 Okay, well as Elise just said, we are not a modern code lab and I'm going to be telling you about our research in a general sense and then getting a little bit into the modern code work. So the question that is of interest to us in a broad sense is how to form organs and how naive populations of cells come together to make a particular structure with defined cell types and very specific kinds of architecture. And the question that I want to talk about today is illustrated here and that is from this classic picture by Waddington to illustrate the kinds of choices that a cell has over developmental time. And so he drew it here where you have a cell that starts its life with many potential outcomes and then through a series of decisions which he suggested would be debinary decisions but just to get the general point that cells ultimately commit to a particular cell fate. And so it's this early, what I want to discuss today is this very early stage where cells have pluripotency or a kind of developmental plasticity and they have multiple choices on what they can become and then that loss of plasticity over developmental time. Now we study this question in C. elegans shown here. We're particularly focused on how the gut is made and I have to say that when I started talking about plasticity in C. elegans people thought I was crazy I think and in part because most people were seeing C. elegans more like this where a cell might have one particular outcome it might not like that outcome but that's what it was stuck with. And of course this perspective derives from the lineage that we just heard about from Bob Waterston and that is this stereotyped cell division pattern that worms have. What we know now is that although C. elegans undergoes this stereotype kind of division in fact that is not reflective of the plasticity that embryos and cells have underneath. And we know that from a number of experiments and I'm going to show you this classic experiment by Jim Priest that I really like and this was what Jim did was he was noticing that these two lineages look very different. They give rise to very different patterns of cells and you can tell this just from the cell division pattern and yet these two cells are sister cells at the four cell stage. So that's ABA and ABP. So what Jim did was the following experiment. He was watching an embryo divide going from the two cell stage into the four cell stage and what he did was to use a blunt needle to rearrange the positions of ABA and ABP. So as the mother cell divides into the two he's pushing ABA back and so now you have ABP in the front. And what this showed is that you could get a completely normal worm out and what this tells us is though the lineages so reproducible that in fact you can get very different kinds of fates out from different cells so they have a lot of plasticity in them we just don't normally see it. So this kind of analysis and other studies that I'm not showing you has told us that the C. elegans embryo and the initial divisions up to about the 28 cell stage or little past have a lot of plasticity and that these cells are capable of adopting very different fates and then what happens over time is that that plasticity is lost and this happens with the onset of gastrulation which is actually very similar to what's seen in vertebrates where it's these early blastomeres that are pluripotent and then with gastrulation cells begin to lose that kind of flexibility. Okay now that's the situation in a worm and of course often when we think about plasticity we think about it in this configuration which is an embryonic stem cell so I just thought it would be good to point out a few differences and similarities. Both have a lot of plasticity the stem cell is sitting in a tissue culture dish of course stem cell ES cells don't exist in the embryo and they can undergo cell renewal and then differentiation of course is controlled by the investigator and that's very different when you're looking at a real embryo where on the one hand we're getting these very rapid changes in cell expression and in fates so there is no sort of stable ES identity in the embryo and then the plasticity that we see is happening despite these changes is happening as I mentioned at this very defined time at the onset of gastrulation. There are differences between the two certainly and for example in ES cells we have these factors that are so critical for establishing pluripotency there's nothing to suggest that worms have anything like this gang of four that have been identified but they also do have similarities for example the repressive transcription complex polycomb. So one aspect that really intrigued us about these early cells in C. elegans has to do with their nuclear organization so these top two panels are C. elegans embryos at different stages of development and so I've ringed and read the nuclei so this is a very early embryo this is kind of a mid-stage embryo and you can see that at the EM level they have they look really very different and in particular the early embryo has very little electron-dense material it looks like it doesn't probably have much heterochromatin and then if you look later there's this clear change in how the nucleus is organized. Now this is not just a feature of worms and so here's some other examples down below this is an ES cell that's the nucleus there but you can see again that the nucleus has this big open look to it and that with differentiation is when you start to pick up this electron dense material and you get a lot more heterogeneity in general and how the nucleus looks and then this is a planarian stem cell an adult stem cell and again you have this big blast cell and it lacks anything that really looks like heterochromatin as you see here. So to begin to look at this kind of nuclear architecture we wanted to develop a way you can't really study things by EM. You can't do a screen by EM. So we've been devising different ways of looking at this and this is the work of a graduate student in the lab Talafakuri and a postdoc Tanya Yuziak and what they did was to take advantage of C. Elegans how it handles DNA. So worms are holocentric as you probably know and if you introduce sequences DNA sequences into the worm it will basically build a big artificial chromosome kind of a pseudo chromosome and so what we and others have done is to adapt the Andrew Belmont system where you can tag the lac repressor and then have it bind to this artificial chromosome and that gives us a very nice way of then being able to visualize this chromosome in vivo and what we like about this assay is it gives a single cell resolution so we know exactly what stage and what cell type we're looking at and we can see the kind of heterogeneity of what's going on and what was immediately striking to us when we did this is that we got this out so what you're looking at here is a panel of cells or nuclei that's what these little purple dots are and black dots and then all these are different artificial chromosomes in these nuclei and what really struck us is how different the morphologies were and just like what we saw in the EM studies we could see a big distinction that happened over developmental time so here we're looking at a single nucleus it's in an early embryo this is this tagged artificial chromosome or extra chromosomal array and what it looked like is it didn't really matter what promoter we put in so these are different lines that we built they all had this kind of very sort of fluffy appearance which we called a floret and by and large the interphase nuclei all had this same behavior however it doesn't last so here we're looking at different cell stages and this is happening quite rapidly so this is 28 to sort of 44 to 100 cell stage and we lose this configuration this big open chromatin look and instead it becomes much more condensed and and these very compacted artificial chromosomes now begin to pick up marks of heterochromatin for example so that was encouraging to us it's it suggested that we could actually recapitulate the kinds of things that have been seen at the EM level and we could see this closing down of the genome so what we think is going on is that there is a global closing down of the genome irrespective these aren't transcribed genes these are just sort of genes in the genome what happens then over time is that kind of opposing this this trend are the so-called pioneer transcription factors or the selector genes and sorry here's the quantitation so this is just to show you that the proportion of nuclei that have one or the other kind of compacted chromatin is increasing over time and so you can see that here so again these are individual lines and just quantifying the number of interface nuclei that have this kind of structure or this kind of structure okay so what we decided to do was to begin oh did this disappear I disappear so what we're going to do is look at one of the selector genes and this is far for so this was the factor that Bob just mentioned it's critical to form the digestive tract in particular the foregut in the worm and it's orthologous to Fox a found in from everything from Hydra to vertebrates and we know a lot about far for and so here is my first modern code slide so I put it up this is was done from Mike Snyder's group and it's looking at the chip data of where far for goes in the genome and then we've done a lot of expression studies and mutagenesis studies and the bottom line is that far for is all over the genome it's in thousands of sites and it regulates essentially everything that selectively expressed in the pharynx so if you make mutations in promoters of those far for sites you will lose or dramatically delay expression in the foregut so it has a very global role now this global role makes sense for a sort of selector gene pioneer factor but it also raised a lot of questions for us and that was that we began wondering how you get specificity if you have this transcription factor and it's going all over the genome and binding to all these genes and critical for their expression and so one way is certainly combinatorial control and I'm not going to talk about that but the other way where far for has a big impact on expression is the affinity for its DNA binding site and what we were able to show and I'm not going to illustrate it since it was published a while ago is that far for has different affinity binding to its DNA target and that higher affinity would promote earlier expression and lower affinity would promote later expression and so you could actually get an input of temporal regulation based just on far for and it's and it's DNA binding affinity and so we've looked at that also by the same spot assay now modified to include both far for and far for promoter to look at the association and the effect of far for on its targets and what we saw was really striking so here what we're looking at these are nascent foregot cells and the purple is the far for Fox a transcription factor and then the green is the lack compressor that just is to let us see this array so in this top panel we're looking at three foregot cells and you can see that far for is clearly binding to the array whereas if we mutate far for binding site in a target promoter we still have lots of far for but it doesn't become enriched on the array as you can see here compared to this so it gives us again a very nice assay to look at the single cell level and know when and where far for is binding what really jumped out at us there were a couple things that really were very striking to us and one is that far for can clearly decompact and open up the chromatin compared to when far for isn't there so if you if you look at this dot it looks huge for example and whereas this looks relatively compacted down and we've actually quantified that by just measuring either the volume or the or the area of these regions and asking cells that have far for how do they compare to cells that don't have far for and what happens to these target regions and the data shown here so green are cells that lack far for their outside of the foregut or the pharynx and that's what these green lines are and the pink are pharyngeal cells that have far for and far for is binding and so you can see what it looks like up here so this is a section where you have far for and it's binding and we're looking at different stages over embryo genesis and then out here are cells that lack far for and we're doing this same kind of plotting the area and the proportion of nuclei and what you can see is that the pink line which is to say the cells that have far for have bigger areas than those that lack it and that we take as evidence that far for is leading to this opening of the chromatin okay so here is the normal that I just showed you and you can see it opening up for example here here we've made a mutation where far for still binds but we know we inactivate the ability of D we get a very dramatic opening that's again this pink line compared to the green line on the other hand if we knock out the if we inactivate the far for site now we get rid of that opening so it needs far for and this is a different gene where we can remove the far for site and yet still get transcription and again you can see that the opening is very much reduced so we think there's something special about far for and its interaction with DNA that it can counteract the closing down that I just showed you okay now far for is not static and I kind of implied that by showing you these different stages in fact what far for does is dramatically shown here so we're starting at early embryo genesis and then continuing to the time of hatching and we've quantified the amount of far for within nuclei and you can see this this huge increase over time and so the readout of that then has this implication for the timing of expression and the affinity of binding for DNA and one of the ways that we know that is from actually measuring the association with far for and this decompaction that I illustrated so what we're going to do in this assay is shown here we have a gene that we know is a far for target we also have made a mutation where we lower the affinity of far for or the far for binding site within this promoter and we know that that will delay the onset of expression and so these are two different promoters and we're going to analyze them in two different strains of worms and look at the association of far for with these targets and what we get are these kinds of graphs where we're looking at the proportion of far for that's associated with the target chromosome and we're also looking at the effect that it has on the morphology of that artificial array and the data shown here so this is the earliest stage that we can pick up far for binding and we're looking at either this normal promoter or this mutated promoter already you can see them begin to shift and then within a little while you can see that the wild type you're getting more far for bound compared to this down mutation and it's having also an effect on decompaction on the other hand if you look out here where the gene is actually transcribed they look very similar so you're seeing the same proportion of far for bound you're seeing very similar effects on the chromatin configuration and so the way that we interpret this is that first of all far for combined to these arrays very early in embryo genesis hours before the gene will actually be transcribed and that the effects of the affinity seem to kick in also very early and that at these later stages even when transcription is actually happening that at that point affinity is no longer a player so let me just sum this up for you so what I've told you is that the early developmentally plastic state in the early embryo is accompanied with a very open kind of chromatin configuration and that over time as you lose plasticity that that's accompanied with a compacting down of chromatin and that kind of counteracting that we can watch selector gene like far for which kicks in at this stage and watch it decompact chromatin and what I didn't have time to show you is that polycomb is actually part of this closing down process so this kind of model is interesting but it raises a number of questions and one question of course is that what I've shown you today is mostly on arrays and those are artificial constructs and so one question is what's really happening in the genome and then the other question is really when I say open versus compacted what do we actually mean and you could imagine a number of different ways that the genome could be opened on the one hand there's the nucleosomes interacting with DNA and there's certainly precedent for losing nucleosomes nucleosomes can exist in this kind of beads on a string configuration or they can begin to count compact down its thought into the 30 nanometer fiber so this is another way that you can compact down DNA and then of course these fibers can begin to loop and depending on how you loop DNA that could also have a big effect on the morphology and at the level that we're looking we can't really distinguish what we decided to do was to focus on this earliest point and to think about nucleosomes interacting with DNA and in part we were thinking about that from a work from Misteli's lab looking at the way chromatin proteins interact with DNA in pluripotent cells in this case ES cells and they were arguing that the interaction of these proteins with DNA is hyperdynamic that there's just a lot of motion so we thought we would ask what worms do and when I say we it's May Chen who's a really terrific postdoc who started recently in the lab and Kareem Carr who's a mathematician and you see him here with the biggest gummy worm ever okay so the approach that we decided to take was actually developed by Jason Leeb sitting here and that is fair and for those of you who aren't familiar with fair it the idea is that just as in chip you crosslink proteins to DNA and you sonicate and bust up your your chromatin but then there is no IP involved instead what you do is just remove any DNA that's bound up with proteins in a phenol extraction and you have this released fragments and that's what you analyze and so what you obtain are these sequences of DNA that are open in that they are not bound to nucleosomes and so what we did was more or less adapt this fair protocol to look at C elegans and we decided to do this very carefully staged where we're looking at this early cohort that we know have plasticity and then comparing it to kind of mid-stage embryos that we know have lost plasticity so the first thing that really jumped out at us is shown here so this is from May's preps and it's just three experiments to show you it's reproducible but it was very noticeable that the amount of chromatin we obtained from the early samples was much higher than what we get from the late samples and typically we were getting somewhere close to four times more DNA from these early samples and so here we're just aligning all the chromosomes and looking at our early fair and our late fair against genes and you can see that in fact there is a lot more chromatin in this early compared to the late what that suggests to us is that in fact by this biochemical method we are getting evidence for a more open more open configuration where it's actually the association of the DNA with the nucleosomes and that that changes by mid-embry genesis as shown by the late fair if we look at this a little more closely one thing that jumped out at us was looking at our data as compared again this is more modern code data from Jason's lab and here he's looking at the association of parts of the chromosome with the nuclear lamina with this protein lem2 so lem2 chip you can think of it as as a way of seeing which DNA is sitting at the lamina and what's very clear is if you look at a particular chromosome the middle of the chromosome is not attached to the lamina but the arms or sides of the chromosome are and that tracks extremely well with what we see in the late fair so you can see this is more open and this is less which works quite well with this lem2 region the early has a little bit of this but not nearly to the same degree so in general we just get so many counts it generally seems open regardless of what's going on at the lamina we can zoom in and so here I think we're looking at part of chromosome 5 as I recall and there are what have been called sub-domains for lem2 so there are regions for example here where you pick up lent to and then regions that have less of it and regions with more and these are interspersed and if we look again at the late fair we we see a fairly nice correspondence where it's these lamina free regions have a pretty nice open region and then the regions that have a lot of attachment to the lamina are are lacking late fair and again though if we just look at this level at the early fair we don't see that kind of what's the word we don't we don't see anything like that it just we generally see this opening so here there's a really a very dramatic dramatic difference in the way that the early fair behaves with the late and that fits actually quite nicely with those EM data I showed you where you could see these attachment points to the to the nuclear lamina okay now there were in addition to this so this is where we see perhaps the most dramatic effect early versus late there were also a number of other places that we expected to get nice fair signals and I'm just going to show you a couple of those one is locations that have histone variants and what Gary Felsenfeld showed a number of years ago is that the histone variants H2A Z when combined with another variant histone 3.3 is a very labile nucleosomal configuration it comes off of DNA very easily and that had been seen in a lot of biochemical experiments as regions that lack DNA so sorry regions that lack nucleosomes and we thought we should actually be able to see this kind of configuration and that is indeed the case so here is more modern code data marking regions that have the histone variant H2A Z in three locations and here is early and late fair data and you can see it it lines up very nicely with these two H2A Z regions but on the other hand it doesn't really seem particularly enriched on the third so it's not a one-to-one it works quite well with transcribed regions uh transcriptionally active regions we also see it in some of these hot regions that you heard mentioned previously so these are demarked again on the modern code website as as hot and you can see we get these whopping peaks here of what seem to be nucleosomal free in in hot regions and then we also get mystery peaks and so this is just an example it's not really sitting near a gene it's not sitting near H2A Z but we're getting really quite strong peaks in certain places and so we're intrigued by by what those actually are so i'm just going to stop there and say what i've told you today is that i think something as simple as a worm is a good way to study developmental plasticity and that one area that we're particularly interested in is the configuration of the nucleus and how it changes over time and we've looking at this both at the cell level where we can do these single cell analyses but then complementing that with this kind of fair analysis and we're really just starting it but i wanted to mention it because the modern code has been really terrific for us as a way of interpreting our data as we begin to compare what we have found and what's up there on the modern code site and and i actually think in talking to other people in the worm community that there are a number of us in this situation we're beginning to look to the modern code as giving us tools that we can use for comparisons as a way of interpreting our data it's not really at the point i think there's maybe one paper that uses modern code data that isn't from a modern code group but i think it'll become more and more useful to see elegans researchers as we as we avail ourselves of that data and then i'll just wrap up by acknowledging tala and tanya who did the initial studies and that's been followed up now by kareem as i mentioned and may and i would like to acknowledge the modern code for their work in particular mike snider who did a lot of far for chip which was great with valerie rynke who is also involved in the transcription factor analysis and jason leeb many of those modern code data spreads that you saw were from the leeb lab they've been really terrific and jason's also given us great advice and this actually reminds me i think i'll put it up as a plug for anyone interested in transcription and development that in a few weeks there'll be a meeting out at facem and we probably still have slots so i thought i would show it to you guys and i'll stop and take any questions thanks i really like this compaction idea i think this is fabulous have you looked in adult self-renewing tissues versus normal tissues in other words is this a would this be a way that would distinguish those two types and i don't know in a worm if they have self-renewing tissues but i guess at least in the germline yeah the only renewing tissue in the worm is the germline and the germline is really very different so you know and we haven't looked at it it would be interesting i think one big difference between the germline and the early embryo the early embryo is really gearing up its whole transcriptional profile is about to just burst forth on the other hand the germline is sort of transcriptionally closing itself down and relies a lot on post-transcriptional mechanisms there's a lot of translational control but i think it would be really interesting because that would help you distinguish it as you say is if this is just a sort of when you need the entire embryo to go off versus individual cells and individual tissues that would be a good way to distinguish that yeah so what i didn't show you and what really got us thinking about this in the first place is there are experiments where people have mis-expressed selector genes throughout the embryo throughout the soma of the embryo and if you do that you can get really dramatic cell fate changes so you can change cells that should be neurons into intestine or body well muscle and what was really striking to us is that the embryos up until the 28 cell stage respond very nicely and they and they respond fairly homogeneously so um i don't have it here so essentially all the selector genes that have been tried were beautifully up until the 28 cell stage and then they begin to falter you know 100 cell stage and on and it was that homogenous response that made us think perhaps there's something changing in those recipient cells and got us thinking about these other experiments so so i think it is kind of homogeneous one thing we do know is for example if we get rid of polycomb we can both change chromatin structure and prolong plasticity they track together but really testing it is in progress so that's yeah uh Susan i have a rather naive question so sorry i don't know i have a rather naive question if as you as you change these you showed this experiment at the very beginning which is very elegant of sort of moving one cell around and get beautiful you know working animal my question is is the gene expression i mean has anybody followed the gene expression patterns to actually understand if the new cell is simply you know leading to the whole new lineage of the cell that was occupying the previous i'm sorry are you saying is the is the is it location the misplaced cell yeah taking on the new expression exactly um it hasn't been done probably to the way you would like it but i think it's fairly strongly implicated put it this way i would bet my one and only beloved child which switch and that's saying something and the reason i say that is for example a b a that front cell makes half of the foregut and that has very distinctive expression patterns and when you flip the but the back cell does not and if you flip the positions of those two cells now you're getting the foregut from that front cell and you're getting on the order of 40 cells from it and that back cell nothing so to have that kind of dramatic change you really need a very gross expression pattern change and and that difference is mediated by notch signaling and then and that's not the only one there are the somatic gonad is very different between the two so there are a number of cells where it would be where to to get those different types of cells out you have to have really dramatic and it is an experiment where you would take a cell from a different embryo and sort of include is that possible to actually get a different genetic background for different so what people have done is they've gotten rid of the eggshell and then kind of mixed and matched blastomeres together and but they they don't you can't make a worm at that point you can see evidence of cell interactions you can particularly see effects of wind signaling on spindle alignment but you what you cannot do is make a chimera take a bunch of different cells and have a chimeric worm come out kind of thing and it's I suspect partly of partly that those embryos those cells aren't as responsive of say a vertebrate embryo and partly it's things like the fact that the egg shell is important for orienting cells but what is really dramatic is this heat shock response where you can mis-express you know a number of these different selector genes and see this complete conversion where you you turn off the endogenous cell expression pattern and you turn on this alternative fate of whatever it is and it and and one thing that's kind of interesting you may be familiar with work of Sophie Giarriot for example or Oliver Hobart and what they've shown is that later in development like in larvae you can convert the fate of cells that are fairly close so for example you can have one neuron cell type become another neuron cell type by mis-expressing um what Oliver would call a terminal selector gene but although you can get those closely related things to switch you cannot get that cell to suddenly a neuron to become a gut cell that seems to be really very specific for this phase of early embryogenesis and it disappears um Susan um so this is great i'm wondering about um when you use the array and you're looking at opening you get the florets um do you have any sense since it's repeated whether there's variability within the chromosome in in how much opening and closing in fact if you look at a floret structure it sort of looks like it's not uniform so i don't know if you've looked at that at all in more detail we haven't looked at it in gory detail but i think you're right so one thing is that when you make um these artificial chromosomes and worms you get different degrees of repetitive dna depending on you know how much of everything you put in and what was shown a long time ago um is that repetitive dna and worms like in other organisms leads to silencing so if you just put in a repetitive array if you build a you would not see this at all i don't think i mean we haven't done it but that would be my hunch it would be much more closed down and it would just stay repressed what we're trying to do is build complex arrays that look more like a real chromosome um and but because of that you certainly see variability whether there's variability in the pseudo chromosome you'd think there would be and actually i was thinking wouldn't it be fun if one could do a like a nano resolution project and and see what's going on there and then i thought it's a worm array like who cares about a worm array so that's when we thought okay we really have to do the biochemistry and prove to people that it happens on real chromosomal dna it's not just some array strangeness and so that's when we started doing all this so the the e-cell becomes committed to a tissue well before any of the other ones and i was wondering if you saw any difference in the in the pattern of those an earlier transition or anything like that um yeah the e-cell is interesting so by the lineage just the wild type lineage the e becomes dedicated to the gut at the eight cell stage and yet at the 28 cell stage you can convert it into skin or you can convert it you know into muscle so even though by the lineage it's only going to produce one cell type already at that early stage you can go later and change it so in other experiments i didn't talk about we actually tried an experiment where we removed pha4 so the selector gene for the foregut and asked is that in fact important for curtailing plasticity so is it really the cell is there a big link between plasticity and cell fate and the surprising answer it was totally not what i expected was no so when we removed pha4 and the foregut cells don't know what to become they still terminated plasticity by that heat shock assay at the same time it wasn't delayed at all and so that's kind of the way i think about it and may i'd be interested to know what what you know if it fits with these networks people were talking about that those very early steps when you establish cell fates whether it's intestinal or foregut or what have you you still are pretty loose and if something else comes in you can switch your fate and it's only as that gets more established you get these feed forward circuits and you really finalize a fate that maybe at that point the cell fate regulators match but at this matter but at these early stages there's still a lot of fluidity and the e-cell it's very happy becoming muscle and it can become skin yeah it's kind of shocking okay thank you