wherever the bacteria are whenever bacteria
00:01:34.25 get introduced. Now, they use a similar mechanism as the sperm
00:01:39.21 in order to find the bacteria, but again, they don't have eyes or brains.
00:01:44.04 So that mechanism is called gradient tracking. And to explain it,
00:01:47.02 I want you to imagine that you walk into your friend's house,
00:01:50.04 and your friend is baking cookies. Now you've never been
00:01:53.21 in his house before, so you don't know where his kitchen is.
00:01:56.01 But you can smell the cookies baking. So the first thing you're
00:01:59.28 going to do when you walk in is sniff, and you're going to sniff,
00:02:04.11 and you're going to go in the direction where there's more cookie
00:02:08.20 smell. So it turns out that the sperm and the white blood cells
00:02:12.04 do exactly the same thing. So in the cookies, the cookies are
00:02:16.15 actually when they're baking, they're emitting a chemical -- actually
00:02:20.19 a number of chemicals, that are diffusing away from the cookies.
00:02:23.28 So in the oven, the cookie chemical is diffusing away and as it gets
00:02:27.28 farther and farther away from the cookies in the kitchen and in the living
00:02:32.02 room, the number of molecules decreases because they're spread
00:02:35.17 out by diffusion. And so when you walk into the house, you smell
00:02:38.25 the cookie smell and the closer and closer you get, the stronger
00:02:41.25 that smell becomes. So the white blood cells and the sperm
00:02:44.19 are doing the exact same thing. Eggs and bacteria secrete
00:02:49.00 chemicals that diffuse away in the environment and the sperm
00:02:52.09 when it's trying to find the egg, is essentially sniffing that egg
00:02:55.23 smell. And saying, "Oh! That's good!" and trying to get to the egg.
00:02:58.19 So, to give you an example to show you a cell actually doing
00:03:04.13 gradient tracking. So let me go back, the gradient is the gradient of
00:03:08.27 the chemical, and the tracking is finding is tracking that target.
00:03:13.04 So here's a cell doing gradient tracking. This is a type of white
00:03:16.29 blood cell called a neutrophil. And that's the funny shaped cell
00:03:20.03 in the middle. And all around are these red blood cells because the
00:03:23.10 -- because this was taken from a blood sample. And the funny
00:03:27.16 shaped dumbbell cell is the bacterium. And you'll see that this neutrophil
00:03:32.12 is really good at tracking the bacteria. In fact, it's so good that
00:03:37.24 even as the bacterium moves around, the cell is able to
00:03:40.28 reorient, and that's actually a really important characteristic
00:03:43.21 for cells doing gradient tracking. The other thing you'll notice
00:03:47.10 is that the cell has a very distinct front and back. Oh! They got it!
00:03:52.03 They're very good at doing this. Okay, so like I was saying, the
00:03:56.29 cells that do gradient tracking have a very distinct front, you can see
00:04:00.09 that here. It's very different than the back of the cell. And that's
00:04:03.25 basically an indication that the cell has cell polarity, there's a
00:04:07.15 a difference from one side of the cell to the other. And I'm going to talk
00:04:11.02 a lot today about cell polarity, the front and the back. And these
00:04:14.27 cells when they're doing gradient tracking always put their
00:04:17.23 front up gradient, trying to get closer to the chemical smell.
00:04:22.26 Now, how is it that cells do gradient tracking? Well, like us,
00:04:27.22 they have noses -- sort of, they don't actually have noses,
00:04:31.00 they have proteins called receptors that bind specifically
00:04:34.06 to the chemical. And actually, what's interesting is that the receptors
00:04:38.02 that these cells have to sense the chemical is exactly the same
00:04:42.04 type of receptor that's in your nose when you sense the cookie
00:04:46.00 smell. So in this schematic, the cell has receptors all over it. And
00:04:51.12 the receptors bind specifically to the chemical. And what's different
00:04:56.05 for cells that do gradient tracking versus us, is that we only
00:05:00.21 have one nose. So when we walk into the room, we have to
00:05:03.23 sniff over here, and over here, in order to get a sense of where
00:05:07.10 the most cookie smell is. But cells, because they have receptors
00:05:10.23 all over them, they actually can take one sniff and get a sense
00:05:14.27 of where the gradient is. So they got receptors all over the
00:05:18.12 cell surface, and they can go *sniff* and get a sense of
00:05:21.25 whether there's more receptors bound to the chemical on this side
00:05:25.17 or this side. Now in this example, it's really easy because the most
00:05:29.13 number of receptors that are bound to chemical are on one side.
00:05:31.28 And so the cell knows it needs to go in that direction.
00:05:34.10 But in most cases, when cells are doing gradient tracking,
00:05:37.17 they're doing it in an environment that looks more like this.
00:05:40.00 Where there's a very small difference from one side of the
00:05:44.14 cell to the other, and the number of chemicals that are bound
00:05:47.10 to the receptors. So in fact, we know from studies in a variety
00:05:52.10 of cell types, that cells are good in doing gradient tracking
00:05:55.22 with as little as 1% difference in the number of receptors
00:05:59.13 that are bound to the chemical on one side versus the other.
00:06:02.18 And in fact, 1% is like if this side had 100 receptors bound
00:06:07.07 to the chemical, this side has 98 bound to the chemical.
00:06:10.13 So at this point, you might think the cells are just really
00:06:13.02 good at sensing the difference from one side to the other,
00:06:15.18 and we think they are. But there's one other complicating
00:06:18.19 factor. And that is, in my schematic, all of these chemical
00:06:22.11 molecules are stuck to the screen. But in reality, they're actually
00:06:26.10 bouncing around because of diffusion. And so these molecules are
00:06:30.03 falling off and coming back on and sticking to the receptor
00:06:32.17 and falling off. And that means that there's a lot of what I'm going to call
00:06:35.18 molecular noise. And that's a problem for the cell because
00:06:39.14 if the cell takes one big sniff at any one moment, then there
00:06:43.20 might be just because of noise and random chance, there might
00:06:47.28 be more receptors bound to receptors on the down gradient
00:06:51.14 side than the up gradient side. And so if the cell simply
00:06:54.26 took that snapshot, it would think it needs to go in the wrong direction.
00:06:57.26 So cells need some way to ask over time, what the concentration
00:07:03.28 is on one side of the cell versus the other, of the chemical that
00:07:07.08 they're sensing. And I was really intrigued by this question,
00:07:10.19 how do cells overcome molecular noise? We know they do,
00:07:13.15 they're really good at it, how do they do it?
00:07:16.19 So, that's the question I set out to answer. But I wanted to do it
00:07:20.03 in a system that was a little easier to work with than white blood
00:07:23.05 cells or sperm. So I decided to work with the budding yeast,
00:07:27.04 Saccharomyces cerevisiae. Now Saccharomyces cerevisiae
00:07:30.00 is the same type of yeast that you use to make bread or beer.
00:07:33.15 And is very genetically tractable, what that means is it's very
00:07:39.00 easy to make mutations or changes to the genes which allows me
00:07:42.28 to ask very specific questions about how is it that the cell
00:07:46.11 does specific things. Now there are a couple differences
00:07:49.24 between yeast cells doing gradient tracking, and I'll tell you how
00:07:53.17 they do it in a second, and white blood cells and sperm.
00:07:56.10 And one of the most notable is the fact that yeast cells
00:07:58.19 are a fungus. And that means that they have a thick cell wall
00:08:02.05 around them, sort of like plant walls but it's a different material.
00:08:04.28 But that means that they can't move, that will be important in a second.
00:08:08.13 So, when do yeast cells do gradient tracking? They do it when
00:08:13.21 they're mating. And if you've never heard about yeast mating
00:08:16.10 before, which it seems like most people haven't, because
00:08:19.05 most people don't think about yeast mating. Let me explain when it
00:08:22.04 happens. So yeast like you and I, can exist as diploid, so they have two
00:08:26.13 copies of every chromosome. And they can proliferate and make more
00:08:30.11 cells as diploids, just like how our cells can proliferate as
00:08:34.11 diploids. They can undergo meiosis, which is the process that we do
00:08:38.11 in our gonads to produce haploid cells that have one copy of
00:08:42.04 every chromosome. Now unlike us, where we have sperm
00:08:46.01 and egg, and those cells can't divide until they come together
00:08:50.08 to create a diploid, yeast cells can actually divide, they can proliferate
00:08:54.26 as haploids. So that means when you're making bread, in your
00:08:58.26 bread dough, there might actually be some haploid cells
00:09:01.19 that have only one copy of every chromosome, and they're just
00:09:04.08 dividing away and happily living their lives. So haploid cells
00:09:08.20 come in two types, a and alpha. And a and alpha cells eventually
00:09:13.21 at some point, if they find each other, they'll come together and
00:09:16.26 physically fuse, just like an egg and a sperm, in order for their nuclei
00:09:20.12 to come together to create a new diploid. But I told you that yeast
00:09:25.07 cells, because they're fungus can't move. So when they are
00:09:29.09 -- so how is it that they get near enough to each other to
00:09:32.15 physically fuse? And the answer is that they grow towards
00:09:35.29 each other. So if I'm a yeast cell and I have a mating partner over there
00:09:39.25 that I want to get to and we're not touching, I'm actually going to
00:09:43.02 grow, sort of like they're reaching out to get to their mating partner.
00:09:46.23 And again, how is it that they know where their mating partner is?
00:09:50.11 We know they're really good at orienting this growth in the right
00:09:54.10 direction, but they don't have eyes. So they use the exact
00:09:57.22 same process that white blood cells and sperm use, in that
00:10:02.02 they do gradient tracking. So it turns out that a cells constantly
00:10:06.08 secrete a small molecule that diffuses away from them, it's called a
00:10:09.14 pheromone. And alpha cells find that smell very sexy. And if they smell
00:10:14.20 that pheromone, if that pheromone binds to their receptors on
00:10:17.06 their cell surface, they'll make a projection up the gradient
00:10:20.19 to try to find that a cell that must be nearby. And vice versa,
00:10:25.29 a cells respond to alpha -- the alpha pheromone. So that
00:10:29.27 was the system I was going to use to study gradient tracking.
00:10:32.28 And just as a proof of principle, let me show you that yeast
00:10:36.11 cells actually can do gradient tracking. So in this movie,
00:10:39.23 I'm showing you a cells that I've put in a special slide called a
00:10:43.25 microfluidics device. And with that special slide, I'm applying
00:10:46.27 a gradient of pheromones, so that there's the most amount of
00:10:49.11 pheromone on top and the least at the bottom.
00:10:51.20 I'm basically tricking these cells into thinking that there's a really big
00:10:55.25 sexy alpha cell up here, and they should reach up to try to get
00:11:00.04 to it. Of course there aren't any alpha cells around. I'm just
00:11:02.29 applying the pheromone. So you'll see in this movie that as
00:11:06.20 the cells grow, they grow up the gradient. And in fact, there are
00:11:11.04 a few cells that can reorient. So they weren't -- so this one
00:11:16.06 wasn't able to go up gradient because a couple cells were in its
00:11:18.19 way. And this one went in the wrong direction originally.
00:11:21.04 So just like the neutrophil, the cells are able to reorient if
00:11:26.02 they go in the wrong direction. Turns out they can reorient
00:11:28.19 if the gradient changes. I'm not showing you that here. And
00:11:31.12 they have a polarized morphology. There's a front, so they're growing
00:11:36.05 only at the tip and the back, which is where they're not growing from.
00:11:39.17 So, just like the white blood cells, I now have a system that I can
00:11:43.23 study gradient tracking using yeast. Okay, so going back to the
00:11:49.04 problem of how cells deal with molecular noise. The problem
00:11:53.28 is that at any one moment, they might have the wrong sense
00:11:57.25 of which way the gradient is, because due to molecular
00:12:00.29 noise, there might be more receptors bound to the chemical
00:12:03.17 on the down gradient side than on the up gradient side.
00:12:06.13 So one of the ways that we think cells that can move like
00:12:10.08 neutrophils and sperm deal with that, is that they do something called
00:12:13.06 a biased random walk. And the idea here is that they take
00:12:16.26 a sniff and maybe they're right, that the most number of receptors
00:12:21.22 are bound to chemicals on the up gradient side, or maybe they're
00:12:24.10 not. But it doesn't matter, because more often than not,
00:12:27.26 the up gradient side will have the most receptors bound to the ligand
00:12:31.12 or bound to the chemical. And so even though they won't
00:12:35.29 always be going in the right direction, net they'll go up gradient.
00:12:40.07 So over time, they'll end up going in the right direction.
00:12:44.28 So that's a very nice theory about how cells that can move do
00:12:48.19 gradient tracking. But it doesn't work for yeast because like I told you
00:12:51.23 before, yeast can't move. So the first question that I have is
00:12:56.17 how do yeast do gradient tracking? They can't do a biased random walk
00:13:01.05 because they can't walk. So how do yeast do gradient tracking?
00:13:05.07 So to try to answer that question, the first thing I did was I looked
00:13:09.11 inside the cells. And I did that using a fluorescent protein.
00:13:12.08 So I fused GFP, the fluorescent protein, to one of the proteins
00:13:17.06 that localizes to the cell front, that protein is Bem1.
00:13:21.00 And Bem1 is what I'm going to call a polarity protein.
00:13:24.02 Obviously it's involved in the polarity of the cell. And in fact, it
00:13:27.08 and a host of other proteins, are localized in a patch -- and I'll talk more
00:13:32.14 about the patch in a minute, localized to a patch at the front
00:13:34.28 of the cell. And in fact, wherever that patch is, that's where growth
00:13:38.17 is happening. So in this movie, you see the cells, the outline
00:13:43.06 of the cells as they're growing in a gradient. So I'm applying
00:13:45.18 a gradient, so again there's pheromone from the top, highest at the
00:13:48.04 top. And here, we're looking at the fluorescent protein, Bem1,
00:13:53.23 which is fused to GFP. And what you'll notice, what I noticed
00:13:58.07 when I looked at this movie is that the polarity patch, the Bem1-GFP
00:14:03.01 patch of proteins is very dynamic, it seems to wiggle around a lot.
00:14:07.09 And in fact, in this montage, we're looking at just one cell.
00:14:10.09 And we're looking at the tip of the cell, and you can see
00:14:12.26 the polarity patch seems to wiggle from one side of the cell
00:14:15.22 to the other over time, as the cell is tracking a gradient.
00:14:19.11 So this immediately evoked for me the hypothesis that
00:14:23.11 maybe yeast do do a biased random walk. Except, of course,
00:14:29.20 the yeast cell can't do it. But what if the polarity patch, what if
00:14:33.08 the proteins that say "This is the front, this is where growth should
00:14:38.08 happen." What if the polarity patch does a biased random walk?
00:14:42.03 So the white blood cell itself can move, but what if in yeast
00:14:46.22 the polarity patch is wiggling around and wherever it is, that's
00:14:51.06 where growth is happening. And if somehow the polarity patch
00:14:54.21 was able to be biased by the concentration of pheromones on the
00:14:59.25 outside of the cell, then it would spend the most amount of time
00:15:03.15 on the up gradient side of the cell and where the polarity patch
00:15:07.08 is, that's where growth is happening. And so that would result in
00:15:10.15 net growth in the up gradient direction. So that was my hypothesis,
00:15:15.19 that the polarity patch does a biased random walk.
00:15:18.26 Now I'm a geneticist, so the way that I want to test this
00:15:22.12 hypothesis is that I want to mess with it. I want to mess up
00:15:25.09 the wandering. In fact, what I'd really like to do is stop
00:15:29.01 the wandering. Because if I'm right, that the polarity patch
00:15:32.14 does a biased random walk, then if I stop the polarity patch from
00:15:37.08 wandering, I should stop the cells from being able to orient
00:15:40.06 in a gradient. Now that's what I want to do and I'm actually
00:15:43.24 gonna do it. But I'm not going to get to it until the end of my talk.
00:15:46.28 Because in order to do that, I need to understand what causes
00:15:50.28 the polarity patch to wander in the first place. Because how else
00:15:53.29 am I going to know how to stop it from wandering? So the
00:15:56.20 second question that I'm going to answer first is, what causes
00:16:00.18 polarity patch wandering? So in order to answer this question,
00:16:05.26 I needed to do a lot of microscopy. I needed to film the
00:16:09.07 polarity patch wandering around. And it turns out that setting up
00:16:13.22 the special slide in order to film cells in a gradient is really
00:16:17.25 time-consuming. It's actually a real pain in the butt, and I'm
00:16:21.02 not a lazy scientist. I'm a practical one. I wanted to film a lot
00:16:25.11 of cells with a lot of different mutations, so that I could really
00:16:28.18 get at the heart of what causes a polarity patch to wander.
00:16:31.11 And I had this rate limiting step of having to acquire the
00:16:35.13 special microfluidics slide, having to set it up, and that
00:16:38.26 was just painful. So I thought, maybe I could use a simpler
00:16:43.08 system to study polarity patch wandering. So it turns out
00:16:46.19 if you treat cells with uniform pheromone, so I simply
00:16:50.12 squirt pheromone on the slide, now they don't think that there's
00:16:53.15 a mating partner over there or over there, because there's
00:16:57.00 pheromone everywhere. But that's okay, because it turns out
00:17:00.14 they still polarize, they just pick a random direction and go in it.
00:17:03.14 So when I treated cells with non-saturating pheromones, so about
00:17:07.20 the concentration of pheromone that they were seeing when
00:17:10.03 they were in a gradient, you can see that the polarity patch was very
00:17:13.29 dynamic as well. So these are cells when in uniform pheromone,
00:17:17.05 pheromone all over, and yet the polarity patch seems to be wandering
00:17:20.13 around a lot. And in fact, I could use a computer to find the
00:17:24.07 centroid of the patch over time. And so this is an example of some
00:17:28.29 tracks. I'm simply showing you for one cell, so in each color
00:17:32.28 is one different cell, the location of the polarity patch over time.
00:17:37.21 So you can see that it sort of wiggles around, but more importantly,
00:17:40.27 I could use this spatial information to quantify the wandering.
00:17:44.20 So I quantified it using a metric called MSD, mean square
00:17:49.07 displacement. And what you need to know about mean square
00:17:52.04 displacement is that it's simply a measure of how far something
00:17:55.01 has moved from the origin over time. And so I averaged a whole
00:17:58.17 lot of polarity patches from a whole lot of cells in this graph.
00:18:02.24 I'm showing you that in non-saturating uniform pheromone,
00:18:06.00 like the movie I just showed you, there is some wandering
00:18:09.00 of the polarity patch. But I wanted to be sure that it was similar
00:18:12.14 to the amount of wandering of cells treated with a gradient of
00:18:15.18 pheromone. And so I compared the number of wandering and
00:18:18.07 non-saturating uniform pheromone with the polarity patch
00:18:21.16 wandering in a gradient. And it turns out that they're almost
00:18:24.02 identical. Phew! So that gives me the freedom to work with
00:18:28.13 cells in uniform pheromone, where I can just squirt it
00:18:31.02 on a slide, which is much easier. And not have to deal with
00:18:34.02 always setting up the microfluidics chamber. So many of the
00:18:36.20 experiments I'll show you today, I did in uniform pheromone.
00:18:39.28 Alright, let's go back to the question, what causes polarity
00:18:43.08 patch wandering? So I came up with this hypothesis, that wandering
00:18:47.17 is caused by vesicle delivery. And in order to understand
00:18:50.14 this hypothesis, you need to understand a little bit about the
00:18:52.26 dynamics of the polarity patch. So I'm showing you down here,
00:18:56.19 simply a fluorescent image of a cell in pheromone, outlined with
00:19:00.28 the dotted line, and the fluorescence, that's Bem1-GFP.
00:19:04.10 So you can see that there's sort of a patch of proteins. And that's
00:19:07.09 what I'm indicating with this graphic here. There's the plasma
00:19:10.08 membrane and then lots of proteins at the membrane.
00:19:15.07 Now, I don't want you to think of this as a raft of all these
00:19:19.15 proteins that are all stuck together, because that's not at all
00:19:22.13 what it's like. In fact, there's two important things going on here.
00:19:25.13 One is that all of these proteins in the patch are actually
00:19:29.09 moving around on the plane of the membrane. None of them
00:19:33.05 are in the membrane, none of them are stuck in the membrane
00:19:36.08 like integral membrane proteins, they're all peripheral membrane
00:19:39.07 proteins. So they're sticky on the surface of the membrane,
00:19:42.21 and they're all sliding past one another. The other thing that I need
00:19:45.27 to point out here is that not every single polarity protein is
00:19:49.18 at the patch. Some of them are down here in the cytoplasm,
00:19:52.16 they're all diffusing around. And importantly, some of them
00:19:57.07 are falling off of the patch, and some of them are actually
00:19:59.28 sticking to the patch or coming on and joining the patch.
00:20:02.21 And an analogy here is if there's a crowd in a park, there's a big
00:20:07.19 crowd all gathered together, so there are some people who are going
00:20:11.08 to come and join the crowd, and there are some people who are
00:20:13.10 like, "Yeah, I'm done here." And they leave the crowd.
00:20:15.14 So the crowd is like the polarity patch, it's this grouping of
00:20:19.16 people that are all sort of milling about, and there are some
00:20:22.16 people who are leaving and there are some who are joining.
00:20:24.26 Okay, so let's get back to the hypothesis, that wandering
00:20:28.09 of the polarity patch, this cluster of proteins is moving around
00:20:32.24 when the cell's trying to find a mate. And I think that it's moving around
00:20:36.03 because of vesicle delivery. So remember, I told you wherever
00:20:39.17 the polarity patch is, that's where the cell is growing.
00:20:42.02 And the reason that the cell is growing there is because there are
00:20:45.11 proteins in the polarity patch that can make actin cables
00:20:50.15 form at that site. So these purple things are actin cables, they're
00:20:53.21 filaments of actin, which is part of the cytoskeleton.
00:20:56.27 And you can think of them sort of like railroad tracks.
00:20:59.15 In that they're used to traffic vesicles. And vesicles are the stuff
00:21:07.03 that the cell needs to actually grow. So in order to grow, it needs
00:21:09.29 to add new cell material, right? And the vesicles are little sacs
00:21:13.26 with membrane on the outside and cell wall material on the inside.
00:21:17.12 So that when a vesicle -- so there's one more thing that's important
00:21:21.12 here, and that is that the vesicle is trafficked, it's literally carried
00:21:25.14 along the actin cable with a motor protein, so that's the little
00:21:29.28 sticks I'm indicating here. And this motor protein actually hangs
00:21:34.12 on to the vesicle on end and actually walks along the filament.
00:21:39.15 And it walks along the filament towards the polarity patch,
00:21:44.01 and it delivers the vesicle at the plasma membrane. And when
00:21:47.01 that vesicle gets to the plasma membrane, it fuses at the
00:21:51.04 membrane, okay? So it's going to fuse at the membrane,
00:21:55.11 and add just a little bit of new membrane. Now, it's important
00:22:00.23 that the polarity proteins which associate with the plasma
00:22:03.25 membrane, they are not associated with the vesicle membrane.
00:22:07.05 So when this vesicle adds to the plasma membrane, it's
00:22:12.00 basically adding naked membrane. So there's this sort of
00:22:15.24 open space, just transiently, for a very short amount of time.
00:22:19.22 Because of course, remember, all these proteins are just sliding
00:22:22.20 around. So some of these proteins are going to slide into that open
00:22:25.13 space. But remember also, that there are proteins that are falling
00:22:28.22 off and coming back on. The people who are joining the crowd
00:22:31.28 and coming away from the crowd. And you can imagine that
00:22:34.10 there's a big crowd in a park, you know there's sort of an
00:22:37.22 empty space on one side, you're less likely to go to that empty
00:22:41.20 space. And the same is true for these proteins. That when they
00:22:44.11 come back on from the cytoplasm onto the polarity patch,
00:22:48.07 they're much more likely to come on where there already
00:22:51.13 are some proteins. And so in this example, when new proteins get
00:22:56.07 added, in the second right after the vesicle got added, the new
00:23:00.17 proteins are more likely to be added here, closer to where there are
00:23:04.27 a lot of proteins than over here, where there are fewer proteins.
00:23:07.28 So there's going to be more added off-center. Now imagine that you're
00:23:11.29 looking down at the crowd, new people get added a little bit
00:23:15.21 off-center from where the center of the crowd was before.
00:23:19.03 And so now the crowd has shifted over a little bit. Now it's not
00:23:23.02 like all of the people in the crowd have moved over, right?
00:23:25.20 It's just that some new members have been added a little bit
00:23:28.05 off-center. Now, when a cell is doing gradient tracking and reaching
00:23:32.20 out to get its mating partner, it's adding a lot of vesicles so that it can
00:23:37.06 grow to meet its mating partner. Which means that one little
00:23:41.02 vesicle probably isn't going to have much effect, but lots of
00:23:44.06 vesicles being added might be enough to cause the polarity
00:23:48.05 patch, the centroid of it, to shift around. Which would look like
00:23:51.19 movement to us when we look down the microscope.
00:23:54.08 So that's the hypothesis that wandering is caused by vesicle delivery.
00:23:57.23 So the way that I'm going to test it is by using a drug called
00:24:02.03 Latrunculin to depolymerize the actin cytoskeleton. So I'm basically
00:24:05.29 going to remove the railroad tracks, so when I treat the cells
00:24:09.09 with latrunculin, the vesicles are still present, and in fact they
00:24:12.25 can still fuse with the membrane, but they're no longer being
00:24:15.22 targeted to the site with the polarity proteins. So if vesicle
00:24:20.18 delivery is causing the polarity patch to wander, then when I
00:24:25.24 treat the cells with latrunculin, I would expect to see a decrease
00:24:28.23 in the amount of wandering. Whereas, if vesicle delivery had no
00:24:33.02 effect, then I'd expect to see no effect. Now before I show you
00:24:37.10 the data, I should mention that I did this experiment in a mutant
00:24:40.26 that has even more wandering than the cells treated with uniform
00:24:44.27 pheromone I showed you before. And I don't have time to explain
00:24:47.20 what that mutant is or why the cells have more wandering,
00:24:49.28 but the reason that I used that strain is because the polarity
00:24:54.00 patch wanders around so much that it allows me to look for
00:24:57.08 small differences in the amount of wandering, which would be harder
00:25:00.21 to see if it didn't wander as much. So in that mutant, I treated
00:25:04.25 it with latrunculin and I saw a decrease in the amount of wandering
00:25:08.23 relative to DMSO, which is the liquid that I applied the drug with. So
00:25:14.25 that was pretty convincing that, or at least it's consistent,
00:25:18.09 with the hypothesis that it's vesicles being added that causes the
00:25:22.07 polarity patch to wander around. But the problem is when you treat
00:25:25.27 cells with a drug, you might have pleiotropic effects, which means you might
00:25:31.02 be affecting lots of things that you don't know what they are.
00:25:34.03 And so, maybe the drug is affecting something else, and that's
00:25:38.20 causing the polarity patch to not wander, as opposed to it
00:25:43.24 affecting vesicles. So I wanted another way to specifically
00:25:47.08 affect vesicle delivery without having to use this drug.
00:25:50.28 So I used instead, a mutation to stop vesicle delivery.
00:25:55.10 Now remember I told you the reason that vesicles traffic
00:25:58.10 along the actin cables is that there's a protein, this with the little legs here,
00:26:04.05 that the motor that hangs on the vesicle on one side and walks
00:26:08.11 along the filament, so that motor protein, I used a mutation
00:26:12.23 in it. A temperature sensitive mutation, so when I put the cells
00:26:16.00 at a high temperature, the motor protein no longer functions.
00:26:19.15 So now, the actin filaments are still present but the vesicles are no
00:26:24.01 longer being delivered along the actin filaments to the polarity site.
00:26:28.08 And again, when I did this in cells with lots of wandering, the high wandering
00:26:34.17 mutant, we saw at high temperatures that cells with this temperature
00:26:39.10 sensitive mutation had less wandering than the wild type control
00:26:42.28 at the same temperature. So together, these data helped convince us
00:26:47.19 that vesicle delivery contributes to the wandering that we see
00:26:52.09 in cells that are trying to sense the gradient. So wandering is
00:26:56.26 caused by vesicle delivery. Good. We've answered question #2
00:27:00.04 so we can return back to question #1, how do yeast cells
00:27:04.26 orient growth up gradient? Like I told you before, I hypothesized
00:27:09.04 that it's the polarity patch that's doing a biased random walk.
00:27:12.17 The polarity patch is sensing what's the concentration over here
00:27:16.00 and over here and over here and over here, and letting itself be biased
00:27:19.23 by the concentration of pheromone. And if that's right, then
00:27:24.05 polarity patch wandering should be necessary for gradient
00:27:28.24 tracking. So what I want to do is stop polarity patch wandering.
00:27:33.01 And I mentioned that in the beginning, I'm a geneticist. So if
00:27:35.07 I were to stop polarity patch wandering, then I should somehow
00:27:38.12 be able to prevent the cell from being able to orient to the gradient.
00:27:41.25 Now I've told you already two different ways that I can impair
00:27:45.20 the wandering, I can treat the cells with a drug or I can use a
00:27:49.06 temperature sensitive mutation. But both of these ways
00:27:52.01 inhibit vesicle delivery, and what I want to do is I want to stop
00:27:56.14 polarity patch wandering, but allow vesicle delivery so that
00:27:59.23 the cells can grow. Because if the cells aren't growing, how do I
00:28:03.12 know which direction they're trying to go in?
00:28:06.14 So my goal is to somehow find a way to stop the polarity
00:28:10.23 patch from wandering around without stopping vesicle delivery
00:28:14.10 so that the cells are still able to grow and I can tell in which
00:28:17.27 direction they're trying to go. So when I thought about what it is
00:28:22.04 that causes the polarity patch to wander, I realized that it's the
00:28:26.25 fact that the vesicles are naked, right? So when a vesicle
00:28:31.15 gets added to the plasma membrane, it adds this hole, right?
00:28:35.19 This short-term space and all of the proteins that are going to be added
00:28:40.15 to the polarity patch get added to the side where there isn't
00:28:45.17 a hole. So what if there wasn't a hole each time a vesicle got
00:28:51.11 added? So I thought, okay, I'll make the vesicles not be naked.
00:28:55.24 What if I forced a polarity protein to be on the vesicle?
00:28:59.24 So I did that by fusing a protein that normally localizes to
00:29:04.17 vesicles, Snc2, to a polarity protein, Bem1. So polarity protein,
00:29:10.01 Bem1, normally localizes to a polarity patch. And that means
00:29:13.14 that because that fusion protein localizes to vesicles now,
00:29:16.21 because of the fusion to Snc2, that when a vesicle gets
00:29:20.06 added to the polarity site, now there's not a naked open
00:29:24.14 space, it's a space filled with Bem1, which is a polarity protein.
00:29:28.01 And so when new proteins get added, they'll get added all over
00:29:31.23 instead of avoiding that open space. So here's the evidence
00:29:37.18 that this fusion protein actually does reduce wandering. Again I'm using
00:29:41.09 the high wandering mutant strain, and when the cells have wild type Bem1 in them,
00:29:45.24 there's a lot of wandering that happens. But when the cells have
00:29:48.24 the fusion protein, there's basically no wandering. And the amount
00:29:52.13 of wandering that you can see here, this is actually equivalent
00:29:54.25 to the amount of growth that's happening. So the polarity patch
00:29:57.28 is moving only in so much as there's new cell wall material
00:30:00.27 and the cell is growing out. So this fusion protein really blocks
00:30:06.00 polarity patch wandering, but it doesn't stop vesicle delivery. And so
00:30:09.27 these cells can still form a projection and I can then measure
00:30:13.12 what direction they're trying to go in. So going back to my question.
00:30:17.24 How do yeast cells orient growth up gradient? Well, I think
00:30:20.16 the polarity patch wandering is important and so I'm going to
00:30:23.27 stop polarity patch wandering with my fusion protein. And I
00:30:26.28 predict that if polarity patch wandering is important, then by
00:30:30.20 stopping the wandering, I would expect the cells would be
00:30:34.00 bad at doing orientation. They'll pick a direction and go in it,
00:30:37.01 even if it's not the right direction. But if polarity patch wandering is
00:30:42.09 inconsequential, if it has nothing to do with orientation, then it my
00:30:46.25 cells with the fusion protein where the polarity patch doesn't wander
00:30:49.27 at all, they'll be fine. They'll still be able to go up gradient.
00:30:53.01 And be able to find a mate. So before I show you the data of
00:30:56.26 cells with the fusion protein, let me remind you what wild type
00:31:00.06 cells look like in a gradient. So this is a movie I showed you
00:31:02.13 in the beginning, remember there's pheromone being applied from
00:31:05.00 the top. So all of the cells think there's the big sexy alpha cell
00:31:08.22 at the top of the screen. And they're all really good at orienting
00:31:12.12 their growth up gradient. And the cells that don't are able to
00:31:15.24 reorient. But when I looked at the cells that had the fusion
00:31:19.22 protein, Bem1-Snc2, so these cells don't have any polarity
00:31:22.24 patch wandering. So they're again treated with a gradient
00:31:25.24 of pheromone in the special slide, the microfluidics device.
00:31:28.12 What I saw is that there are a lot of cells that go in the wrong
00:31:31.20 direction. And when they go in the wrong direction, they keep
00:31:34.20 going in the wrong direction, they don't reorient. And that
00:31:38.15 led us to believe -- to conclude that the polarity patch wandering
00:31:42.15 actually is important for the cells being able to orient in a gradient
00:31:46.25 and find a mate. So in summary, what I've told you is that
00:31:50.04 polarity patch wandering is necessary for gradient tracking.
00:31:52.28 I think that it's an example of a biased random walk on the
00:31:56.20 inside of a cell, as opposed to the whole cell doing the biased
00:31:59.27 random walk. The other thing, which we weren't expecting to find
00:32:03.00 is that the polarity patch is wandering around because of vesicle
00:32:07.09 trafficking. And that actually is really important, because as a
00:32:10.25 basic cell biologist, I set out to answer one question, how do
00:32:14.15 yeast do gradient tracking? And as a byproduct of what I was
00:32:18.24 studying, I discovered this phenomenon where vesicles
00:32:22.24 cause the polarity patch to be disrupted and cause it to wander
00:32:26.14 around. And that actually has implications for human health
00:32:29.24 possibly. So this is a picture of Candida albicans, which is the
00:32:35.14 type of fungus that causes a yeast infection. And when Candida
00:32:39.21 albicans is in the human host, it makes these long projections
00:32:44.25 that are called hyphae. And they look a little bit like yeast cells
00:32:49.10 trying to mate. And in fact, we know that they grow in a very
00:32:53.00 similar way. They grow just from the tip of the hyphae, just like
00:32:56.27 yeast cells grow just from the tip of the mating projection.
00:32:59.29 And they use a similar set of proteins, polarity proteins.
00:33:04.17 And in fact, they're adding vesicles at the polarity site just
00:33:08.06 like mating yeast cells do. And so it would seem like they might have
00:33:11.14 a problem, because shouldn't the adding of the vesicles cause
00:33:14.27 the polarity patch to wander around? We don't think that these
00:33:17.28 cells are doing gradient tracking when they're growing, they're just
00:33:20.22 growing out. So isn't it a problem that the vesicles should be pushing
00:33:25.22 around the polarity patch? And it doesn't seem to be a problem.
00:33:28.20 So if we were able to understand how these cells are able to
00:33:32.28 overcome the problem of vesicle delivery not pushing the
00:33:36.19 polarity site around, then possibly you can imagine a scenario
00:33:40.09 in which we could harness that information to prevent these
00:33:43.03 cells from being able to make hyphae when they're in humans.
00:33:45.20 And therefore be able to treat some diseases.
00:33:48.02 It's fairly speculative, but you can see how a basic science
00:33:51.29 question led us to some interesting results that have implications
00:33:55.11 in fields beyond what we were anticipating.
00:33:59.19 So with that, I need to thank some important people.
00:34:01.25 First, my PhD advisor, Danny, who was incredibly supportive
00:34:05.09 throughout my PhD. Some other important people that helped me
00:34:08.15 with this project, Trevin and Natasha and Allie, all worked with me
00:34:11.23 in Danny's lab, various parts of the project.
00:34:15.03 I also need to thank our collaborators at UNC Chapel Hill,
00:34:17.28 who did some important computer modeling and provided
00:34:21.05 the specialized microfluidics device for us. And finally,
00:34:24.24 I want to thank Sam and Gao, who helped with the microscopy
00:34:27.26 that I did. And you'll notice that I did a lot of microscopy,
00:34:30.06 so they were very helpful in that regard, as well.
00:34:32.20 And with that, I thank you for your attention.

This Talk

Speaker: Jayme Dyer

Audience:

General Public
Educators of H. School / Intro Undergrad
Student
Educators of Adv. Undergrad / Grad
Researcher

Recorded: May 2015

Talk Overview

Jayme Dyer explains that some cells have the ability to detect and “drive” toward other cells. Sperm, for example, can swim to an egg during fertilization. How can it do this without eyes? The answer is by following a gradient of chemicals emitted by the egg, from lowest to highest concentration, through a process known as “gradient tracking.” Dyer points out that most chemical gradients found in biology are shallow, making it hard for cells to detect a difference in concentration. So, how then are cells able to overcome this challenge? Dyer decided to tackle this question by studying how cells detect and respond to a physiological (shallow) gradient in yeast. Even though a yeast cell can’t move, it uses gradient tracking to sense and grow toward another yeast cell during mating so that the two cells can fuse. A yeast cell has a patch of proteins (called a “polarity patch”) that lives right on the inside of the cell membrane. This polarity patch is very dynamic and moves around to continuously sense the direction of most concentrated chemical. Dyer characterized the dynamics of the polarity patch and uncovered the mechanism that causes it to move.

This talk is part of the Young Scientist Seminars, a video series produced that features young scientists giving talks about their research and discoveries.

Speaker Bio

Jayme Dyer

Jayme Dyer is a postdoc in Michael Laub’s lab at Massachusetts Institute of Technology (MIT), where she is currently studying a cell cycle checkpoint in the model bacterium, Caulobacter crescentus. She completed her PhD in Daniel Lew’s laboratory at Duke University, where she studied polarity and gradient tracking in the yeast Saccharomyces cerevisiae. Her scientific… Continue Reading