Cell Motility and the Cytoskeleton
Transcript of Part 2: Mechanics and Dynamics of Rapid Cell Motility
00:07.2 Hello. I'm Julie Theriot. 00:10.0 I'm a professor at Stanford University. 00:12.2 And for the second part of my 00:14.0 iBioSeminars presentation today, 00:15.2 I'd like to delve into the details 00:17.1 of the mechanics and dynamics 00:19.1 of rapid cell motility. 00:21.0 Now, to get us started here, 00:23.0 we see two images of a particular kind of 00:26.0 very rapidly moving cell 00:27.1 that comes from the skin of fish, called a keratocyte. 00:29.3 Over on the left, we have a fixed cell 00:31.2 that's been labeled for filamentous actin, 00:33.2 so you can see the distribution 00:35.1 of the actin cytoskeleton throughout the cell. 00:38.1 On the right, we have a video image 00:40.3 of the same kind of cell moving 00:42.1 in the same direction 00:44.1 that the fixed cell was at the time that it was 00:46.1 frozen with formaldehyde. 00:47.3 And what I'd like to talk about particularly 00:49.2 is the way that all of the different molecular machines 00:51.3 that have to operate in the context 00:53.2 of a moving cell 00:55.0 are able to coordinate with one another 00:57.0 in order to get this incredibly smooth and elegant 00:59.1 gliding motion 01:00.2 that you see in rapidly moving cells. 01:03.1 So, focusing a little bit on the details 01:06.0 of these molecular machines... 01:07.2 these are things that have actually been intensively studied 01:10.0 by biochemists and cell biologists 01:11.1 for many decades, 01:12.2 and at this point we're pretty familiar 01:14.3 with many of the proteins that are necessary 01:17.1 and, to some extent even maybe sufficient, 01:19.1 for generating the forces and dynamics 01:21.3 that we see associated with cell motility. 01:24.1 And in general, for large-scale cell biological processes 01:26.2 like motility, 01:28.0 the proteins within the cell 01:29.2 that are responsible for that kind of behavior 01:31.2 arrange themselves into nanomachines 01:34.3 where a number of different proteins will work together. 01:37.2 Now, one nanomachine that's very important 01:39.2 in cell motility 01:41.0 is the assembly of branched actin structures 01:42.2 at the leading edge of a motile cell, 01:45.1 and in the first part of this presentation 01:48.0 I talked in some detail 01:50.2 about how these different components have been identified 01:52.2 and how they're thought to work together. 01:54.2 But this particular little nanomachine 01:56.1 of growing actin filaments 01:57.3 pushing against the membrane r 01:59.2 eally only operates 02:01.1 in this very front part of the cell, 02:02.3 just a few microns back from the plasma membrane 02:04.2 at the very leading edge. 02:07.1 Another nanomachine that's also been quite intensively studied 02:09.2 and is very important for motility 02:11.1 is shown here. 02:12.2 This is the adhesion that actually binds the cell to the substrate, 02:16.1 enabling it to generate traction 02:18.2 so that it can move itself forward, 02:20.1 and here too a lot of the protein components have been identified, 02:23.2 and in this particularly beautiful example 02:25.1 from Clare Waterman's lab 02:27.0 even the spatial orientation of all the different components 02:29.1 have been very carefully measured 02:31.0 with respect to one another. 02:33.0 But this particular nanomachine 02:34.2 that makes this well-organized adhesion 02:36.2 really operates only in, again, 02:38.2 a very small zone right at the back of the cell. 02:41.0 And in order to understand 02:42.2 the overall process of cell motility... 02:44.1 when the cell is moving forward, 02:46.0 it's incredible how well all of these things 02:48.1 seem to be coordinated with one another. 02:50.0 The cell looks like it's 02:52.0 gliding over the substrate without changing shape, 02:53.3 even though it's got to be assembling actin filaments like mad 02:57.1 at the leading edge in order to push that membrane forward, 02:59.1 and then it's got to be building and then disassembling 03:01.2 the adhesions at the rear. 03:03.0 All of these things have to be coordinated 03:04.3 to happen at the same pace, 03:06.2 so that the left side of the cell 03:08.1 moves at the same rate as the right side of the cell 03:10.0 so it can go straight, 03:11.2 so that the front extends 03:13.2 at exactly the same rate that the back retracts 03:15.0 so that it can appear to move forward 03:16.2 without changing its size. 03:18.2 So it's a really fundamental and interesting question, I think, 03:21.3 about how all of these different nanomachines 03:23.1 can coordinate with one another 03:25.2 over the entire span of the cell, 03:27.0 which is tens of microns across, 03:29.0 so many of orders of magnitude larger 03:31.2 than, certainly, the individual proteins 03:33.0 that make up the machine, 03:34.1 but even any of these individual assemblies on their own. 03:37.3 Now, in order to address these kinds of questions 03:40.2 about how you get large-scale coordination 03:41.3 in moving cells, 03:43.0 we've taken a lot of advantage 03:44.2 of these fish skin cells, 03:46.0 so I'd just like to give you a bit of background 03:47.1 about where they come from 03:48.2 and how we isolate them. 03:50.1 It turns out that pretty much all fish 03:52.1 and many other amphibians 03:53.2 have a bilayered epidermis, 03:55.2 and the basal layer of that epidermis 03:58.1 is made up of these cells 04:00.1 that seem to be specialized 04:01.2 for very rapid wound healing. 04:03.2 So in the context of a fish, 04:05.0 when there are scales 04:06.3 coming out of the flank of the fish, 04:08.2 the epidermis actually wraps around the scale, 04:12.0 so if you go in with a pair of blunt forceps 04:13.3 and pluck a scale off of a fish 04:16.1 and then put it in culture, 04:17.2 the scale comes along with just a little bit of skin 04:20.1 right along the edge. 04:21.2 Now, the fish is not happy about this, 04:23.2 but it can grow a new scale, 04:24.3 and actually turning over the scales 04:26.1 is part of its normal skin regenerative process, 04:29.2 so it doesn't cause any significant damage. 04:32.1 But in the meantime, in culture, 04:33.2 now we have this scale 04:35.0 with a little bit of skin that's just wrapped around the tip of it, 04:36.2 and the cells at the edge 04:38.3 of that little bit of tissue 04:40.1 essentially think that the surface of the fish 04:42.1 has been wounded, 04:43.3 and so their response 04:45.1 is to try to start crawling outward 04:47.0 to close that gap, 04:48.2 and so you can see here, 04:50.0 both at the bottom and at the top of this particular scale, 04:53.0 these big clumps of cells 04:54.2 that start coming off first as epithelial sheets 04:56.3 and then eventually break up to make 04:59.0 all these little individual cells that seem to go 05:01.1 buzzing around more or less on their own. 05:03.1 Now, here we can look at that same process 05:04.3 at higher magnification, 05:06.2 here at the edge of a sheet 05:08.0 as if first starts coming off, 05:09.2 and then here, again, 05:11.0 this is what the isolated cells look like 05:12.2 once they break away from that epithelium. 05:14.2 And I hope you can appreciate 05:16.0 from looking at these movies 05:17.2 why these cells are such a spectacular model system 05:19.1 for studying the mechanics of cell motility. 05:21.2 They move extremely fast. 05:23.1 They are among the fastest animal cells that are known. 05:26.1 And they also have this very characteristic, 05:28.1 stereotyped geometry, 05:30.0 which is very well illustrated by the cell over here. 05:32.1 It has a very large, flat, 05:35.1 broad lamellipodium, which is its motile organ, 05:37.2 and then it carries all the rest of its organelles, 05:40.0 its nucleus, its Golgi apparatus, 05:41.2 even all its microtubules, 05:43.1 it carries in this little package of a cell body 05:45.3 that it just keeps right behind it. 05:48.2 The movement is very fast, 05:50.0 the movement is very persistent, 05:51.2 the movement is essentially unidirectional 05:53.1 - that is, they don't really have a very strong tendency 05:55.0 to change. 05:56.1 So they're essentially moving at steady state. 05:59.1 The fact that the movement is so regular 06:01.0 and is so stereotyped 06:02.2 makes it very good for biophysical analyses 06:04.1 of the kind that I'm going to be delving into today. 06:08.2 This shows you a little bit more detail 06:10.2 about how the cytoskeleton is organized in these particular cells, 06:13.1 and in this beautiful 06:15.1 structured illumination micrograph 06:16.3 taken by my student Sunny Lou, 06:18.2 you can see the distribution of actin filaments, shown in green, 06:21.1 myosin-II filaments, that is, contractile myosin, shown in red, 06:25.2 and we're going to come back to the role of myosin 06:27.1 in this coordination quite a lot. 06:29.2 And then, also, you can see labeled in blue 06:31.2 the focal adhesions 06:33.1 that are adhering the cell to its substrate. 06:35.2 And diagrammatically, 06:37.1 shown down here looking at the cell from the top, 06:39.3 you can see the actin filament branched network 06:42.1 is primarily oriented towards the front of the cell 06:45.2 and then in the back of the cell, 06:47.1 down here at the bottom, 06:48.2 you can see the filaments have become rearranged 06:50.1 to form these parallel bundles 06:51.3 that are being organized by the myosin. 06:54.1 Now, if we take a cross-section 06:55.3 through one of these cells and look at it sideways, 06:58.1 it looks like a baseball cap, 06:59.2 where the lamellipodium is very, very thin and flat, 07:02.2 only about 200 nanometers from top to bottom, 07:05.0 and the the cell body can rise up several microns high. 07:11.0 As these cells move forward, 07:12.1 they follow the same general steps 07:14.1 of actin-based cell motility 07:15.2 that are shared by many other motile animal cells 07:18.1 and also a large number of 07:20.0 eukaryotic unicellular organisms, 07:21.2 such as amoeba. 07:23.1 Overall, the first thing that has to happen 07:25.2 is the cell has to establish polarity, 07:27.2 that is, it has to distinguish its front from its back. 07:30.0 And then it has to be able extend the leading edge, 07:32.1 and in this cell, 07:33.2 as in many other motile cells, 07:35.1 the force for that extension 07:37.0 is thought to be driven by actin polymerization itself. 07:40.1 As it's extending the new leading edge, 07:42.0 it needs to form new adhesions to its substrate, 07:45.1 and at the same time be able to contract its rear, 07:49.0 to bring the cell body forward, 07:51.2 and then retract and disassemble 07:53.2 the adhesions that are at the back. 07:57.3 One of the fun things about keratocytes 08:00.2 is it's actually possible to demonstrate in these cells 08:02.2 that all of the components necessary 08:04.2 for that whole cycle of motility 08:06.1 are contained only in the lamellipodium 08:09.0 - you don't actually need any contributions from the cell body. 08:11.2 And that was first proved in this really classic 08:14.1 1984 experiment by Ursula Euteneuer and Manfred Schliwa, 08:18.0 where they sliced little bits 08:20.2 off of the lamellipodium of a keratocyte, 08:22.2 leaving the cell body behind, 08:24.1 and were able to see that those small fragments 08:26.1 of the keratocyte lamellipodium 08:27.3 were able to continue to translocate on their own, 08:30.2 and move at just about the same speed 08:32.1 and just about as persistently 08:34.2 as the whole cell was when it was intact. 08:38.0 And this movie shows a modern reenactment of that experiment 08:40.2 that was done by my student Erin Barnhart. 08:42.2 Here you see a fragment that's been 08:45.0 isolated away from its cell body 08:46.2 that's nothing but lamellipodium, 08:48.3 with all of these dynamic cytoskeletal structures inside of it. 08:51.1 And when the movie plays you can see 08:53.2 it's crawling along very nicely, 08:54.2 it's got a clear front and a clear rear. 08:56.2 It's about to crawl over a little piece of schmutz on the coverslip 08:59.1 that actually is going to separate 09:02.0 this crawling lamellipodial fragment into two bits. 09:05.1 The membrane connection between them resolves 09:07.0 and they're both able to crawl off on their own, 09:09.1 until they eventually get sliced into some unit 09:11.2 that's too small to movie. 09:13.3 So, with this system 09:15.2 we have favorable geometry 09:17.0 -- it's very, very simple, very reproducible from one cell to another -- 09:19.2 and we also have a fairly simple self-contained system 09:22.0 where we know that it's only the 09:24.2 components of the lamellipodium 09:26.1 that are necessary for persisten motility. 09:28.2 So, from analyzing the behavior of these cells 09:31.2 over many years, 09:33.0 my group has been able to identify and specifically 09:35.1 measure the contributions of all of the 09:38.1 different force-generating elements 09:40.0 that help the cell to move, 09:41.2 and those are all illustrated here. 09:43.3 At the leading edge, we have actin polymerization, 09:46.1 that's shown in red, 09:47.2 which pushes the membrane outward, 09:49.1 and that polymerization is actually opposed 09:51.2 by tension in the plane of the membrane. 09:54.1 And that tension serves 09:55.3 both to act as the barrier 09:58.0 that the growth of the actin filament pushes against, 10:00.2 and also, in fact, 10:02.0 helps to coordinate the motion 10:04.2 over the entire surface of the cell, 10:06.1 as we'll see in a little bit more detail. 10:08.1 Now, there's also adhesions that have to contribute, 10:12.0 and those adhesions are assembled in the front 10:14.0 and then disassembled in the back. 10:16.1 And then there's contractile forces that are driven by myosin, 10:18.2 primarily acting at the back. 10:21.0 Because that myosin contraction is happening 10:23.1 at the back and squeezing the cytoskeleton inward, 10:25.2 that actually creates a forward 10:27.3 hydrodynamic fluid flow 10:29.1 that squirts fluid through the meshwork of the lamellipodium 10:32.3 to deliver components up to the front of the leading edge. 10:36.1 Now, as I said, we've been able to measure 10:38.2 the quantitative contributions of each of these different forces 10:40.3 within the context of this very simple kind of motile cell, 10:44.0 and what I'd like to do over the next few minutes 10:45.2 is share with you a couple of highlights of things 10:48.1 that we've learned 10:49.2 that are somewhat surprising in retrospect 10:51.1 as to how this coordination is able to work 10:53.1 over such a large scale. 10:56.0 So, in order to make these kind of measurements, 10:57.2 we've had to develop methods 10:59.3 both for measuring behaviors of the cells very precisely, 11:02.3 and also methods for perturbing the behaviors of the cells 11:05.2 so that we could understand what aspects 11:07.3 were dependent on what other aspects. 11:09.3 So, one example of a kind of measurement that we can make, 11:12.3 shown here in a movie from Cyrus Wilson, 11:16.0 is tracking of the overall motion of the actin network 11:20.0 using a technique called speckle microscopy 11:22.2 that was originally developed by Clare Waterman. 11:24.2 And in this technique, 11:26.1 the cells are electroporated with a small amount 11:28.2 of a fluorescent dye that binds to the actin filaments, 11:31.0 but a sufficiently small amount that 11:33.0 rather than labeling the whole cell uniformly 11:35.1 you instead see this little speckly, textured pattern. 11:38.2 And then if we take the movies 11:40.0 as the come off the microscope, 11:41.2 which is what you see up top, 11:43.2 and then convert them into a different frame of reference, 11:47.1 where instead of looking at the cell in the lab frame of reference, 11:50.0 we now translate everything 11:52.2 as if the cell had a GoPro camera attached to its head, 11:55.2 and we're looking just right down at the cell itself 11:58.1 from its own point of view. 12:00.0 Then you can see, now, quite a bit more detail in terms of... 12:03.2 both in phase contrast and then with this fluorescent speckle microscopy, 12:06.1 how everything inside the cell is moving. 12:08.1 So, looking at the fluorescent speckles, 12:10.0 I hope you can now appreciate 12:11.2 that the whole actin network is sort of raining downwards 12:14.1 from the front towards the back of the cell 12:16.1 in the cell's frame of reference, 12:18.1 and is also being gathered inwards on the side, 12:20.1 down at the back here, 12:21.2 where the myosin is able to contract it. 12:24.1 Now, once we can do that frame of reference shift 12:26.3 and look at things from the cell's point of view, 12:28.3 Cyrus Wilson was able to work together with 12:31.3 Gaudenz Danuser and people in his lab, 12:34.0 including Lin Ji, 12:35.2 to develop quantitative methods for 12:37.3 measuring the flow of all of this material 12:39.3 in the lamellipodium very precisely, 12:41.2 and was able to map, overall, 12:44.0 how the motion of these particles 12:45.2 depend on the location inside of the cell. 12:47.2 So, here from the lab frame of reference, 12:49.2 what you can see is that the motion of the particles 12:51.2 with respect to the substrate is actually very little, 12:54.3 that is, the actin is actually pretty stationary 12:56.2 with respect to the glass that the cell is crawling over, 12:59.1 except at the very back where you see this massive 13:02.0 inward sweeping driven by myosin. 13:03.3 However, if you look from the cell's point of view, 13:06.0 you see there's a low of flux, 13:08.0 a lot of turnover of constantly treadmilling actin network, 13:11.2 where it's assembling at the front 13:13.1 and then disassembling under the cell body. 13:17.0 So, in order to try to understand this process 13:18.3 of assembly and disassembly a little bit better, 13:21.0 we also wanted to be able to 13:22.2 manipulate the behaviors of the cells, 13:24.1 to perturb them so that we could look at 13:26.1 how they responded to changes in their environments. 13:28.2 And one kind of perturbation 13:30.1 that was actually very informative 13:32.0 for understanding how these things couple together 13:33.2 was worked out by Erin Barnhart, 13:35.2 specifically where she was able to 13:37.2 change the degree of adhesivity, 13:39.2 or the degree of stickiness, 13:41.1 of the substrate that the cells were crawling over. 13:43.2 And what she found is that when cells 13:45.3 are on a sort of moderately sticky substrate, 13:47.1 they're able to move exactly the same way 13:49.1 that they would on glass 13:51.1 or in fact on the surface of an aminal. 13:53.2 When they're put on substrates that were less sticky, 13:55.2 so, ones that were more slippery, 13:57.3 you can see the cells actually change shape 13:59.2 #NAME? 14:01.0 and you can see the accumulation of these characteristic 14:03.1 pleats in their lamellipodium, 14:04.3 where the inward flow of the actin 14:06.2 is now actually faster than the motion of the cell, 14:09.1 so it's really spinning its wheels 14:11.1 because it can't quite get a grip on its surface. 14:14.1 Now, most interestingly, I think, 14:16.1 when they're put on high adhesion substrates, 14:18.1 their behavior changes very dramatically, 14:20.1 and instead of now having this steady state motion 14:23.0 where they glide forward uniformly, 14:24.3 they now do this completely crazy thing 14:26.1 of putting out small bits of lamellipodium 14:28.0 that seem to sweep sideways. 14:30.3 And we're in the process of trying to figure out 14:33.0 how all of these different things work, 14:34.1 but I hope you can appreciate that even this 14:36.3 very, very simple motile cell 14:39.2 that seems like, you know, sort of the stripped down, 14:41.0 minimalized, like, soapbox derby version of a motile cell, 14:44.0 even this is able to extremely 14:46.1 expand its behaviors depending on cues 14:48.3 that it's getting from the environment, 14:50.1 in this particular case, 14:51.2 mechanical cues in the form of the stickiness of the substrate. 14:56.3 So, the several examples 14:59.0 that I want to tell you about 15:00.3 mostly have to do with surprising roles for myosin. 15:03.2 Now, myosin, of course, 15:05.0 we're mostly familiar with in the context of skeletal muscle, 15:07.2 where it's able to contract sarcomeres 15:09.3 by sliding stable arrays of actin filaments 15:12.1 relative to one another. 15:14.2 Myosin in non-muscle cells, 15:16.0 myosin II in non-muscle cells, 15:17.2 also acts as a contractile protein, 15:20.1 and its assembly is regulated, 15:22.1 so the monomeric state of the myosin 15:25.0 in non-muscle cells 15:26.2 is folded up on itself, 15:28.1 and then when it receives an appropriate signal, 15:30.2 the phosphorylation of the 15:34.1 regulatory light chains on myosin 15:35.3 enables it to extend outwards 15:37.3 so that it can then assemble into bipolar thick filaments 15:39.2 that are much more similar to the organization inside of muscle. 15:42.2 And so we look at the myosin in keratocytes... 15:45.1 what you can see is there's very little myosin at the front, 15:48.1 where the actin is actively polymerizing, 15:50.2 and instead there's actually quite a lot of myosin 15:52.1 right at the back, 15:53.2 and in particular it's in these very bright spots 15:55.2 right on either side of the cell body. 15:59.0 Okay, so bearing all that in mind, 16:00.2 now let's go back to this question of assembly and disassembly. 16:03.2 One of the things we're able, now, to measure, 16:06.0 that we can track the motion of the actin 16:07.2 and know where all these other elements are located, 16:09.3 is Cyrus was able to actually figure out 16:12.1 how to make a map of, 16:14.0 quantitatively, 16:15.2 how much assembly and disassembly of the actin cytoskeleton 16:17.2 takes place over the context of the whole cell. 16:20.1 And what he found was that the assembly 16:22.1 is very much biased towards the leading edge, 16:24.1 specifically right in the middle of the front of the leading edge, 16:26.3 which is very much what we'd expect, 16:28.3 but the disassembly, unexpectedly, 16:31.0 was found in these two very intense spots 16:33.3 right on either side of the cell body. 16:36.0 And Cyrus recognized that 16:38.2 those locations were actually 16:40.1 very similar to the locations where we found myosin. 16:43.1 Now, we can also look at the distribution 16:45.0 of mysoin II in these cells. 16:46.2 Here, using a cell that's been transfected 16:48.2 with myosin light chain carrying YFP. 16:51.1 And now, in the cell frame of reference, 16:53.1 you can actually see the motion of these little speckles, 16:55.3 which now are mini-filaments of myosin, 16:58.3 that is, bipolar filaments that have been assembled. 17:01.2 And what you can see is they seem to 17:04.0 stick onto the actin network 17:05.2 and then they rain backwards 17:07.1 towards the back of the cell body, 17:09.1 essentially riding on the actin network 17:11.1 until you get right all the way to the back, 17:13.1 where they then start forming these contractile cables, 17:15.2 pulling in the actin network 17:17.0 and making these bundles that go from one side to the other. 17:20.1 Now, it's suggestive that the spatial distribution of myosin II in these cells 17:26.3 is exactly the same as the foci of disassembly, 17:30.0 and we can also inhibit disassembly of actin in these cells, 17:34.1 for example, by inhibiting the motor activity of myosin. 17:36.2 So, we hypothesized that 17:38.3 the myosin itself is actually contributing 17:40.2 to the disassembly of the actin cytoskeleton 17:42.3 by buckling and breaking and ripping apart the actin filaments 17:46.3 using, directly, its force-generating capabilities. 17:49.1 And one of the strongest pieces of evidence 17:50.3 in favor of that hypothesis 17:52.2 is this very nice experiment done by Mark Tsuchida, 17:55.1 where instead of using moving living cells, 17:58.1 he used extracted cytoskeletons. 18:00.2 So, if you take a keratocyte 18:02.3 as it's crawling across a substrate 18:04.1 and then sneak up on it with a little bit of detergent, 18:06.2 you can get the membrane to dissociate, 18:08.2 leaving behind only the insoluble parts of the cytoskeleton, 18:12.0 so, the assembled actin filaments, 18:13.2 whatever actin binding proteins are bound to them, 18:17.0 but having now gotten rid of all soluble components, 18:19.1 including actin monomers, ATP, everything else. 18:24.2 Mark was then able to label those extracted cytoskeletons with phalloidin 18:28.1 to see where the actin filaments were, 18:30.1 and then add back ATP 18:32.1 to those extracted cytoskeletons. 18:34.0 That added ATP was able to then 18:36.2 activate the myosins that were left behind, 18:38.2 so he could see if, 18:40.0 in this sort-of semi in vitro environment, 18:42.1 myosin activity could actually 18:44.2 drive destruction of the actin filament network. 18:47.1 And so that's what you're going to see in this movie 18:49.0 #NAME? 18:51.2 When the movie starts to play, 18:53.2 the ATP is going to be added, and you can see the network 18:55.3 just melted right in the back, 18:57.2 right where the myosin is located. 18:59.1 And you can also see that by comparing 19:00.2 this before and after shot, 19:02.0 where the blue shows the places where 19:04.0 the actin network disappeared 19:05.2 when the myosin was activated. 19:07.1 So, although we normally think of myosin 19:09.0 as actually contributing to contraction, 19:11.1 in this context, at least, 19:12.3 it seems like one of its more important functions 19:14.1 is destroying the actin network 19:16.1 when it gets to the back of the cell. 19:19.1 So, putting that together, 19:21.1 we came up with this idea for 19:23.2 myosin driving overall network treadmilling in the lamellipodium, 19:27.1 as illustrated here, 19:28.3 where initially, towards the front, 19:30.3 there's very little myosin in the network, 19:33.1 it's hard for the myosin mini-filaments 19:34.2 to diffuse through the network, 19:36.1 as it's actively assembling and 19:38.0 essentially pushing everything backwards, 19:39.2 but a few fo them get ahold of the filaments, 19:41.0 and then as they start contracting 19:42.2 they start rearranging the actin filaments 19:44.1 to form more parallel structures 19:45.3 that are of more favorable geometry for force generation by myosin. 19:49.2 After that goes on for a while, 19:51.0 by the time you get to the back of the cell, 19:52.2 the actin is now all in parallel bundles 19:55.0 rather than a dendritic network, 19:56.2 and the high concentration of myosin 19:58.1 that's able to accumulate there over time 20:00.1 is enough to rip that network apart. 20:02.3 So, overall, we think that's a major mechanism 20:05.1 for determining what the distance is 20:07.1 from the front of the cell to the back of the cell. 20:08.2 It's just determined by how much time it takes 20:11.0 for myosin to incorporate, 20:12.2 and for myosin to destroy the network. 20:16.0 Now, so far, I've been talking about keratocytes 20:19.1 as if they're all exactly identical, 20:20.3 and certainly that's one of the useful things about them 20:23.0 is that they're similar, but, 20:24.2 like any other organism, 20:25.3 if you look at the closely enough, 20:27.1 you'll see they actually have very interesting differences 20:29.0 from one another. 20:30.1 So, this shows a gallery of a whole bunch of different keratocytes 20:33.1 that were collected by Kinneret Keren and Zach Pincus, 20:36.3 showing that from even one scale 20:39.3 of a particular individual fish 20:41.1 you can have quite a lot of variation, 20:42.2 both in terms of the size of the individual cells, 20:45.0 and then also their shape. 20:46.1 So, some of them are quite round 20:47.3 and some of them are quite elongated 20:49.2 and almost canoe-shaped. 20:51.1 And to summarize a lot of work, 20:53.0 what we've been able to find is that 20:55.2 these cell-to-cell shape differences 20:57.1 are both persistent 20:59.0 -- so, if you follow a cell over time, 21:00.1 it keeps its shape -- 21:02.0 and they're also cell-intrisic 21:04.0 #NAME? 21:05.3 and let it grow back, 21:07.2 it will grow back to exactly the same shape it was before. 21:10.3 And from quantitative analysis of those kinds of measurements, 21:13.2 what we found is that these shape differences 21:15.1 are essentially extremes of a continuous spectrum, 21:18.2 where some cells are very large and wide and smooth, 21:22.0 and these are the ones that are canoe-shaped, 21:24.0 and those are also the fastest moving cells. 21:26.2 And some of the other cells, 21:28.0 such as the ones over on the left side of this gallery, 21:30.1 and rounder, they tend to be smaller, 21:33.2 their leading edges look kind of rough, 21:35.2 and they also move kind of slow 21:37.1 and in a less persistent manner. 21:39.0 And so we call the wide, smooth cells, 21:41.2 we call those coherent cells, 21:43.1 and the smaller, narrower, rough cells, 21:45.1 we call decoherent cells. 21:47.1 But overall, we can find every behavior in between, 21:50.3 so we think the differences that we see among these shapes of the cells 21:53.3 basically just has to do with the 21:56.1 exact quantitative amount of all of these 21:58.3 force-generating elements they have present 22:01.1 within their cytoplasm 22:02.3 that balance each other in slightly different ways 22:04.2 to give overall cell shapes. 22:07.3 And overall, we can 22:09.3 quantitatively measure these variations in cell shape, 22:11.2 particularly identifying principle modes of shape variation, 22:14.2 and the mode I've been most frequently referring to 22:18.1 is this second mode, 22:19.2 where we go from the wide cells to 22:22.2 the rounder, more D-shaped cells. 22:24.2 And our modeling that we've done together with Alex Mogilner, 22:28.0 together with experimental work, 22:29.1 has suggested that really the variation in those shapes 22:31.3 is primarily due to 22:34.1 the back-and-forth force balance 22:35.2 between actin polymerization pushing on the membrane 22:37.1 and membrane tension restraining the actin polymerzation. 22:41.2 And the short version is that 22:44.1 cells that have very forceful actin polymerization 22:46.2 are able to assume this coherent, wide lamellipodium, 22:50.0 and cells that have weaker actin polymerization, 22:51.3 for whatever reason, 22:53.2 are the ones that end up in the rounder D-shape, 22:55.2 and also move slower. 22:57.2 Now, if that idea is true, 22:59.2 then it should be the case that we can 23:01.2 take an individual cell 23:03.2 and somehow increase or decrease 23:05.2 its overall rate of actin polymerization, 23:07.1 and have that individual cell 23:09.1 change across this entire shape spectrum. 23:12.2 And so that experiment actually was done by Greg Allen, 23:15.1 where the method he chose to change the rate of actin polymerization in the cell 23:19.3 was simply to lower and raise the temperature. 23:22.2 So, here we're going to watch a movie 23:25.1 of a cell and as it moves along 23:27.1 you can see it's fairly slow, 23:28.2 it's got this more sort-of D-shaped pattern, 23:31.1 and Greg is first going to start 23:33.3 dropping the temperature, 23:36.1 and as the temperature drops 23:38.1 you can see the lamellipodium gets rounder and rounder, 23:40.3 and the cell is moving slower and slower. 23:45.0 And at this point, 23:48.1 when we get down to just about 7 degrees Celsius, 23:50.0 he's now going to start raising the temperature, 23:53.1 and as the temperature comes back up 23:55.1 you can see the cell not only goes faster and faster, 23:57.2 but it also assumes a wider shape. 24:01.1 And so following cells like this quantitatively 24:03.1 using a variety of different metrics, 24:04.3 what we were able to find is that, in fact, 24:06.2 individual cells can explore 24:08.2 this entire range of behavior 24:10.0 that we see in the context of cell-to-cell variation, 24:12.0 and it's all consistent with the idea that 24:14.1 the primary determinant of the shape of the cell 24:16.2 as well as the speed of the cell 24:18.1 is simply how fast the actin is polymerizing. 24:23.1 Now, another really fun thing about keratocytes 24:25.1 is they have the ability to sense and respond to electric fields, 24:29.2 and this is something they actually have in common 24:31.2 with many other motile cells. 24:32.3 Pretty much any motile cell 24:34.3 that you put in a DC electric field 24:36.2 will choose either the anode or the cathode 24:38.2 and will head in one direction. 24:41.1 This was first described for keratocytes 24:43.1 by Cooper and Schliwa back in 1986, 24:45.1 and Greg was able to replicate this 24:47.3 using a setup that he built in our lab 24:49.3 to look at motion of individual cells 24:52.2 as the electric field was switched from one direction to the other. 24:57.0 So, in this movie, we see a cell 24:58.3 that's moving along in an electric field 25:01.1 that is oriented in this direction 25:03.1 #NAME? 25:04.2 and also the magnitude listed over here -- 25:06.2 and you can see the cell is following that line. 25:08.2 Now, when the label turned red, 25:10.2 that was when the field was flipped around, 25:12.2 and you can see the cell has turned 25:15.1 and is now heading back in the other direction. 25:17.1 And now, once again, the orientation of the field is flipped, 25:21.1 the cell flips back around, 25:22.2 and now heads back in the direction that it has been told to go. 25:26.1 So, the cells obviously... 25:28.1 even though I've been emphasizing 25:31.0 how good they are at balancing forces 25:32.3 across the front of the cell and between the front and the back of the cell, 25:35.0 they are able to also initiate imbalances in their forces 25:38.2 so that they're able to turn. 25:40.2 So, Greg Allen looked a little bit more deeply into the mechanism 25:43.1 of how they turn 25:44.2 and he found a couple of interesting things. 25:46.1 So, for example, 25:47.3 if we look at a single turning cell 25:49.3 -- in this case we're looking at it both in the lab frame of reference, 25:52.1 like it looks on the microscope, 25:53.3 and then in the cell frame of reference, 25:55.1 where we've repositioned everything 25:57.0 to see things in the cell's point of view -- 25:59.0 you can see there is a physical asymmetry 26:01.0 in a turning cell, 26:02.2 where the part that's on the inside of the curve, 26:05.0 that is, the part that's going slower, 26:06.3 has this very round shape that we call decoherent, 26:10.1 and it's characteristic of slow motion. 26:12.3 And on the outside of the turn, the part that's going faster, 26:16.0 has a much more elongated, coherent shape 26:18.2 that we associate with fast motion. 26:20.3 So, this variation that we see, 26:22.2 both at the population level 26:24.0 and in individual cells as the temperature 26:26.2 is raised and lowered, 26:28.1 can actually also happen even within the context of an individual cell, 26:31.0 where one side can end up being much faster 26:34.0 than the other side. 26:36.1 Now, there's a number of different things 26:37.3 that contribute to this asymmetry, 26:39.2 but at this point you won't be surprised 26:41.0 that one of the major things that contributes 26:42.2 is the left-right distribution of myosin. 26:45.2 So, I showed you before that the myosin 26:47.1 accumulates in these two spots on either side of the cell body, 26:50.2 and those two spots aren't always necessarily equal in size. 26:54.1 So, this is an example of a cell 26:56.2 where there's a relatively low amount of myosin 26:59.0 in the spot on the left side, 27:01.0 and a much higher amount of myosin in the spot on the right side. 27:03.2 And looking in the movie, 27:05.0 you can see the consequences of that, 27:06.2 here with the myosin labeled: 27:08.0 the side that has more myosin is moving faster 27:12.2 and is therefore sweeping around the outside part of the turn. 27:14.2 Now, there's of course other elements 27:16.2 that also contribute to this turning 27:18.1 -- there are differences in adhesion, 27:19.3 there are differences in traction force, 27:21.1 there are differences in rates of actin polymerization -- 27:23.1 but they all seem connected to one another, 27:25.1 and specifically connected to one another 27:27.2 through the mechanism of myosin action. 27:30.0 So, to summarize what we think is going on here, 27:32.2 we think as the cell begins to turn 27:35.3 the actin network flow starts to flow... 27:37.2 instead of just flowing straight back to the back of the cell, 27:40.0 starts to flow at a slight angle. 27:41.3 Because the myosin is carried along 27:44.1 on that flowing actin network, 27:45.2 the myosin then accumulates 27:47.2 at the outside corner of the cell. 27:50.1 That myosin is able to contract faster, 27:52.1 pull in that side of the 27:55.3 back of the wing of the lamellipodium 27:57.1 and help the cell sort of flip around. 27:59.0 At the same time, because the myosin 28:01.1 is depolymerizing the actin filaments, 28:02.2 it's generating a gradient of G-actin, 28:04.3 such that there's more actin available for polymerization 28:08.0 on that same side of the cell 28:09.3 where you have more myosin 28:11.1 and where you have faster motion. 28:12.3 And all of these things, we think, 28:14.2 are able to actually feed back on one another 28:17.0 in a positive sense, 28:18.1 such that once a cell starts making one of these turns 28:20.3 it actually is able to continue to make that turn 28:22.3 in a persistent way 28:24.1 actually for quite a long time, 28:26.0 until it's then forced to turn in another direction. 28:30.1 So, so far what I've shown you 28:32.1 is that in the context of these very simple 28:34.2 steady-state moving cells, 28:36.1 myosin in the back of the cell 28:37.2 is actually doing a tremendous number of exciting things 28:40.0 that help the cell move overall and that help coordinate the front and the back. 28:43.3 It's helping to disassemble the network 28:45.1 and it's also specifically contributing to 28:47.2 left-right aymmetries that help the cell to turn. 28:49.3 Now, the keratocytes are fairly unusual cells 28:53.0 -- they're unusual in their appearance, 28:54.2 they're unusual in the steadiness of their motion -- 28:56.3 and so it became very natural then for us to ask 28:59.3 whether similar mechanisms might be at play 29:01.2 in more complicated cells 29:03.1 that are doing more complicated tasks. 29:05.2 And one of the very interesting cells that's been well-studied 29:09.0 in the context of motility 29:10.2 is the neutrophil, the human neutrophil, 29:12.2 a white blood cell, 29:14.1 whose job is to go after and engulf the 29:16.2 bacteria that are invading the human body. 29:19.1 And you can actually isolate neutrophils 29:21.2 from your own blood 29:23.1 and watch them crawl around and eat things 29:24.3 -- it's really very gratifying -- 29:26.1 but also we have a call line, 29:28.1 a neutrophil-like cell line, 29:30.0 that is able to behave much like a neutrophil 29:31.3 but that we can also transform 29:33.2 and look at protein distributions 29:35.2 in moving versions of the cell. 29:38.1 So, Tony Tsai in the lab 29:40.1 decided that it was time 29:43.0 to actually break out of the keratocyte mold 29:45.0 and start looking at motion in more complicated kinds of cells, 29:48.1 including neutrophils, 29:50.1 and just to show you how dramatic 29:52.2 the behavior of these cells is, 29:53.2 this is one of these HL60 cells 29:55.3 that's been put in a chamber with some Candida albicans, 29:58.1 which is a pathogenic yeast, 30:00.2 and what you see as the movie loops 30:02.0 is the neutrophil starts off over on the right side, here, 30:05.1 and then runs across 30:07.2 to this little pile of yeast 30:08.3 and is able to actually phagocytose and engulf them. 30:11.0 So it really is a very 30:13.0 neutrophil-behaving tissue culture cell. 30:16.0 Now, looking at the shapes, 30:17.3 they're obviously much more complicated than the keratocytes, 30:20.1 and putting in labels for the actin, 30:22.2 which is shown here in green, 30:23.3 the myosin, shown in red, 30:25.1 and then DAPI to stain the nucleus in blue, 30:27.2 what you can see is that the shapes are not only 30:29.1 much more variable than keratocytes 30:30.2 but also much, much more dynamic. 30:32.3 All of the cytoskeletal elements 30:34.2 are drastically rearranging themselves 30:36.2 over periods of just a few seconds 30:38.1 as the cell is crawling around. 30:40.1 So, although this makes it a more interesting question, 30:42.1 I think, 30:43.2 to figure out what is going on in terms of 30:45.2 the mechanics and dynamics of this behavior, 30:47.2 it's also a much more challenging problem 30:49.1 as far as quantitative analysis goes. 30:53.1 So, Tony so far has been able to work out 30:55.1 a number of quantitative techniques 30:57.1 to be able to break down this complex motion 30:59.1 so we can actually watch changes over time. 31:02.0 So, for example, he can track the edges 31:04.0 of one of these moving neutrophils 31:05.1 and then go back and calculate, 31:07.1 for the cell as it's moving, 31:09.1 how much of the cell is extending in every time frame, 31:12.0 in this case, every couple of seconds, 31:13.2 how much is retracting, 31:15.2 calculate the overall area of the cell 31:17.1 as well as the extension of its leading edge, 31:19.1 and the amount of retraction of its body. 31:22.3 At the same time, we can look at labeled proteins inside the cell, 31:26.1 and obviously one of the ones we're most interested in looking at 31:29.3 is myosin, 31:31.1 and look at the overall fluorescence intensity distribution 31:33.2 and see how that changes as the cell moves around. 31:36.1 And what you'll probably be able to see 31:38.1 is that the myosin localization itself is also very dynamic. 31:40.2 It's often in the back of the cell, 31:42.0 sometimes in these bright spots, 31:43.2 but then those bright spots will disassemble, 31:45.1 the myosin will become more uniform 31:47.2 or will move to a different location within the cell. 31:50.3 And tracking all those things quantitatively over time, 31:53.1 what Tony was able to find 31:55.2 was that when a cell was speeding up, 31:57.2 that the accumulation of the myosin 32:00.2 in response to that change in cell behavior 32:03.0 happens later, happens about 12 or 15 seconds 32:06.0 after the initial movement of the leading edge. 32:09.0 So, whereas in the keratocytes 32:10.2 it seems like the actin and the myosin 32:12.2 were always in perfect balance 32:14.1 so that the cells were always gliding forward, 32:16.1 in the neutrophils the story is a little bit different 32:18.1 #NAME? 32:20.1 and the myosin is then reacting. 32:22.1 So, as the cell initially extends, 32:24.3 it then activates the accumulation of myosin at the rear 32:27.2 to pull the back together. 32:29.1 So, instead of gliding, 32:30.2 it's doing more of an inchworm motion. 32:34.2 How does this affect cell turning? 32:36.1 Well, the same way that Tony was able to come up with 32:38.3 quantitative metrics for the localization of myosin, 32:41.0 he was also able to come up with 32:42.2 quantitative descriptions for the orientation change of the cell 32:46.0 and then the left-right asymmetry of the myosin 32:48.3 with respect to its immediate orientation. 32:51.0 And you can already see the answer, actually, 32:53.1 quite dramatically, 32:54.3 with this maximal intensity projection, 32:56.2 where this is just a low magnification movie 32:59.2 of a cell that's undergoing a sinusoidal path, 33:02.1 and we're looking now only at the myosin 33:04.1 that's accumulating at its rear, 33:05.2 and what you can see is that the myosin 33:07.1 always accumulates on the outside of the turn, 33:10.0 and when it changes its direction, 33:12.0 the myosin then changes which side of the cell it's on. 33:16.1 So, this is actually very reminiscent of the keratocytes, 33:18.2 where we saw, again, 33:20.1 the myosin on the outside of the turn, 33:22.0 except in the case of neutrophils, again, 33:24.1 there's a little bit of a time lag 33:27.0 between when the turning initially starts 33:30.1 versus when the myosin accumulates on the outside of the turn. 33:33.0 So, here too it looks like 33:34.3 the actin is calling the shots in terms of direction, 33:36.2 the flow direction of the actin 33:39.0 that's changing 33:40.2 then gathers the myosin on the outside of the turn, 33:43.1 that then causes disassembly of that cytoskeleton 33:46.1 in a way that enables the neutrophil 33:48.1 to essentially swing its tail around 33:50.1 so that the whole cell is now oriented in the proper direction. 33:54.1 So, overall, comparing these two stories... 33:56.2 if you look at a movie of a keratocyte versus a neutrophil, 33:59.2 they seem like they're behaving rather differently, 34:01.2 but what we understand when we dissect 34:04.1 the mechanics and the dynamics of this behavior 34:05.2 is that they're strikingly similar. 34:07.1 In particular, I've shown you 34:09.2 recent data that myosin accumulates at the cell rear 34:11.3 due to this actin network retrograde transport 34:13.3 and mediates actin network disassembly 34:16.0 in a way that is then able to coordinate 34:17.3 not only front-rear motion of the cell 34:20.1 but also give you asymmetries 34:22.1 that can lead to turning. 34:23.2 And both of those things seem to happen 34:25.2 in very similar ways 34:27.0 in both the keratocytes and the neutrophils. 34:29.2 Now, one last little hint I want to leave you with is, 34:32.2 as I showed you before with this movie, 34:34.1 we can force keratocytes 34:36.0 to behave more like neutrophils 34:38.0 in terms of changing their shape all the time 34:39.2 if we simply change the environment, 34:41.1 and particularly if we put them on 34:43.1 very, very sticky substrates. 34:44.2 So you might wonder, can we do the flipside; 34:47.2 can we make a neutrophil behave more like a keratocyte, 34:50.2 into something that will have a steady-state motion 34:52.1 where everything is happening at the same rate? 34:55.1 Well, it turns out when you take a neutrophil 34:57.1 and you confine it to a very narrow channel 35:01.0 and then watch it move over time, 35:03.0 these guys now are moving at steady-state. 35:05.3 The speed is an absolute constant, 35:07.3 the distribution of myosin is constant, 35:09.2 it's always found at the rear, 35:11.2 and they'll continue to move like this 35:13.0 for many tens of minutes 35:14.2 without any obvious changes 35:16.0 in terms of their overall shape or overall behavior. 35:18.2 And taking this same movie and now making a kymograph of it, 35:21.2 where the time is moving from top to bottom 35:23.3 and each one of these slices 35:25.2 is an individual frame of this movie, 35:27.0 you can see, really, how constant this speed is over time. 35:30.1 So, not only can we force keratocytes 35:32.2 to become more crazy 35:33.2 and change their direction like neutrophils, 35:35.1 we can also force neutrophils 35:37.1 to behave more in a steady-state like keratocytes. 35:40.0 And moving forward, I think the combination 35:42.1 of our ability to both measure and manipulate 35:44.0 these different kinds of motile cells 35:45.2 will help us to understand 35:47.3 the general principles that govern motility 35:49.3 for all animal cells 35:51.3 that use actin polymerization to drive their movement. 35:56.1 So, as an overall summary, 35:58.2 I think the main point here is that 36:00.3 actin and myosin have to cooperate 36:02.2 in order to make cells move, 36:04.1 not just in order to generate force, 36:05.2 but also just in order to do things 36:07.2 like steer them and determine their shape. 36:09.0 And we found that myosin plays actually 36:12.0 several very unexpected roles at the rear of cells, 36:13.2 not just contributing to contraction, 36:15.2 as we might have expected, 36:17.1 but also contributing specifically to actin network turnover 36:20.1 and to asymmetries that lead to cell turning. 36:22.3 And overall, we've been very surprised 36:24.2 by how similar the mechanics are 36:27.0 between fish skin keratocytes and human neutrophils. 36:33.2 So, obviously, there have been 36:35.3 a lot of very talented people 36:37.2 who have contributed to the work that I just described, 36:39.1 and I've listed here the many members of my group 36:41.2 who have contributed to different aspects of cell motility projects 36:44.2 over the last 15 years or so, 36:46.1 and also our wonderful collaborators. 36:48.2 And I'd particularly like to mention this in this context 36:51.2 our very productive long-term collaboration 36:53.0 with Alex Mogilner, 36:54.2 who has done a lot of the quantitative physical modeling 36:56.2 that has driven the thought processes behind our experiments. 37:00.0 Thank you.