Patterning Development in the Early Embryo: The Role of Bicoid
Transcript of Part 2: Stability of Morphogen Gradients & Movement of Molecules
00:00:02.15 So I'm Eric Wieschaus, I'm a HHMI investigator and a professor at Princeton University. 00:00:08.15 And in the first part of this presentation, I already talked to you about how the Drosophila 00:00:13.19 embryo develops and in particular, how spatial patterns of gene expression 00:00:19.08 and cell behavior arise, and one of the things we saw was the importance of maternal RNAs 00:00:26.24 and maternal proteins that are localized to particular regions of the egg, 00:00:31.06 and provide positional information to the cells that are found in individual positions. 00:00:38.15 What I'd like to do in this second part of the lecture is to actually focus in a little bit more 00:00:44.26 on these maternal RNAs and in particular, I'd like to talk about some work that 00:00:49.23 a graduate student, Thomas Gregor has done in the lab in collaboration 00:00:56.18 with two biophysicists at Princeton, Bill Bialek and David Tank. And I think the reason that 00:01:02.07 I wanted to talk about these experiments is that they provided a wonderful example 00:01:08.29 I think of the importance to us as biologists right now, beginning to try to view 00:01:18.00 problems more quantitatively, to try to establish actual numbers for the biological 00:01:24.18 phenomenon that we've come to understand in a vague-ish kind of cartoon way, but that 00:01:28.28 if we want to bring our knowledge to the next level. If we want to test our understanding 00:01:36.10 one of the things that we really have to be able to do is to measure and supply numbers. 00:01:41.02 And the experiments that I'm going to present, although you'll see their not fully complete 00:01:45.27 in the numbers, in several cases even when one has numbers one still has to struggle with 00:01:52.09 the meaning for those numbers, I think the important thing 00:01:55.12 is the direction that they point out. So, what we saw and what we've learned 00:02:05.23 in the past twenty years in Drosophila development is that patterning 00:02:10.16 along the anterior to posterior, that is head to tail axis of the embryo depends 00:02:16.13 on the presence of a maternal protein called bicoid that's graded such that 00:02:22.20 individual genes like hunchback are activated at particular times and at particular places 00:02:31.07 in the embryo. Now the interesting thing about this information rich maternal gradient 00:02:41.06 of the bicoid protein is that it arises from the synthesis of the bicoid protein from 00:02:47.14 the localized RNA at the anterior end of the egg. And it's that process 00:02:51.27 that I'd like to talk about. What we were interested in, what Thomas in particular 00:02:55.07 was interested in was in this simple cartoon sense, of how an information gradient 00:03:03.25 like bicoid arises, you have a localized RNA, synthesis, and then movement 00:03:10.19 of the protein. And most of our knowledge and all the pictures 00:03:13.20 that you see back there were made in fixed embryos where you fix an embryo 00:03:19.11 you stain it with something that allows you to see particular proteins like bicoid protein 00:03:23.27 or do in situ hybridizations to identify the localization of the RNA, and that approach 00:03:32.18 of fixed material has been extremely valuable in almost all of developmental biology 00:03:38.04 over the past 20 years, but it has two essential flaws. One is that most of these 00:03:44.07 techniques are indirect, so you don't see the molecule itself, you see something 00:03:48.17 that binds to something that binds to something that binds 00:03:51.22 to the molecule, and you are never really sure then about the levels of staining 00:03:57.03 or the intensity of the staining that you see and how that relates to absolute 00:04:00.28 concentrations. And in a model, like the one we are dealing with here, where 00:04:05.24 the fates of cells are controlled by concentrations that you want to measure you'd really 00:04:11.15 like to have, as direct as possible, a measure of actual concentration in the given 00:04:17.27 region of the embryo. The other problem with fixed material obviously is that it's fixed. 00:04:21.29 You look at a given embryo, it is fixed at a given stage, even if you could measure it 00:04:26.18 at that stage, you don't really know what the levels were before or afterwards. 00:04:32.01 And so for those reasons, what Thomas Gregor decided to do in the lab was to develop 00:04:37.20 a living probe, a probe that could allow us to follow the bicoid protein in living embryos 00:04:44.26 and to do that he established a fusion protein of bicoid to the fluorescent EGFP protein 00:04:53.04 and followed its development in embryos and I'll show you that in the next images 00:04:59.14 and these are actually living photographs at different levels of a single Drosophila embryo 00:05:05.06 that expresses this EGFP bicoid transgene that Thomas Gregor made. Now the transgene 00:05:12.27 that he made is under control of the normal bicoid transcription controls during 00:05:18.29 oogenesis. It also has the normal 3'UTR of the bicoid RNA, which is the part of the RNA 00:05:25.15 that is involved in localizing that RNA to the anterior end of the egg. And so 00:05:30.03 the EGFP-bicoid is expressed as RNA in the mother, localized appropriately, and in fact if you 00:05:39.15 put it into a mutant embryo, mutant for bicoid, it will rescue and produce perfectly 00:05:45.02 normal development, indistinguishable from that which you see in wild-type embryos. 00:05:50.14 Now, while I've been talking, this embryo has established a bicoid gradient, and I think 00:05:59.16 what we need to do is to go back one time and watch this movie one more time. 00:06:06.02 So what we're going to see is the bicoid gradient in this living embryo and in the embryo 00:06:14.16 immediately after fertilization you can see some movement in the cytoplasm here. 00:06:18.18 What you can't see yet is an obvious difference in the concentration of fluorescence 00:06:25.11 at the anterior end and the posterior end. And over time though, you can begin to see 00:06:29.27 over the two hour period of time when this, this is the nuclei now have made it out to the surface 00:06:36.17 and you can see in those nuclei at the anterior end of the egg, you can see an 00:06:42.20 accumulation of bicoid. You can see that when the nuclei divide, bicoid protein goes out 00:06:47.25 of the nuclei. When the nuclei reform after each division the bicoid protein goes back in. 00:06:54.00 What you can clearly see in this embryo is that along the anterior/posterior axis of the egg 00:06:59.16 is a graded distribution. So the nuclei at this end have much more bicoid protein 00:07:04.18 and the protein levels drop off, such that a cell really could figure out where it was 00:07:10.25 in the egg. Whether it was the hundredth cell at this end or the zero cell or the one cell 00:07:16.27 or the tenth cell just based on what its position is here. Now, using this construct 00:07:28.07 because we can accurately time and say exactly how old the embryo is how many minutes 00:07:36.19 it's been since it's done any particular division, we can image these embryos with 00:07:42.03 two photon microscopy that allows us to get very good in-depth resolution. We're able to 00:07:49.11 and because we're actually looking at EGFP bicoid, we're looking at the protein that is 00:07:54.28 actually responsible for the pattern, we can measure the intensity and we can measure 00:08:00.22 the profile of the gradient. So what it allows us to do is to ask relatively simple questions. 00:08:06.24 If the RNA is localized at the anterior end of the embryo and it begins translation at 00:08:14.20 fertilization, protein is being made, translation is continuous if the embryo is making more 00:08:22.12 and more bicoid protein, then how is it, how does the profile of bicoid concentration 00:08:31.21 along the length of the embryo change? Does it continue to rise? Does it stabilize? 00:08:36.09 Stabilization would actually be really important because if the nucleus wants to know 00:08:41.17 where am I along the anterior/posterior axis of the egg, the easiest way of doing that is if 00:08:47.23 the concentration at that point, 30% egg length really were constant and were constant 00:08:53.10 over time. So one of the things that Thomas did by following individual embryos from 00:08:59.08 the time when the nuclei first migrate out to the surface, and you can measure the bicoid 00:09:03.06 concentration. Measuring exactly the same position, even when nuclei divide, looking 00:09:09.24 at the daughter nucleus that comes to lie closest to the position of the mother nucleus 00:09:16.07 and ask when does the concentration in a given area stabilize? And what Thomas found, 00:09:22.04 and this was quite remarkable, is that at the earliest points when we could measure the RNA, 00:09:26.24 that is at cycle 10, just after 10 cycles to just when the nuclei made it out to the surface 00:09:32.12 the concentration in a given nucleus was fixed for that position and remained constant 00:09:41.04 at that position and remained constant throughout all the remaining cleavage divisions, 00:09:47.11 that is the concentration is stable all the way up to cycle 14 when we see massive 00:09:53.07 visible expression of the downstream targets. The other thing that was really important 00:09:58.16 for us was to look at many embryos on the same slide and ask in 10 different embryos, 00:10:05.21 if you look at the bicoid concentration, how similar is this concentration at 30% from one 00:10:12.27 embryo to the next to the next. Because we know that ultimately all 10 embryos 00:10:16.18 will develop with exactly the same gene expression patterns, and so if the concentration 00:10:21.22 of bicoid really is directly reflecting or is directly controlling those gene expression patterns, 00:10:29.29 then in individual embryos somehow at 30% egg length you've got to have the same bicoid 00:10:35.27 concentration in all those different embryos and quite remarkably 00:10:39.07 and something that we had never really been able to convincingly see in fixed material 00:10:45.08 because of the variabilities of fixation and staining was there's very very little 00:10:50.05 variation from one embryo to the next. And all of those observations were really 00:10:55.17 important because it pointed out to us that bicoid and the nuclear concentration 00:11:02.15 of bicoid can provide stable sufficient information about position to activate precise 00:11:11.27 patterns of gene expression and that it could do this from cycle 10 all the way 00:11:17.12 up to cycle 14. Now, another thing that you can do if you believe that 00:11:26.18 you can actually measure the concentration of the molecule, you can actually measure 00:11:30.20 and describe the distribution. And if you look at the bicoid distribution in embryos what 00:11:38.10 you can see is that it's as expected, highest at the anterior end here and falls 00:11:44.05 and you can show that this fall can be fitted to an exponential decay which is what one 00:11:50.07 would predict if you had a source and you had simple diffusion from that source. 00:11:57.01 Because the concentrations are so reproducible, particularly in the region right here 00:12:02.23 where we're activating hunchback gene expression you can measure those concentrations 00:12:10.04 measure the concentration of bicoid in a given nucleus and the output 00:12:15.29 of that concentration in terms of hunchback protein expression. One of the interesting 00:12:23.16 things is that if you look in these embryos at precisely at the region where hunchback 00:12:30.16 is being expressed right here, and you ask how much of a difference is there 00:12:38.26 in bicoid concentration in nuclei between those that are going to express hunchback 00:12:45.08 and those that aren't going to express. It's basically a 10% difference. The cells that are 00:12:51.13 making the choices, their neighbors either will not express hunchback protein 00:13:01.13 or express hunchback protein based on only a 10% change in the concentration of bicoid. 00:13:07.25 So what that means is that cells have an extraordinary ability to measure accurately 00:13:12.21 what these concentrations are. You can plot out that response. You can see that it's highly 00:13:18.11 precise you see hunchback expression within one or two cell diameters at 48% egg length 00:13:26.05 and that it's also highly non-linear, meaning that a small change in bicoid concentration 00:13:34.04 results in a huge on/off change in the response in terms of gene activities. So you could 00:13:43.06 think about how this non-linearity arises and there's a number of experiments from 00:13:48.28 a number of different labs that indicate that if you look at the control regions of the 00:13:53.20 hunchback gene, that there are multiple bicoid binding sites and you can begin to think 00:13:57.20 about this in terms of cooperativity of bicoid binding. And when you do that you model 00:14:03.03 the non-linearity and you come up with a Hill coefficient for this cooperativity of 5, 00:14:08.00 which is a very high degree of cooperativity, which means that very little change 00:14:13.13 in bicoid results in massive changes in gene expression. Now all that means 00:14:22.13 is that the relative numbers, because really if you think about what we are doing 00:14:25.03 where we are looking at these embryos is that we are looking at EGFP, we're measuring 00:14:28.29 the EGFP that's attached to individual bicoid molecules, measuring the intensity 00:14:34.06 of the signal and trying to extrapolate concentrations of functional protein 00:14:40.12 in the nucleus that are activating genes. What we'd really like to be able to do is not to talk 00:14:46.24 about bicoid concentration in terms of intensity, but to actually talk about it 00:14:50.16 in terms of number of molecules, in terms of absolute concentration. And the experiment 00:14:55.02 that Thomas did to begin to approach this is something that you can do uniquely in flies 00:15:01.14 in part, or at least it was easier to do in flies, is that we're using EGFP and many argue 00:15:08.21 that you can use type proteins in many organisms, but fly embryos develop 00:15:14.17 inside a water impermeable but completely transparent egg shell called the 00:15:21.07 vitelline membrane and so what Thomas was able to do was to take these EGFP-labeled 00:15:26.29 bicoid embryos and immerse them and image them in a solution that had 00:15:34.03 defined molarity of bicoid, in fact 36 nM bicoid and so what this meant is that he could 00:15:39.20 look at the concentration or measure the intensity of the bicoid EGFP signal 00:15:45.24 inside the embryo and compare it to an absolute concentration of bicoid 00:15:51.06 immediately outside the embryo. And this allowed him then to show that the actual 00:15:55.14 concentration of bicoid in the nucleus in those nuclei that are actually making 00:16:00.28 the choice whether to activate hunchback expression or not, was of the range 00:16:05.06 something like 8 nM. That's an interesting number because it's close to what might be 00:16:11.16 predicted from in vitro binding studies for what would be required 00:16:15.13 for activating and binding to and activating the hunchback gene. But it's actually interesting 00:16:21.11 from an additional standpoint, and one that raises I think one of 00:16:26.15 the really fundamental questions about transcription. If you take the concentration 00:16:30.13 of bicoid, this 8 nM value, and you know the actual volume of the nucleus 00:16:37.27 because you can measure that optically from the images that you have, 00:16:42.18 you can calculate the total number of bicoid molecules in the nucleus, 00:16:46.26 in a nucleus that is making a decision. When Thomas did that, he found that 00:16:50.08 there were about 697 molecules per nucleus. This is a remarkably small number 00:16:57.22 if you think about it, 600 or 700 molecules in a nucleus if we consider that 00:17:05.17 the neighboring nucleus that is choosing not to activate hunchback 00:17:13.00 or choosing to activate hunchback only varies by 10%, and so what it requires 00:17:21.29 is that individual nuclei distinguish between whether in their total nuclear volume 00:17:28.23 there's 690 molecules or 630 or 770. The relatively small number of bicoid molecules 00:17:39.19 that are present in the nuclei, that are making accurate decisions raises a lot of questions 00:17:45.21 about how bicoid or how any transcription factor controls or activates transcription 00:17:53.15 in a concentration dependent way. Part of the question is that we don't really know, 00:17:58.04 we assume that bicoid activates transcription of hunchback by binding to 00:18:04.16 a hunchback control region and activating transcription by interacting with 00:18:09.18 other proteins that are the actual transcriptional activators. What we don't know 00:18:14.00 about bicoid and hunchback is the extent of this binding and occupancy 00:18:21.06 how long it lasts, and whether it's permanently associated when genes are active 00:18:27.06 whether bicoid molecules that are bound or that bicoid molecules responsible 00:18:32.13 for that activation remain permanently bound or whether there's a dynamic 00:18:36.07 exchange between bicoid and individual transcription factors. And that becomes important 00:18:43.15 because actually hunchback is not the only gene in the Drosophila genome that can 00:18:48.21 respond to bicoid transcription, and there are various informatics estimates that suggests 00:18:54.03 that there's anywhere from several thousand based on informatics to several hundred 00:19:01.16 based on biochemical studies, several hundred genes capable of binding to 00:19:07.06 and responding to bicoid. So if all of these genes are trying to measure bicoid concentration 00:19:12.22 and binding bicoid concentration at the same time, and there are only 630 molecules 00:19:17.19 in the nucleus, the question of occupancy becomes really important and it drives 00:19:22.15 us to models where what's actually required is, for cells to make choices,. The 00:19:35.01 concentrations are based in some way by averaging collisions over time, averaging 00:19:44.13 interactions between bicoid molecules over time, such that molecules of bicoid diffusing 00:19:51.29 throughout the nucleus will bind or interact with a promoter and then leave and be able to 00:19:57.21 interact with other genes in the genome. But to measure concentration in those molecules 00:20:03.18 what you have to be able to do is you have to have some memory, some way of recording 00:20:10.06 individual collisions and averaging them over time to come up with an accurate 00:20:15.00 measurement of concentration. The problem is actually, to make the problem graphic just 00:20:23.08 one last cartoon like illustration, but what a nucleus or what a hunchback gene, what any 00:20:32.12 gene in a nucleus has to do with response to transcription is similar to what a piece of DNA 00:20:40.03 or say my knuckle here, would have to do if sitting in the middle of this room 00:20:45.03 where I'm giving this talk, if there were 690 molecules of bicoid flying around 00:20:50.19 and it's trying to judge from collisions how many molecules are actually in this room, 00:20:56.06 and trying to figure out whether it is in a room with 690 or 630 because it's activation 00:21:04.17 is going to depend on that. And other pieces of DNA, if we accept 00:21:12.06 the transcriptional threshold model, accept the idea that the bicoid gradient is providing 00:21:17.18 pattern along the entire access of the embryo, other genes responding and 00:21:22.25 measuring concentrations are sensitive to activation at other concentrations. 00:21:28.00 So one of the other things that we don't know, and I think it's really important to know 00:21:31.16 about the whole response is to get a more global sense of what genes respond to bicoid, 00:21:40.02 how do they respond, and do they show the same levels of accuracy 00:21:49.07 and their ability to measure molecules. One possibility would be that the embryo 00:21:53.27 uses certain genes like hunchback as guideposts, establishes them with great accuracy 00:21:59.20 and then determines other genes which appear to respond, like these genes 00:22:05.01 that I've listed here, this orthodenticle, giant, or Krüppel, obviously are also responding 00:22:12.20 in some way to the bicoid gradient, but whether those responses 00:22:20.10 are necessarily as accurate as hunchback, or whether part of the accuracy if 00:22:24.16 they are as accurate, whether part of that accuracy depends on interactions or 00:22:29.26 subsequent interactions that depend heavily on the interactions of guidepost genes 00:22:34.20 like hunchback. The way that the bicoid gene activates transcription 00:22:42.20 is a fascinating problem but it’s not the only thing that we don't know. 00:22:45.19 Another really interesting problem is how it is that the gradient is actually established 00:22:54.06 in a stable form. How does bicoid protein move. If it's being made at the anterior end 00:22:59.23 of the egg, how does that movement of bicoid protein from the site of its synthesis 00:23:07.11 how is that able to establish a stable gradient. And in particular, in that simple 00:23:13.09 cartoon models that we have, that those protein molecules unlike the RNAs 00:23:19.03 are not anchored would be able to move by diffusion. So what we'd like to know is what are the 00:23:24.22 mechanisms that actually move bicoid molecules newly made from the anterior end 00:23:31.24 and establish the gradient. In simple kinds of experiments where you follow 00:23:37.14 the establishment of the gradient over time. Look at how soon you can begin 00:23:44.12 to detect molecules at different distances from the anterior end of the egg, 00:23:48.12 and you model the kinds of diffusion constants, the rates that molecules have to move 00:23:54.07 within the egg to establish a stable gradient. Most of the modeling that's been done 00:24:01.03 suggests that you need diffusion constants of the order of 00:24:05.17 about 4-8 microns squared per second. Once we had an EGFP-bicoid molecule 00:24:14.09 though, that allowed us to do photobleaching experiments 00:24:17.07 or to tag individual molecules and follow them. What Thomas Gregor 00:24:21.29 was able to do was to actually measure the movement of bicoid molecules 00:24:27.14 in small little spaces using photobleaching experiments over small volumes 00:24:32.16 and over small times to ask how fast does bicoid molecules actually move in the surface, 00:24:40.14 and the remarkable conclusion, the one that was surprising to all of us was that 00:24:45.08 if you look at those movements and those measurements that you can measure 00:24:48.28 you find that they are very small, that the diffusions constants are very small. 00:24:52.18 The molecules move very slowly. The best measurements that we have 00:24:58.18 from Thomas's data and constructs suggest that the diffusion constants are 00:25:04.02 on about the order of 0.3 microns squared per second, which is ten-fold, twenty-fold 00:25:08.20 less than what you'd actually need to visibly establish the gradient. 00:25:13.07 So a value of 0.3 microns squared per second is too small, molecules would move too 00:25:21.05 slow to, it would require several hours to produce a gradient of the kind that we see 00:25:27.17 in the Drosophila egg, and yet we know that the gradient is already there 00:25:33.00 and stable by cycle 10 and already the effects of that gradient in terms of 00:25:40.21 gene expression patterns, are also already clearly visible at that stage. 00:25:48.04 So we don't simply have enough time in development for a gradient to be established 00:25:54.00 with diffusion constants that we see. And there are other theoretical considerations 00:25:58.04 which one argues that the establishment of a stable gradient in some way balances, 00:26:06.20 the shape of that gradient somehow balances the movement of molecules versus 00:26:11.21 their constant degradation and under those circumstances you would need 00:26:15.10 half-lives to produce the visible shapes of the bicoid gradient that we see 00:26:20.23 you would require half-lives that are extremely long for the bicoid molecule 00:26:28.17 to produce through gradients, at any time in development, to produce 00:26:34.01 the gradients that we see with a diffusion constant of 3.5. So all of that basically 00:26:39.23 just raises the basic problem of that we don' t know really how molecules 00:26:46.23 move in the egg, and we don't really have good handles or good strategies 00:26:51.26 for tracking individual molecules or even whole populations of molecules 00:26:56.22 and knowing at what scale and over what time frames we have to 00:27:00.06 measure molecular movement. One of the other strategies that Thomas Gregor 00:27:06.02 took to follow movements of molecules involves experiments where 00:27:13.16 rather than looking at EGFP and making assumptions about its translation 00:27:19.07 RNA distribution and movement from the eggs, what Thomas did was to take embryos 00:27:25.27 wild-type embryos and inject fluorescently labeled compounds into 00:27:30.05 the anterior ends of the eggs, and you can see in this panel right here 00:27:34.07 the consequence of when you inject a dextran into the anterior end of the egg, 00:27:38.29 you can see this biologically inert but fluorescent molecule moving through 00:27:44.29 the whole volume of the egg and you can follow the change in the distribution of 00:27:52.03 these molecules over time you can model them. And what Thomas was able to show 00:27:56.07 is that when you take and inject dextran into the egg, it moves, it can be modeled 00:28:01.14 by its distribution over time, and it follows what would be expected of 00:28:08.00 simple diffusion and allowed Thomas to calculate for dextrans of different sizes 00:28:16.00 the diffusion constants for molecules. What this graph here shows is just some of 00:28:25.05 Thomas's data for trying to model bicoid movement by using dextrans 00:28:34.06 of approximately the size of bicoid. So bicoid with EGFP would be about 00:28:40.03 a 70 kDa protein, and so if you inject a dextran of 70 kDa into the egg 00:28:47.11 and ask how fast does it move compared to what he was measuring with EGFP bicoid, 00:28:53.11 what Thomas saw was that in contrast to the 0.3 microns squared per second 00:29:00.12 diffusion constants that he obtained for bicoid at the surface of the egg and 00:29:05.26 when photobleaching small volumes of cytoplasm, the dextrans apparently move with 00:29:13.18 diffusion constants with speeds of 15 microns squared per second. 15 microns is 00:29:21.08 more than enough of what you need to make a gradient. And so this means that 00:29:25.06 at least some molecules moving from the anterior end of the egg 00:29:29.09 can produce gradients of the kind that we see with bicoid. 00:29:40.09 Now, strictly speaking, what Thomas's measurements argue for dextran is that 00:29:49.09 you can model the movement of injected dextran as though it were diffusion. 00:29:54.17 It doesn't actually tell you that molecules are moving by diffusion rather than 00:29:59.18 being transported or binding to something or being moved by other mechanisms. 00:30:05.12 One of the approaches, one of the strategies, for testing whether 00:30:08.12 the molecular movement that you are looking at in a biological system is due to 00:30:14.10 diffusion or to Brownian motion if you will, is its size dependence in that the 00:30:20.24 Stokes Einstein relationship argues that small molecules will diffuse faster, 00:30:28.02 larger molecules will diffuse more slowly and that the diffusion constant 00:30:32.08 that you measure will depend on the size of the molecules. And so the curve 00:30:36.29 as you see here, you see the behavior of the diffusion constants that you measure 00:30:43.10 for dextrans that have small radii versus large radii and you can see 00:30:52.05 the size dependence of those diffusion constants which argues that 00:30:55.29 a significant component of the movement that he's looking at really fits 00:31:00.09 Stokes/Einstein, really is a physical diffusion. But one of the other interesting things 00:31:05.01 that arose from Thomas's analysis and by fitting these curves is that if you follow 00:31:11.20 the shape of this curve out as molecules get bigger and bigger, what you find 00:31:21.12 is that while the curve can be fitted to a Stokes/Einstein relationship 00:31:26.22 it doesn't go down to zero. There's a floor, a fraction of the movement 00:31:33.27 of every one of these particles, which is size independent, meaning that 00:31:38.27 about 6 microns squared per second of the movement of a molecule of about 00:31:43.07 the size of the radius of bicoid is independent of its actual size and all molecules 00:31:53.09 are going to move in the egg at about 6 microns squared per second 00:31:58.07 independent of their size. This is actually a very intriguing observation for us 00:32:03.05 because it suggests that a molecule, even a molecule like bicoid 00:32:08.22 that may show very slow diffusion constants perhaps because it’s interacting 00:32:19.21 with or being bound to other molecules and therefore ultimately we can explain that 00:32:25.10 as having an increased size. If you look at values at the behavior here, 00:32:30.26 this size independent movement might actually account for the movement of bicoid. 00:32:37.04 And we don't know what the nature of the size independent movement is, 00:32:40.14 but one model, all that we know is that it can't be explained by Brownian diffusion. 00:32:47.20 But one of the interesting models may come back from 00:32:52.27 the biological understanding of the phenomenon, because if you actually 00:32:56.29 look at eggs during the process when this gradient is being established, 00:33:06.07 in a simple sense, the way we've talked about it before, 00:33:10.17 we've had a localized RNA, which is translated into a protein and you have diffusion 00:33:14.28 of this molecule through what you can think of as a stable cytoplasm 00:33:18.17 such that diffusion could establish a stable gradient. If you actually look though, 00:33:23.22 at embryos during the process when this bicoid gradient is being established, 00:33:30.06 and you look at the cytoplasm, we're looking here at an early cleavage division 00:33:33.06 in embryos that carry a GFP-histone so we'll be able to eventually 00:33:39.29 to see the nuclei, but you'll also see some GFP-histone in the cytoplasm of these 00:33:44.12 syncytial embryos. You can see the nuclei migrating out to the surface 00:33:49.18 and you can see this pattern of division of the individual nuclei 00:33:55.04 that I talked about in the first lecture that give rise to these synchronous divisions 00:33:58.27 that give rise to the syncytial blastoderm, but one of the things that you can also see 00:34:07.17 and I'll point this out to you when we watch the move one more time, 00:34:15.11 is that associated with these nuclear replications are massive movements 00:34:21.02 of the cytoplasm. The cytoplasm moves forward and moves back and this swishing 00:34:26.13 kind of turbulent patterns that appear to have no overall directionality 00:34:34.27 are not going to move molecules in any particular direction, but 00:34:39.22 will we believe contribute, if bicoid is being associated with the cytoplasm, 00:34:46.14 will contribute to the movement of bicoid in a non-Brownian, non-diffusion 00:34:51.18 sense, but in a non-directed sense that will be essentially equivalent 00:34:56.04 to the random walks that are produced by diffusion. So what we're 00:35:01.08 beginning to think now is that the bicoid gradient that arises in the egg 00:35:06.20 does not arise necessarily directly from diffusion or from the diffusive movements 00:35:14.29 of bicoid molecules per say, but is actually established by random movements 00:35:20.07 in the cytoplasm. If that is true then what it will argue is that the establishment 00:35:30.25 of the gradient can't be understood from simple biophysical properties 00:35:37.17 like diffusion. That it requires that the egg cytoplasm and the motors in the cytoplasm 00:35:43.11 maybe in a totally undirected way but still that these cytoplasmic flows 00:35:48.28 establish and move molecules like bicoid and that the stable gradients that we see 00:35:56.24 are the products more of that movement. We're beginning to try and test those 00:36:01.23 models by asking can we inhibit this movement, can we follow the establishment 00:36:06.25 of gradients in unfertilized eggs where we see different patterns of movements 00:36:10.19 but the possibility that cytoplasmic movements rather than simple physical diffusion 00:36:18.23 establish the bicoid gradient is an intriguing possibility for us because it suggests 00:36:23.29 that if biological parameters like cytoplasmic flows control the ultimate shape 00:36:33.27 and distribution of bicoid, then those biological parameters 00:36:38.03 in fact can become immediate targets in a biological process 00:36:43.20 if you wanted to change the shape of the bicoid gradient or have the bicoid gradient 00:36:47.26 move during the course of evolution. As eggs change or times change, 00:36:57.04 if you want to maintain or continue bicoids usefulness to use bicoid as 00:37:04.00 a morphogen, as a molecule whose concentration establishes gene expression 00:37:09.07 you possibly want to be able to manipulate its distribution beyond those things 00:37:16.11 that are possible by simple physical parameters like diffusion. So in the last 00:37:20.07 part of this lecture, I'll talk a little bit about what we learned about how 00:37:25.21 bicoid distributions have changed during evolution.
