Patterning Development in the Early Embryo: The Role of Bicoid
Transcript of Part 3: Evolution of Bicoid-based Patterning in the Diptera
00:00:04.17 My name is Eric Wieschaus and I'm a HHMI investigator and professor 00:00:09.08 at Princeton University. For the previous two parts of this lecture we've 00:00:15.20 been talking about how pattern is established in the early Drosophila 00:00:18.22 embryo, and we focused on a molecule called bicoid that is deposited as an RNA 00:00:25.12 at the anterior end of the Drosophila egg during oogenesis in the mother. 00:00:30.25 And then is translated into a protein and forms a protein gradient in 00:00:36.19 the early embryo. And the model and then the reason why this protein, 00:00:43.11 this bicoid gradient is important is that it is thought to be the major 00:00:50.11 determinant in establishing the pattern of gene transcription in the embryo 00:00:55.03 such that different concentrations of the bicoid protein at different points 00:00:59.29 along the anterior posterior axis activate expression of particular genes 00:01:06.25 like the hunchback gene in green here, or the Krüppel gene in red. 00:01:11.18 Now, we've seen that if you examine this gradient in Drosophila embryos 00:01:17.12 or if we examine the expression of the downstream targets, we see that their 00:01:21.12 extraordinarily constant from one embryo to the next. And this is probably 00:01:26.24 what you want if you want to have a system which is establishing pattern 00:01:30.00 and controlling the behaviors of individual cells in the embryo. Now 00:01:35.10 in the last lecture we talked about some of the biophysical parameters 00:01:40.22 and cell biological parameters that might give rise to these constant 00:01:45.05 distributions of bicoid or constant transcription patterns. 00:01:51.05 Now one of the underlying assumptions for all of that work 00:01:59.07 is based on actually a fact. If you look at fly eggs, not only are the expression 00:02:06.01 patterns constant, but the actual sizes of the eggs are constant. And that's 00:02:10.29 an important idea because if you think about any of the mechanisms 00:02:13.29 that we think about when we talk about how one would establish a gradient, 00:02:17.12 they are very sensitive to the size of the egg. If you want to establish 00:02:25.02 hunchback expression up to 48% egg length, and you are doing that 00:02:30.12 by having a gradient of a molecule that diffuses with a particular 00:02:35.24 diffusive constant and establishes a particular constant shape 00:02:39.21 in its distribution as bicoid appears to do, then that mechanism for 00:02:45.13 patterning would be very sensitive to variation in egg size. 00:02:52.19 And if you look at Drosophila eggs in the particular way you raise them 00:02:55.21 in the laboratory, in the way that we measure them, in the stocks 00:02:59.17 we've measured them to establish these gradients and look at them. 00:03:01.29 What we see is that actually individual fly eggs are remarkably similar, 00:03:08.05 wild-type normal eggs are remarkably similar in their size, and so it's basically 00:03:12.09 consistent that such a mechanism could function to pattern 00:03:19.01 embryonic development in Drosophila melanogaster. The problem though 00:03:24.02 with that model arises if you go outside of Drosophila. If you go outside of 00:03:29.12 fruit flies and extend your observations to other kinds of embryos 00:03:35.26 from even other embryos of other insect species, even other fly species. 00:03:39.28 As we all know, from our own personal experience, flies come in different sizes. 00:03:46.13 There are the nice little small Drosophila fruit flies with red eyes 00:03:52.01 that we kind of raise in the lab and have such affectionate feelings for 00:03:56.05 and there are also disgusting flies like house flies and blow flies that 00:03:59.21 kind of fly around and invade our picnics and are less attractive. 00:04:07.06 The bigger flies are the uglier ones, the smaller flies are nicer and sweeter, generally. 00:04:13.29 Now what is also true though, not only is the adult flies that we see 00:04:19.25 are of different sizes, but also if you look at the eggs in the embryos 00:04:23.27 that these different fly species make, they are also different sizes. 00:04:29.16 So Drosophila melanogaster for example makes eggs that are about 00:04:34.04 500 microns long. There are even Drosophila species like 00:04:37.16 Drosophila busckii that make eggs that are even smaller than Drosophila, 00:04:43.20 but the big flies like Musca domestica, the house fly, 00:04:49.01 or Calliphora, blow flies or green bottle flies, Lucilia, that are big 00:04:55.29 obviously as adults much bigger than Drosophila, and the eggs 00:05:00.01 that they make are substantially bigger. Now all of these flies 00:05:08.15 these higher Diptera are closely related. And even though 00:05:14.01 bicoid as an RNA or a gene product was a newly evolved solution 00:05:19.25 to the problem of patterning in the embryo and arose during the evolution 00:05:23.17 of the Diptera, all of these insects here share the common feature 00:05:29.13 that their anterior/posterior patterning depends on bicoid. 00:05:33.20 And yet the bicoid gradients that are forming in these eggs are forming 00:05:39.18 in eggs which are very large or very small. Now, one of the interesting 00:05:47.08 features of all these eggs though is even though they are different sizes 00:05:51.12 if you actually look at the development of these embryos 00:05:54.08 you can see that the early development is remarkably similar in that if you 00:05:58.24 go back and remember how early development in Diptera starts with 00:06:02.18 fertilization followed by these nuclear cleavages in the syncytial embryo 00:06:06.21 and then a pause at 2.5 hours to form the syncytial blastoderm 00:06:14.25 that then transforms itself by the formation of cell membranes between 00:06:19.01 nuclei and into a cellular blastoderm and a gastrula, that process 00:06:22.17 that takes about 2-3 hours in Drosophila melanogaster 00:06:26.11 is also observed in all of these other insects and it also is observed 00:06:31.03 with exactly the same time, 2.5 hours, the same kinetics 00:06:35.17 and if you look at divisions of the nuclei, if you look at individual, 00:06:41.04 say this is an embryo, at the syncytial Drosophila embryo from 00:06:46.14 the big green bottle fly, Lucilia versus Drosophila melanogaster 00:06:52.08 or Drosophila busckii, the eggs have the same shape 00:06:56.14 and although you can probably barely make it out, you can barely see the nuclei 00:07:02.28 in the Lucilia and the nuclei in Drosophila melanogaster or busckii 00:07:07.04 are somewhat smaller, if you blow up the pictures of the individual eggs 00:07:12.00 and look at the nuclear distributions you can see that all these eggs 00:07:16.09 at 2.5 hours have the same number of nuclei. They all have about 100 nuclei 00:07:21.00 going from the anterior to the posterior end of the egg. 00:07:23.23 They're all being patterned over the same time constraints 00:07:27.15 and they're all being patterned by bicoid. If you look at transcription 00:07:35.13 the other remarkable similar thing between all these insects that 00:07:38.25 are all so closely related even though their sizes are different, is that 00:07:42.15 all of them activate transcription at this critical 2 hour period in response to bicoid 00:07:50.19 and if you look at the patterns of gene expression 00:07:52.23 if you look at say hunchback or giant, the two different gap genes 00:07:56.02 in Drosophila or in Musca or pair-rule genes like 00:08:02.23 paired or evenskipped, they show exactly comparable scaled patterns. 00:08:09.00 Even though the eggs are bigger and even though the cells are bigger 00:08:13.13 the patterns per cell are exactly the same. Now these are transcriptional 00:08:18.20 responses, they're genes which are transcribed at the blastoderm stage 00:08:24.05 directly or indirectly in response to the bicoid gradient. And so the question 00:08:29.17 that you'd like to ask, is how is it during the course of evolution, 00:08:34.03 as egg size changes, how does the embryo or the species adjust to using 00:08:43.03 a bicoid gradient to establish pattern. There are really two simple ways 00:08:48.01 that you can think about it. One way is that each of these genes, 00:08:52.05 like hunchback and any of the targets of bicoid activation 00:08:56.14 is going to have a control region which will respond to bicoid concentration. 00:09:03.04 And as you change the length of the egg, one strategy 00:09:08.04 would be to keep the bicoid gradient the same shape 00:09:11.20 and the concentration distribution the same, and yet change 00:09:16.20 the cis-acting control regions of each of these genes. Adjust them 00:09:19.28 during evolution and we know that that's generally what happens 00:09:22.18 during the course of evolution. Alternatively, during evolution 00:09:30.05 you could adjust with the size of the egg not by changing 00:09:35.15 the control responses of individual genes, but by somehow 00:09:39.03 changing the manner or changing the physical properties 00:09:43.07 that establish the bicoid gradient itself such that in bigger eggs 00:09:50.10 the gradient extends longer and in smaller eggs it is shorter. 00:09:57.15 This seemed to us initially a less likely alternative, partially because 00:10:03.19 many of the cases in evolution that we know about involve 00:10:06.26 changes in cis-acting control regions. But when we actually looked 00:10:10.16 at the bicoid distribution in these other insect species, what we observed 00:10:15.14 was that not only does bicoid form gradients in big eggs and the small eggs 00:10:24.14 but if you look at the big egg, if you compared the distribution to say 00:10:29.14 in melanogaster. In melanogaster the gradient falls exponentially 00:10:34.10 over an area like this, and if you look in Lucilia the distribution 00:10:41.11 of bicoid protein extends much longer, much farther 00:10:46.17 into the length and into the egg. In terms of microns, that is the bicoid gradient 00:10:53.13 in a bigger egg, is bigger, proportionately bigger, because if we then 00:10:59.03 replot the data, not in terms of absolute lengths as here 00:11:02.29 but in terms of relative lengths along the eggs you can see that 00:11:07.01 the bicoid gradients in the big eggs and the small eggs are 00:11:10.08 exactly equivalent scaled to the size of the egg. What that means then 00:11:16.14 in turn, is that somehow during the course of evolution of these insects 00:11:23.03 the bicoid gradient has changed. The properties that establish the gradient 00:11:30.18 have changed to allow this gradient to now span 00:11:35.03 a bigger or smaller egg and can provide positional information along 00:11:44.05 the whole length of the egg. So how does this happen? One simple strategy 00:11:52.15 would be to change bicoid as the species evolve they change bicoid, 00:12:00.07 they change the properties of the bicoid protein such that it moves faster, 00:12:03.13 such that it is degraded less rapidly for example, such that it ultimately 00:12:07.19 the gradient that you get out of this bicoid would extend farther 00:12:15.23 and thus establish gradients of comparable shape when 00:12:21.05 scaled back to the actual shape of the egg. To begin to test or 00:12:27.22 think about those models we've cloned the bicoid genes from these 00:12:33.12 different species and compared their structure to 00:12:40.06 that of Drosophila melanogaster bicoid, and if you look at that 00:12:43.06 bicoid is reasonably well conserved particularly in regions of the protein 00:12:48.28 which are functionally well defined. The homeodomain that binds DNA 00:12:52.23 or other regions that have been implicated at least suggestively 00:12:56.11 as being involved in protein stability. Most of the sequences are the same 00:13:06.04 but not surprisingly when you look at any particular sequence, 00:13:10.09 any particular region of these proteins, there are amino acids differences. 00:13:14.25 And so one possibility is that these different species specific bicoids 00:13:20.02 have evolved and the changes in their sequence, either the ones 00:13:24.08 that I've indicated here or changes in other regions of the protein 00:13:28.15 are actually responsible for adjusting the shape of the gradient 00:13:35.10 such that it can now function in larger eggs or smaller eggs. 00:13:40.10 So to test that possibility, you need to begin to hope to identify 00:13:45.04 the regions that have changed in the bicoid protein. 00:13:48.08 What Thomas Gregor and Alistair McGregor in the lab did was to 00:13:53.00 take these cloned bicoid genes from the other species, 00:13:58.03 tag them with EGFP and put them back into melanogaster to ask 00:14:04.07 what type of gradients that they make. And the surprising result here, 00:14:08.19 one that we hadn't anticipated was that if you compare, 00:14:13.29 if you take a bicoid protein from Calliphora for example 00:14:17.14 that will make a large gradient that extends more than a 00:14:23.05 millimeter through the entire Calliphora egg which is 00:14:27.19 one and a half millimeters long, and you put it into a Drosophila egg 00:14:33.01 which is only five hundred microns long, one possible result would have been 00:14:38.25 that this bicoid protein because of its changed properties, the protein 00:14:43.00 that it moves faster or that it degrades less, would make a Calliphora 00:14:47.10 sized gradient in a Drosophila egg and you can imagine 00:14:52.07 that would result in a catastrophe for development 00:14:56.16 for the Drosophila embryo that was depending on that gradient. 00:14:59.09 But what you actually see, the amazing thing is that these bicoid gradients 00:15:03.12 that are established in these eggs, actually using the Calliphora protein 00:15:09.20 are identical, surprisingly identical not to the gradients 00:15:16.15 that those same proteins would have made in Calliphora 00:15:18.13 but to the gradient that is made in Drosophila melanogaster. 00:15:22.27 So, for example in this figure here we can see the Calliphora bicoid 00:15:28.25 extending out in a visible sense to about 48% where we'd be activating 00:15:33.27 hunchback, and that's very similar to the distribution that you'd see 00:15:40.02 when you took the Drosophila EGFP and put it in the same egg. 00:15:45.28 And you can measure that and show that these distributions overlapped 00:15:50.10 each bicoid that regardless of the source of the bicoid protein put into a 00:15:58.24 Drosophila melanogaster egg that protein will make a gradient 00:16:03.06 of the same shape, the same size, as the melanogaster bicoid. 00:16:08.20 So we know that what that's telling us is that the fact that Dipteran bicoids 00:16:18.07 expressed in Drosophila make Drosophila sized gradients is that during 00:16:24.23 the course of evolution it's not bicoid that has changed to allow for the 00:16:31.23 adjustment of the egg shape and the egg size, but some other property 00:16:36.29 of the egg. So these proteins put in a melanogaster egg will produce 00:16:44.10 melanogaster type gradients. Now we've done the same kinds of experiments 00:16:51.02 using other tagged proteins, EGFP, we've altered the bicoid area, 00:16:56.00 just put straight, if you can imagine just taking EGFP or EGFP 00:16:59.18 with an NLS, a nuclear localization sequence, and localizing the RNA 00:17:05.10 to the anterior end of the egg and ask will any protein, 00:17:08.10 any GFP-tagged protein put into a Drosophila egg make a gradient 00:17:13.04 the shape of bicoid. And what we found is that's not true. 00:17:15.17 That each individual protein put into the egg makes a gradient of a particular 00:17:21.21 size and a particular distribution. But all of the bicoid proteins 00:17:25.17 put into the Drosophila egg at the anterior end make gradients 00:17:30.17 of a particular size. So the interesting conclusion then from these 00:17:35.03 experiments is that during the course of evolution, it's not that bicoid 00:17:42.17 has changed to adjust for the size of the eggs, but actually 00:17:46.23 bicoid has been conserved. What's been conserved in bicoid is 00:17:51.00 the property of the protein that allows it to build gradients of a 00:17:55.28 particular size when put into Drosophila eggs. And all the bicoid molecules, 00:18:04.06 but other proteins do not have that property. 00:18:09.04 What's actually diverged during evolution has been not the protein itself 00:18:18.27 but the environment that we put the protein in 00:18:24.10 and then that raises the kinds of questions that you'd like to answer now. 00:18:28.15 What are those properties? What could influence bicoid? What's the 00:18:37.19 features of bicoid, the bicoid protein itself, that allows it 00:18:42.11 to respond and make gradients and there are obvious experiments 00:18:45.09 that we're in the process of doing where you can identify the regions 00:18:50.28 of the bicoid protein that are essential for it to make 00:18:55.25 a gradient of a particular size and shape that's characteristic of bicoid. 00:19:00.24 And then the other question is what are the features that change 00:19:07.02 as you change egg size that change those distributions. 00:19:11.26 How is it that you are able to maintain the property of the protein 00:19:15.26 on the one hand and then change the size of the egg 00:19:18.14 and change the movement of the proteins. So those are 00:19:24.28 really essential questions for understanding how during the course 00:19:34.07 of evolution you are able to use the same system, the same protein, 00:19:39.23 over and over again. They will require different kinds of measurements 00:19:47.21 and being able to work with multiple species and multiple variants 00:19:52.13 both of being able to put bicoid proteins into melanogaster 00:19:57.25 but able to also put variant proteins into the bigger and smaller eggs. 00:20:01.26 But what they'll also require is to distinguish between these different models 00:20:06.20 to give it the underlying mechanisms that are controlling the distribution 00:20:11.16 is again the kind of quantitation that we talked about in the second lecture 00:20:16.22 and it's my own belief that the future of developmental biology 00:20:21.16 in general and the refined understanding of problems 00:20:28.23 in development will depend heavily on our ability to combine 00:20:34.01 both those quantitative visual techniques for analyzing distributions 00:20:40.21 of molecules with the powerful techniques of molecular biology 00:20:45.00 that allow you to manipulate the sequences and structures 00:20:50.03 of those proteins and also the other features of the egg. 00:20:56.12 So I'll stop there and thank you for your attendance.