Scientist Stories: Terry Orr-Weaver, Cell Division and Chromosome Segregation

Okay, this is the largest, this is the most packed I've ever seen this room, so I think we better get started before, nope, yeah. No, I think you should turn it off. Can you hear me? No. Can you hear me? You can hear me anyways. Thank you all for coming today. It's really great to see this packed turnout. I think we've had this room full a lot of times. This is the fullest I've ever seen this room, so thank you all for coming today. So, in particular, so this is now our keynote talk in our first ever Whitehead Seminar Series this year. It's been a great series of talks, been really nice turnout, and I'm particularly excited that we can finish with Terry Orr-Weaver today. So, I've been looking forward to this talk for a while, and I'm excited to hear this. I feel kind of nervous giving this introduction, but I'm going to do my best because I think that there's some things that I really want to say and highlight about Terry and about her contributions, about who she is. First, I'm going to give you the kind of introduction that we would always give when somebody comes through here. Teri was a graduate student with Jack Szostak at Harvard Medical School. There she did really important work and foundational work on understanding the basis for recombination and using sort of a double-strand break model for how recombination works. She did her postdoctoral fellow with Alan Spradling at the Carnegie Institute in Washington. And there she transitioned to working in Drosophila where her lab has continued. And then she started here at Whitehead as an assistant professor in 1987, and then she's sort of grown up through the ranks since then. Along the way, she's made a lot of accomplishments. There's 3 things I really want to highlight about Terri today. First, her contributions to science, and then her really strong contributions as a mentor, and then also her contributions as an educator.. And each of these I've really loved being able to see and benefit from myself. So first on the science side, as you'll hear today, Teri has been really a strong leader and advocate for basic and foundational science, fundamental science. And she's done that through her work on chromosome segregation, on understanding the oocyte to embryo transition, and on understanding the basis for DNA replication and its control, and particularly how that is controlled across a multicellular organism. As part of this work, so she's been honored in a lot of different ways. She's an American Cancer Society professor. She was a past president of the Genetic Society of America. And she was elected to the National Academy of Sciences in 2006. And really she's been just this fierce advocate for good science, high-quality science, particularly model organisms and genetic approaches for that. The work from her lab has identified a lot of things and also understood them, and things that have really affected the way that our lab also cares about cell division. So her lab was the first to identify a protein called MYOS332, which plays a really important role in controlling how meiosis occurs in the pengu and Plutonium genes, in a DNA replication factor called ORC. She's made a lot of fundamental discoveries in the key players that drive forward these processes and then also how they work.

Okay.

Not only has she really led, I think, in terms of understanding the fundamental science, but to me she's really been sort of a consummate mentor and role model. So I've been here about 10 years, and I don't think I would exist if it hadn't been for everything that Terri has done along the way. So I think she, at least for the first 7 years of my— of our lab here, read every single paper, every single grant that we had. And she's the person who taught me how to teach and how to be an educator. She's the person— so when my first daughter was born, she's the person I went to talk to about things like that. And so she's been the person along the way that I have at every— whatever it be, big or small, I seek her out for advice. And she's been that way for me, but I think she's also been that for essentially every single member of our community, both in this building and throughout this department. So she actually was awarded a Whitehead Postdoc Mentoring Award. I actually think it's the only time we've given that out maybe. But I think they had to come up with it just because of the strong respect and strong appreciation that everybody has had for really who Terry is as a very thoughtful, compassionate, and giving mentor to those both in her lab and people like me who just seek her out for advice. And then finally I'd say I've had a huge pleasure of being able to teach with her for the past 8 or 9 years, and just watching her devotion to education and training, it was just inspiring to see what she did for the undergrad cell biology class that we taught together and the other classes that she's had. And then I think I'm going to mention one more accomplishment amongst us before I let Terri begin, which is we have two excellent examples of her success and productivity in the front row. Matthew and Nathaniel are here today, Terri's sons. Actually, Matthew graduated with his PhD last week in physics. So they're both here to also join Teri, and things that she should be very proud of as well. So it's really a pleasure to have Teri present this keynote. I can't think of anybody more fitting for finishing off this great series, and she's going to talk about research taking flight from foundational biology.

Thanks, Ian, for the wonderful introduction and for organizing this whole thing. It's— thank you all for coming. It's just incredibly fantastic to see you, to see former members of the lab here, to see people who came from across the river.— and especially to thank Matthew and Nathaniel and Kelsey for coming. Matthew flew in on a red-eye flight this morning. So it's great to have you here. Ian told me I could talk about anything I wanted to talk about. So what I decided was rather than giving an in-depth detailed science talk, that instead Instead, what I would try to do is highlight some of the work from the lab over the years. But I thought what might be most useful to the students and the postdocs is, since there are so many different cultures and so many different styles of running labs and doing science, what I thought might be actually useful was to try to use some of these examples of projects from the lab to give you a sense of how did we choose what we were going to work on, how did we approach the problem, and, you know, the whole time the lab was doing research, I also myself was developing as a researcher. So to give you kind of a sense of how I grew and what I learned from these projects. So let me begin then, and I think the major message I want to leave you with is to choose a problem in biology as the foundation for your research. Make sure that's one that you're excited about. And the most important message for you to get from this talk today is that it's still the case, it still remains the case that doing basic foundational research, figuring out how fundamental problems in biology work, is a great way to do research. It's a way to really generate unexpected new insights, and things are so conserved that it's certainly the case that those insights and knowledge that you gain from basing your research in fundamental biology ultimately will have important applications. So in saying— in recommending this approach, of course, one has to be pragmatic, and I think it's important also to choose a a system that has— that's ideally suited to answer the question that you're interested in, that has the right biology, and that also has exploitable experimental advantages. So most of the work in my lab has worked on this critter, the fruit fly, and you'll see why that has powerful experimental advantages and why the biology is fascinating and there are fascinating biological problems. The other thing I want to note is that Drosophila, to me, really stands out as a research community. This is an extraordinary community of researchers that really were committed to working together as a group, and there were a number of key selfless leaders in the field who devoted tremendous time and generosity to generate resources and toolkits for everyone in the field to use, and you'll get a sense of that also. So what have we worked on? When I was starting my lab, it was in 1987. It was kind of a pivotal moment in two areas of biology. There were explosions happening both in the developmental biology field and the cell cycle field. So what was happening in the developmental biology field is this collection of mutants that identified key regulatory genes for development had been identified in Drosophila, and these ultimately ended up being many of the critical signaling pathways you know about that are important in mammalian development, like hedgehog signaling, notch signaling, like the HOX genes. And so this collection of mutants was available in Drosophila. The community was working their way through identifying the proteins identified— encoded by genes affected by these mutations. And by looking at where those genes acted, where those genes were expressed, like in this example of an expression of some key regulatory— regulators in the embryo, we were really getting a sense of how pattern gets regulated during development. Now, while that was happening, there was a pivotal explosions happening in the cell cycle field. And this came about because genetic approaches in yeast synergized with biochemical approaches in Xenopus and some of the marine invertebrates to begin to identify the key regulators that control the cell division cycle and to really establish this conservation. And so among them were a family of regulators, the cyclin-dependent kinases. The first members of those families identified were the cyclin B-CDK1 complex, which is conserved from yeast to human and is critical for driving the onset of mitosis in a cell cycle in yeast from G2 into M. I'll talk about these embryonic cycles where it also is involved in starting mitosis. So when I was thinking about what would be the theme and the focus of my independent research program, what I realized is it could be very powerful to work at the interface of these two fields, to really combine them, because it was going to be important not only to understand how cells knew what they should become and where they should become that, but also how was— how did development interface with cell cycle control to make sure that the right number of cells were formed at the right place. So I decided it would be good to think about working at the interface of these fields. And this is what I actually looked like when I started my lab. And so we decided— What I wanted to do was I wanted to run a style of lab where everybody in the lab would have an independent project that they would have ownership over and responsibility for. But the idea was to put all of those under this umbrella of how does development— how is cell division coordinated with development. Now, it was really clear that Drosophila was going to be great for this. Of course there was the genetics, there was this very supportive community. But the other reason it was clear is that it was— it was clear that in addition to regulation of the normal cell division cycle, right, which you're familiar with, where there's a discrete period of DNA synthesis with segregation of the replicated chromosomes punctuated by these gap phases when gene expression and growth occurred. That in addition to this canonical cycle, that there were likely to be variations in the cell cycle that achieved particular developmental strategies. And sort of the first one recognized was this cycle in early embryos where any organism that has— any animal that has a reproductive strategy where the mother lays the embryos and then leaves them, needs to undergo very rapid embryogenesis to get to a motile stage. And so what those organisms do is to use an accelerated, stripped-down embryonic cell cycle where there's a rapid DNA synthesis phase that oscillates with mitosis but without these growth gene expression phases. And this is made possible by the mother depositing everything into the embryo that's going to be needed for the early divisions. So this embryonic cell cycle was there. And in addition, there was— there was the possibility of another variant cell cycle because in Drosophila, most of the differentiated cell types actually didn't have a diploid DNA content. They had many copies of the DNA. And so somehow these cells had to be able to increase their DNA. So what we— I decided to do was put my first two students on these. The one worked on understanding the S/M cycle. And actually the first person in my lab was my first student, Albert Smith. He decided to try to figure out how cells could increase their ploidy during development. And so he's— was quite brave. He actually unpacked the boxes when we started the lab. Okay. So what we ended up doing was really working in 3 research areas. Ian already mentioned this. A bulk of the lab worked on how the oocyte becomes an embryo. We also had a number of projects looking at how during development DNA replication could be altered as a developmental strategy. And then there was a major effort in the lab to understand how progression through meiosis was regulated, how that's coupled with development, and also the mechanism of chromosome segregation. So what I'm going to tell you about today is a few highlights from work on the oocyte-to-embryo transition and the regulation of DNA replication. So let's begin with vignette 1 about the oocyte-to-embryo transition. That's what I view as the beginning of development. And it's this really profound change where oocyte undergoes a plethora of changes to become an embryo. Those are cell cycle changes. The mature oocyte is arrested in meiosis. Specialized cell division. And so this transition involves completion of meiosis and changing so that it can switch to mitotic divisions during embryogenesis. It involves this change in potency. This is a differentiated cell and it has to change to become totipotent during embryogenesis. Now what's kind of amazing about all these changes is that in nearly every animal, this hap— and that includes mammalian embryos— this happens in the absence of any transcription. So that means that all of these changes have to be controlled post-transcriptionally. They all have to be controlled by changes in mRNA translation and changes in protein stability. This is a great problem to look at in Drosophila. Because of the genetics but also because of how development works. So all of these stages of oogenesis with— that mark progression through meiosis are all morphologically distinct. And so that means that you can actually isolate sufficient quantities of each stage of meiosis to do biochemistry. The other thing that's really great is that the release from this arrest and this process of the oocyte to embryo transition happens as the oocyte moves through the oviduct and into the uterus. And that is accompanied by mechanical stress and by hydration. And so the transition from oocyte to embryo doesn't actually require fertilization. All those changes in mRNA translation, protein stability, expression of developmental regulatory genes don't require anything from the sperm. The only thing the sperm's doing is giving the cytoskeletal apparatus to allow the first mitotic division. And so what that means is we can do this all in vitro, and we don't have to be confounded by any regulatory impact from the zygote's own genome. So the way we began was to do something very simple, to just ask, can we just do a visual screen through mutant collections and ask if we can identify mutants by sight, by staining for DNA, that have a defect in the oocyte-to-embryo transition. And again, the community made this possible for us because there were— 3 labs that had generated female sterile maternal effect lethal collections, and they gave them all to us for us to screen. So Trudy Schupbach, who was at Princeton, Janine Islen-Wohlhardt, who together with Eric Wieschhaus had generated a collection, Dawson Muller had generated a collection, and later, you know, some years later in the lab, we redid this on a much larger scale, and doing that we used a collection from Charles Zucherkowski. In Peter's lab. So we went through and looked, and this particular phenotype understandably really jumped out at us. So what you're looking at here is embryos that have been stained with a DNA stain. Here's a wild-type one from a wild-type mother. You can see it's filling up with these diploid nuclei doing those SM-SM cycles. And here's a phenotype that really grabbed us because what's happening is there's this giant nucleus filling up with DNA, but there are no mitotic divisions. And we actually recovered quite a few mutations that gave this phenotype. And they ended up falling into 3 complementation groups. So the 2 complementation groups that we identified were called Pangu and Plutonium, and the 3rd complementation group that gave this phenotype was called giant nucleus and it was identified in David Glover's lab. Now, they all gave the same phenotype. If we looked for genetic interactions, they all genetically interacted, which was consistent with all three of these genes regulating the same process. So, you know, Drosophila researchers are famous/notorious for their gene names. The genes are always named after the mutant phenotype. Giant nucleus is obvious. Pangu is way too long of a story to tell you. But we named the other plutonium because it had a giant nucleus and was lethal. So that was the basis of that name. All right. So we really wanted to understand how do they act at a molecular level. And we thought we had it cracked. So what Lori Lee and Leah Vardy in my lab were able to establish is the reason you get that phenotype is that these mutants complete meiosis and they start DNA replication, but they can't do the first mitotic division. So they're locked in DNA replication and just continue to replicate and replicate but don't do any divisions. And what we were able to show is that could all be explained by these three genes being required to translate cyclin B mRNA. So after meiosis, cyclin B proteins drop, and then you need to restore them by new translation. And these genes are needed for that. We could completely— not completely, we could partially rescue the phenotype by adding back in cyclin B. So we thought we had it, but then what happened is Howard Lipshitz's lab was looking at a later event in embryogenesis when there's a handoff from maternal control to zygotic control and the maternal mRNAs become degraded. And what the Lipshitz lab found was that these three genes also were needed later in embryogenesis genesis for this degradation of maternal mRNA. And what we were able to show in collaboration with them is that's because there was a second target of these genes, and that was that they were also required to translate this protein SMOG, which, yes, is named after the dragon. And what SMOG does is recruits the deadenylation machinery to lead to degradation of maternal MRNAs. So we thought we had it. It's very specific, regulates translation of these two MRNAs specifically at the oocyte to embryo transition. All right, so what happened then is Eva Kronja decided to take a step back and to see if she could get a bigger view of all of the changes that happen during these stages through meiosis and during the oocyte to embryo transition. And so she did these kind of global studies where she used quantitative mass spec to identify the proteins whose levels change during oocyte maturation and egg activation. And then in collaboration with the Bartel lab, she studied looked at all of the mRNAs to see which ones changed in their translation levels during these developmental windows, and then later helped the Bartel lab to look at how polyadenylation changes. And so let me tell you what she found out about translation. So if she looked at translational efficiency for each of the mRNAs present in the in a mature oocyte and compared that to those in— once egg activation has happened, she found that there were extensive changes in translation. So about 1,000 mRNAs become activated for translation and about 500 become repressed. So there's massive reorganization of what mRNAs are being translated, and the idea that developmental biologists scientists had, that the mature oocytes are just sort of sitting there quiescently, really wasn't the case. There was extensive translation happening in the mature oocyte, and then that got remarkably changed in response to egg activation. So once Eva had this, we— I thought, well, let's just look because maybe there'll be a couple more targets of PANGU, PLU, and GENU. So let's— now we know what the reference is, let's look in these mutants and see if they have some other mRNAs whose translation they affect. And Eva did this, and I don't know if I looked like this, but this is certainly what I felt like when she got the result, because what she found is that if we look at the mRNAs that are translated traditionally activated, 85% of them are partially or completely dependent on PANGU. And even those that are inhibited, also 90% of them are partially or completely dependent on PANGU. So now we had to think about these genes in a completely different way, that these were really acting in a global way to regulate translation of maternal mRNAs. So of course we wanted to know how it worked molecularly. We thought that— we also thought that would crack really readily because when we cloned these genes, what we found is PANGU was predicted to encode a serine/threonine kinase. So we thought, okay, this is going to be pretty easy. We'll somehow sort out how this— by acting as a kinase, how does this all work? That didn't turn out to be the case because we had— I think it's like childbirth, I think I've sort of blanked it out, but we had 3 to 5 years of unbelievable frustration where— sorry, Kim— where No matter what we did, we couldn't get this kinase to be active. But we— it had to be a kinase because, you know, we had a lot of alleles. The alleles that completely ablated function were in conserved residues in the catalytic domain. Alleles that had a partial effect were in unconserved regions. It had to be a kinase. And so we kept hammering at it, and thankfully we did because it had a really beautiful biological explanation. Explanation in the end. And that was that what Laurie Lee was able to show is that the reason these genes all genetically interacted is they encode proteins that are physically in a complex. PANGU is the serine/threonine kinase subunit, but it's completely dead unless it's physically bound to these two activating subunits, which are encoded by Plutonium. And Gnu. So now we understood the genetics, we understood why we had struggled so much with the biochemistry. So here's where we were then. We had identified a new protein kinase complex. This protein kinase complex acts at this window of development to regulate the translation of maternal messages. And so now the questions were, how is it developmentally regulated, and how does it control translation? And we feel like we've really solved this problem, so I'll tell you the answer. I think it's a really elegant developmental answer. I'll tell you where we've gotten to in terms of understanding how it controls translation. So to think about how this kinase could be developmentally regulated, what Masatoshi Hara in the lab did was to first ask, when in developmental time are the proteins present. And so what you're looking at here are Western blots of extracts from these stages that we can so nicely isolate of progression through meiosis. And the thing to remember is the kinase is dead unless it's in a physical complex with its two activators. And what Masatoshi found was that all of the subunits are actually present in the mature oocyte. But what he found is Plu and Pangu are together in a complex, but Gnu is not. And so they're all there, poised, ready to go at egg activation, but the kinase is not active. So the question then is, when does it become active, and can we get a molecular handle on what might be responsible? For activating it. And here's where there's— the Drosophila has another tremendous advantage that you can get the whole activation process to happen in vitro simply by taking oocytes and swelling them in hypotonic buffer. This is something that Tony Mahowald worked out some years ago. So we did that and we asked— our assay for when Pangu became active was to ask for the appearance of SMOG protein because the translation of this mRNA is completely dependent on active PANGU. So here's the results. It's a complicated slide. Let me just walk you through this. It's a series of Western blots together with H1 kinase activity. So what we could do is at time points after in vitro activation, we could ask when does PANGU become active by when we see see the appearance of smog protein. So you see it's right here. And then we could also look at these oocytes and ask what's happening in terms of meiosis. And so coincident with Pangu becoming active was the completion of meiosis. Now there was another striking event coincident with Pangu becoming active, and that was the state of Gnu. So Gnu went from being slower mobility form protein to a higher one, and we could show that that difference was phosphorylation. So let me just sort of cut to the chase and tell you what we were able to sort out about this developmental regulation. Here's how it works. In the mature oocyte, which is arrested in meiosis, Cyclin B-CDK is active. That phosphorylates Gnu, and phosphorylated Gnu cannot bind to Plu and Pangu. And so at egg activation when completion of meiosis is triggered, cyclin B is degraded, this becomes inactive, we see that Gnu becomes dephosphorylated, and now we see the formation of the kinase complex and its activity, and it regulates translation. But this kinase is restricted to being active for a very narrow window, less than an hour, because what happens is once the complex forms, Gnu becomes degraded in a Pangu-dependent way. So the onset, the developmental onset of this active kinase that regulates maternal mRNA translation is the completion of meiosis because of the requirement for dephosphorylation of Gnu. And then it self-destructs to limit it to not have a role later in development. So we were really pleased with this, how exquisite this developmental control is. And so the question then is, how does it control the translation of 1,400 mRNAs, a third of the mRNAs present in the oocyte? The way we're going after this is to look for substrates. And, you know, what I think the answer ultimately is going to be, this being a kinase, that it probably has multiple substrates and a myriad of ways of regulating translation. We did a substrate screen, an unbiased biochemical substrate screen, and found that many of the components of the translational machinery are substrates for Pangu kinase. One of the ways that I think it acts is that it— we found that it can phosphorylate translational repressors, and that appears to inactivate them. So we think that at least one of the ways it can activate translation is by phosphorylating and inhibiting the activity of translational repressors. So before I leave this vignette about the oocyte to embryo transition, let me just tell you a couple of other important things that I think came out of the work. One is I've been talking about mRNA translation, but proteolysis is really critical also in this window. And what we found is that there's developmental regulation of proteolysis because there's a meiosis-specific form of the anaphase protease-promoting complex that has meiosis-specific targets. So it's a way to clear out proteins that are involved in meiosis and need to be removed to not hinder mitotic divisions. And this was largely the work of two graduate students in the lab, Julian Pessin and Zach Whitfield. Emir Aviles-Pagan, who's a current student in the lab, is looking at Gnu. What is Gnu doing in the mature oocyte when it's not with its buddies? And he's looking at its function, whether it has Pangu-independent functions, and whether its localization to cytoplasmic RNPs affects its regulation and activity. And in a really different project, what Boryana Petrova in the lab did was to ask whether there are changes in redox during this stage that could also regulate this transition. And what she's found is that there are both global changes in redox, that the oocyte is in a reduced state and becomes more oxidized in this transition, and there are also specific proteins whose redox state is changed during this transition. And that's actually critical for the completion of meiosis. In the activation of the embryonic divisions. So let me— what did I learn from this? What I learned from all of this was that it's important and powerful to use diverse approaches. Genetics, I think, is a great way to start a problem, but to really understand what PANGU was doing, we had to do biochemistry. And our doing omics approaches gave us kind of the catalog of information that was important for sorting it all out. The other thing I learned from the pangu kinase battle was it's important to be resilient and to keep pushing on a project. I learned that students and postdocs are incredibly resourceful so that, you know, they— Eva was fearless about finding somebody to do the quantitative mass spec with and about approaching that and approaching translation at an omics level. We used collaborations in two ways. One was a pragmatic way. We collaborated to avoid collisions, so we collaborated with the Lipschitz lab when it seemed like we were colliding on PANGU regulation of smog. But really a more powerful way to use collaborations is to be able to use them to synergize different expertise to bring it to bear on a problem. And so the quantitative proteomics was done in collaboration with Jeroen Kriegsfeld's lab at the EMBL. I mentioned that we collaboration— collaborated with the Bartel lab, with Steven Eichhorn and Alex Telny to do the translatome. And we did two PANGU substrate screens. The first was in collaboration with Mark Kirshner. The one that really worked the best for us was in collaboration with Sebastian Laredo here, and Ben Turk at Yale really helped us to begin to understand the substrate specificity of PANGU. So let me turn to the second vignette, which is another— based in another cell cycle variant. So this canonical cell cycle maintains the diploid content of a cell by always linking partitioning of the replicated sister chromatids to DNA replication. And it seemed like there was likely to be a different cell cycle variant in Drosophila because most of the differentiation activated cells actually had an increased DNA content. And so we went after this initially to really understand its development, and then I'll tell you how we've exploited it to understand some other aspects of biology. And ultimately, we established this really is due to an altered cell cycle that we've termed the endocycle, and it's one where there's a discrete period of of DNA replication followed by a gap phase of growth and gene expression, but these aspects— but mitosis is shut off. So, you know, our reason for looking at this was curiosity. You know, we had been looking at these incredibly beautiful polytene chromosomes and really wanted to know how they are made. So what you're seeing here is that each of these chromosome arms has 2,000 DNA helices that are replicated copies, and they're physically aligned together to give this banded pattern. So Albert Smith, as I said, decided to take this on. We didn't know anything about whether there really was a cell cycle and how it was developmentally regulated. So he started with a very basic question. Could he look in embryogenesis and could he find evidence during embryonic development of some kind of SG cycle? And so he did it by pulse labeling with a nucleotide analog, BrDU, and looking at successive stages of embryogenesis. The details here aren't important. What is important is that he was able to show that these developing larval tissues in the embryo began S phase at a certain time and an S phase that was never then coupled to mitosis. So what he was able to show is that there was this stereotypic onset consistent across embryos of the endocycle in embryogenesis. He was able to show that there was tissue-specific control of how long the S phase would be and how long the G phase would be in these differentiating tissues. And so he established a map of when the endocycle gets set up during embryogenesis. And then in some really important work from Bob Daronio and Pat O'Farrell, they were able to show that there's a transcriptional program of S-phase genes that exactly mimics the pattern of the S phase. So the question was then, can we do a genetic screen to isolate mutants that affect the endocycle? And we initially tried to be clever about this. We tried to do this with a reporter. And when that completely fell apart, and so my sense at the time was we should just give up. But I had this unbelievably determined graduate student student, Irina Roisman, and she was determined to do a screen. And she convinced Allison Whitaker, another graduate student, to join her. And what they did was they did a visual screen. And so they screened 4,500 mutants by in situ hybridization from one of these S-phase genes looking for defects in the endocycle. Actually, while this was happening, when I would go to meetings, I would never tell anyone that we were doing this, not because I was afraid of competition, but I was too embarrassed to say, you know, my students are looking at 4,500 mutants by in situ hybridization. But anyway, they did it. And it paid off. They got lots of things. They got mutations in G1S regulators. Two examples here, the two subunits of the E2F transcription factor completely block this endocycle S phase. They got mutations in DNA replication components, but because this was a visual screen and they were looking at each of these embryos, they also got lots of things we hadn't set out to look for, namely things that were involved in mitosis and cell division. So they got the mutations in the separase, in the condensing complex, in several centromere proteins, and proteins needed for cytokinesis. And so as you can see from these publications, this was a goldmine for us. This gave us things to work on for years and years when, you know, for years after this when new people would come to the lab, I'd go to the incubator and take out some mutants and say, here's Here are 5 mutants you could study. Would you like to study these? So it gave us a lot to do. Okay. So the— let me tell you 2 big picture things about the endocycle that came out of this screen. The first was the role of these G1 to S regulators, E2F and cyclin E, in promoting the transition from the G phase to S phase in the endocycle. You may know these are the key regulators that in the canonical cell cycle control the transition from G1 to S. And they're also used in the endocycle. And I want to point out that some really beautiful work subsequent to this from the Edgar and DeRogna labs showed that oscillations of protein levels of the E2F transcription factor are what drive the oscillation in the endocycle between the gap phase and the S phase. Allison Whitaker identified a mutation that completely blocked DNA replication, and when she cloned that, she named the gene DUB, which stands for double parked in a tow zone. This is a lab joke, but anyway. So, and so she looked for the human homolog and found that there was a human homolog, and when she found those because she recognized that this was a new protein family also composed of a protein from fission yeast called Cdt1. And this turned out to be a new component of the replication machinery that's essential for the initiation of DNA replication by recruiting this complex of proteins that makes the active helicase. Okay, so what I want to tell you about is where we've gone from this foundation of understanding the endocycle to learn some more biology. And I'd like to tell you sort of two biological principles that we've learned. And that is that the endocycle is a way to grow large cells, and I'll tell you some examples of that. And second, that we can use the way S phase is regulated in the endocycle as to give us some very powerful models to be able to directly look at replication origin activity and fork progression in a metazoan. And I'll tell you some principles we've learned from that. Okay. So what we know throughout biology in plants and animals is that when there are large cells, that is invariably accompanied by an increase in DNA content. And that's true in plants. It's true in mammals as well. For example, these trophoblast giant cells in the placenta, these keratinocytes in the skin, and megakaryocytes. So it seems to be the case throughout biology that if you need a big cell, that's accomplished by increasing DNA content. Both nuclear size and cell size scales with DNA content. And there really appears to be some kind of minimal nuclear-to-cytoplasmic ratio that's essential. Okay. So I want to tell you what we've learned about how to grow a big cell and why to grow a big cell. And let me just make sure we're on the same page about size. I'm not talking about some, you know, diddly 1.7 7-fold or 2-fold change in size. I'm talking about orders of magnitude differences in size. So, for example, these are a type of glial cell in the Drosophila larval brain, and the surface area of one of these cells is 15,000 micrometers squared. So if we compare it to the area of an epithelial cell, this is 10,000 times bigger. Okay? So what you need to think about in terms of thinking about tissue and organ growth is to not simply think about it happening by an increase in cell number, but to realize that in biology an alternative strategy is to have tissue growth happen by a constant cell number with those cells increasing in size. I'm going to give you two examples from the larval nervous system of two types of glial cells that use this strategy for growth of their tissue layer because these, I think, are wonderful examples. They've really given us some insights into why this might be an important developmental strategy. Strategy. So the two types of glial cells I'm going to talk about are called subperineural or SPG glia and wrapping glia. These glia are present throughout the nervous system. Wrapping glia are present in the peripheral nervous system. And so the question is, why would a tissue layer grow by increasing cell size rather than number? Here's a schematic from Christian Klamp's group of the different types of glia in the Drosophila larval nervous system. This is basically looking at a section through the ventral cord of the nervous system. The glia that are important for us are these subperineural glia. They're glia that lie at the surface of the nervous system, and these wrapping glia that are or just in the peripheral nerves. So what these subperineal— I'll just use SPG— cells do is they're these giant flat cells, and they're joined all along their edges by these septate junctions, which are a type of tight junction. And these SPG, via those junctions, make the blood-brain barrier. So the blood is floating out here, and these prevent ionic ions and other molecules from passing into the underlying neuronal mass. The wrapping glia basically surround the axons in the peripheral nerve and ensheathe them. Now, what we found is the SPG number doesn't increase during development, but they just keep getting bigger and bigger, and the way they keep getting bigger is by increasing their DNA content. And so we could test for functionality of that because we had the tools from the community to specifically block the ability of the SPG to increase their DNA content. And when we did that, we could see their function was defective. Their function is to make a barrier, a blood-brain barrier. If the blood-brain barrier is working, if you inject the DNA, a fluorescent dextrin into the body cavity, it won't cross into the brain. If we blocked polyploidization of these glial cells, the blood-brain barrier is defective. So here's how you can think about it. Basically, if you don't allow these SPGs PG to become bigger and bigger by polyploidization, but the underlying neuronal mass continues to increase from neuroblast division. Think about the clothing here. What happens is this envelope, this like Saran Wrap envelope, basically ruptures and the blood-brain barrier is defective. And we could show that, you know, in fact, the septate junctions that are making the seals between the cells actually ruptured if that was the case. So here's the biological explanation for why this could be a good strategy. An organ in which there's a tissue layer that needs to act as a barrier, if those cells divide and undergo cytokinesis, at least transiently there's going to be a leakage in that barrier. And so we think that's the likely explanation explanation for why cells like the giant trophoblast cells, which make a barrier between the fetal and maternal compartments of the placenta, use a strategy of increasing cell size rather than cell number. Now, the wrapping glia, it's a similar story. We think that in a situation in which tissue architecture gets established, that cytokinesis could really disrupt that architecture. So here's a schematic of these wrapping glia. This is a section through a peripheral nerve. They're interior to the SPG, and they ensheath the axons that are running through the peripheral nerves. And here's some enigmas about the wrapping glia. There's only 3 wrapping glia per peripheral nerve, and that's despite the fact that if you look at an— a nerve in the anterior versus the one that extends to the posterior of the animal, there's a 50-fold difference in length of these nerves. And if you think about development, there's up to a 50-fold difference in length that happens in development. And this is work from von Hiltschin using lineage tracing. So the question is, how can you have this constant number of wrapping glia— but yet have such vast differences in the size that they need to— sheath. And so you're probably catching on now. The answer to this was worked out by Laura Frauley in the lab, and that is that they don't increase number, but they increase size by ploidy. And that is reflected in the length of the nerve, that the more posterior longer nerves have a higher ploidy. And she could do the same kind of experiment of blocking their becoming polyploid and asking what happens. And when she did that, she found that that leads to axonal achievements. So here's another biological example where this strategy makes sense, and that's if you've established a— wrapping glia and sheathing cytokinesis may really disrupt that ensheathment. Okay. So these are developmental contexts in which this is a useful strategy, we think, taking advantage of the endocycle. So for the last point, let me turn to a more mechanistic story, and that's how we exploited the endocycle to really understand specific aspects of metazoan DNA replication. And the two I want to tell you about is addressing can the progression of replication forks be inhibited, and what happens if replication forks collide. So let me give you the background for this. What I haven't told you is that during the S phase of the endocycle, cycle, there's not necessarily an integral doubling of the genome. So in Drosophila tissues, what happens is there's a region of— there are regions of the genome, usually about 30% actually of the genome, that's not replicated. So in those polyploid cells, there's these regions that are under-replicated. And the opposite can happen, that there can be specific regions that repeatedly initiate DNA replication to have an amplified gene copy number relative to the overall ploidy. And so this is what it looks like if we assay this on microarrays where we're comparing copy number across the genome relative to an embryonic standard. We can find these valleys of decreased copy relative to overall ploidy, or the flip side, these regions where there's a gradient of increased copy number. And what we realized is this would be very powerful to understand how you can regulate initiation and fork progression, because here initiation is repressed, forks are somehow inhibited, and here initiation is hyperactivated. Interactive, if you will. And the reason it's so powerful is we did a survey of tissues to look at this, and so we knew exactly in the genome where these regions were. We focused on single copy regions that we could handle molecularly, and we also knew exactly during differentiation when these replication changes would happen. So let me tell you what we decided— what we went after. We wanted to understand how initiation is regulated. This is conserved from yeast to humans. There are regions at which initiation can happen, are marked by an origin recognition complex that then can load an inactive form of what will become the DNA helicase. Initiation of replication involves activating that helicase to melt the DNA, to recruit the polymerase machinery. But then DNA replication also requires that these forks might continue to replicate until they run into another fork to ensure complete replication of the genome. And we understand a lot about how initiation is regulated from yeast, very little about it from metazoans, and in both— in tells all these eukaryotes almost nothing about what can regulate fork progression. So here's what we— I'll just summarize the key points that we've learned and then really focus on regulation of fork progression. What we found from these developmental models is that in Drosophila there are multiple mechanisms to activate replication origins, that when we look at where ORC is positioned, where potential origins are. We see there's extensive tissue specificity of where it localizes. But the converse of that is there are constitutive genome regions across tissues that lack ORC. So these lack the ability to initiate replication. And that doesn't— that's not sufficient to block replication, but whether these are replicated or not depends on active control of replication fork progression. So the key message I want to give you from this story is there's active tissue and region-specific control of whether forks can cross a region or not. And the reason this is important is this tells us that in thinking about things that can impact genome stability and lead to instability and progression to cancer, this tells us that it's now really important to think not just about initiation and copy number control from that aspect, but also for progression. So how did we get here? The way we got here is that, again, these developmental events gave us very powerful models to be able to, in a very direct way, in this case a visual way, to look at what happens with fork progression. And so let me illustrate that for you with amplification. So the amplification event happens in these somatic cells in the ovary called follicle cells at 6 specific specific sites, and it's precisely regulated with respect to follicle cell differentiation. So this— none of this amplification happens until replication of the entire genome is shut off. And what that means is if you pulse label with a nucleotide analog, during these stages when origin firing and fork progression are happening. You'll see specific sites in the genome where you see a spot of nucleotide incorporation. This is a single nucleus. But then there's a developmental transition where there isn't any more initiation, but the forks continue to migrate bidirectionally. And so now when you look at where the nucleotide analog is incorporated, now you see this resolved to these two double bars. And the distance between the double bars tells you the distance the bidirectional forks have migrated, but importantly it tells you where the active forks are. So now we had a way to directly visualize and know where the forks are. And so that gave us a way to identify the fact that there are proteins that actually can load onto the fork and block fork movement. And the one I'll tell you about is a protein called SUUR. It's a dosage-dependent repressor of fork progression. The way you could see that is from work from Jared Nordmann where he looked at what these profiles of amplified gene copy number look like. This is the distance along the genome. Here's the wild-type extent of one of these amplicons. If you don't have this protein, now there's a much longer amplified region. And if you overexpress the wild-type protein, now there's no effect on initiation, but now you see that you don't get replication fork movement to the same extent and you get a very narrow window. So what this told us is this is a protein whose wild-type function is to impede fork progression. And you could think about this happening mechanistically two ways. One, it could be like a concrete barrier sitting in the chromatin that when forks arrive there blocks them from moving further. But the other interesting possibility was maybe it was localized to the forks. And I think you can see this is a difficult problem except that we know where the forks are, and we can see them. And so Jared Nordman took advantage of that and asked, where's the localization of this SUUR protein? And what he found is that when amplification first starts, it's not there. But then there's a window when it becomes recruited to where the forks are. And if you look later when you can visually distinguish where the forks are, you can see it's clearly tracking with the forks. And he was able to confirm this by chromatin immunoprecipitation. So what we've identified is the fact that there are proteins that can get onto the fork. Interestingly and importantly, when and where they get on the fork is regulated by tissue and also in a region specific way. And so they can set up a domain that will— in which replication will be impeded. And so there are lots of interesting questions to understand. One, how does this actually get recruited to replication forks? How does it inhibit forks? And are there other proteins that act with it? And really an important question is, is this the mechanism that in diploid cells is responsible for setting aside regions of the genome to not be replicated until late in S-phase. So these are fascinating questions, and Jared is going to solve them all in his own lab at Vanderbilt University. I want to tell you one more thing, and that is how we've used this system to address an important question in cancer biology. So what— cancer biologists who think about DNA replication have proposed is there might actually be two reasons why refiring an origin is a problem. The first is the one you're probably used to thinking about, that you get this increase in gene copy number. But the second is that you set up this situation where you have two forks traveling on the same helix, and if this fork runs into that one, it generates a double-stranded break, which leads to genome instability. And of course the problem in mammalian cells is that you could trigger re-initiation and see globally there were double-stranded breaks, but since you don't know where the forks are, it wasn't possible to know if this was really the event happening. And so what Jessica Alexandra, a graduate student in the lab, did was taking advantage of knowing where the where the forks are, she asked what— whether these cytological markers of double-stranded breaks appeared where she knew the forks were in the position to collide. And I'll just show you the cell biology answer. She confirmed this by ChIP-seq. And that is when you look at where the forks are, you see that this modified histone accumulates at those forks. And this other marker of double-stranded breaks, RPA, also accumulates there. So, indeed, we think that these fork collisions resulting from refiring result in double-stranded breaks. She was able to ask what are the mechanisms for restoring it by looking at restoration of fork progression, and this was another surprise that rather than repairing these in a way that's not going to be mutagenic, she found they're exclusively repaired by a mechanism that generates further mutations, non-homologous end joining. So this is problematic by multiple mechanisms in generating genome instability. Okay, so what we've learned by exploiting this differential replication that happens in the endocycle is the genome is surprisingly organized with these regions that constitutively lack ORC across tissues. Struhl's lab has shown this is also true in human cells. And so whether or not these regions get replicated is actually the consequence of active tissue-specific control and region-specific control of fork progression. So let me just finish this second vignette by what did I learn from this research. I learned to keep your eyes open for interesting biology. The dogma in Drosophila that was that there weren't any polyploid cells in the nervous system, there are at least 3 types of glia that are polyploid. Don't be afraid to enter a new field. You know, we decided we needed to go into neurobiology to understand what was happening with these glia. And again, I really was educated by my fearless students and postdocs who learned these new fields and learned— went and sought the right expertise to know the experimental techniques to advance them. Community resources were critical for us. We relied on these Gal4 driver lines that the community had generated, as well as RNAi lines against all the genes in the genome for a lot of our experiments. And the final thing I want to emphasize is that, you know, when you look for a biological problem and you keep your eyes open for some unusual biological examples, it may be easy to dismiss them as some sort of strange Drosophila a thing or some esoteric thing. And, you know, what we've really learned is that if you explore even these unusual biological processes, they give you these insights. In the case of looking at fork progression, I think really important insights in how we have to think about this in thinking about genome instability in cancer cells. We, again, collaborated on a lot of these projects. The wrapping glia work was done in collaboration with Megan Cordy and Mark Freeman. We've had a number of collaborations over the years on DNA replication with Steve Bell and David McAlpine. We collaborated with Igor Jemilev's lab to do the work on SUUR. And we collaborated with Mitch McVey to look at the mechanism of repair of these double-stranded breaks from fork collisions. You know, this is really a great career. It's exciting to be able to investigate these problems, to make discoveries, to constantly be learning new things. But really, certainly the best part of it is the personal interactions. And so I've had so many wonderful people come through the lab over the years. A lot of them shown here. I tried to point out the students as I— and postdocs as I went along on these projects. I haven't told you about the people who worked on chromosome segregation and meiosis. And I also want to point out that the past few years it's been really wonderful for everyone in the lab to have Tony Mahowald in the lab as a visiting emeritus professor. So here's the PhD students who completed their theses in the lab and the postdocs. Really, none of our research would have been possible without the technical support by these two long-term technicians. Raissa Fomina kept our Drosophila stocks, thousands of them, and kept them healthy, and that was essential for our experiments. And Helena Koszewski contributed on many of the experiments in many technical levels. I don't want to stop without really highlighting the Whitehead Bioinformatics Group. So as we made this transition from genetics to really looking at things in a more omics level, we absolutely could not have done it without support from the bioinformatics group. George Bell played a pivotal role in many of the projects, and also in Memburasa and Bingbing Yuan. And right now the lab is funded by grants from GM and NINDS, and I'm an American Cancer Society professor.

Thanks very much. I'm sure Terry will be happy to take questions, but one thing I wanted to say before we begin that is that for those of you who are available, there will be cocktails available on the second floor following the talk, so you're welcome to join us there as well.

In case you need a drink after.

Yeah, up here. You mentioned the targets of the active fangu complex is the core translation machinery itself as well as some of the translation repressors. I'm wondering if the core translation machinery targeting is essential or just a part of it. For example, if you look in double mutants with some of the translation repressors in fangu, are mutations in the repressor sufficient to suppress the translation block?

Yeah, so that's a great question. We haven't done that experiment. So the genetic analysis we did, the two translational repressors that we've really shown a functional relationship with Pangeru, with our conserved one called Trailer Hitch and another one called Prunellia, and when we looked at that, we looked at the cell cycle effects, so we looked at translation of cyclin B and restoration of the mitotic divisions in the embryo, but we didn't do a global look at translation. It would be a good thing to do. Where's Amir? Yes. [Speaker:MALE_SPEAKER] Why does it use non-homologous end joining? [Speaker:KATHY_HANSON] Yeah, I was— I was so surprised by that result because why use non-homologous end joining? Because you have 80 copies from amplification, right? And they're also polyploid cells, so you have 80 times 16 templates. So why wouldn't the cells use homologous recombination? Homologous recombination is actually a problem. Things get better if you get rid of homologous recombination. Homologous recombination, it's really competing for those ends with non-homologous end joining. And so we were perplexed by the result. You know, one thing that— what mammalian cell biologists say is that it takes a whole lot longer to do double-strand break repair using homologous recombination than non-homologous end joining. And, you know, what we're asking is repair of those breaks to restore fork progression. And so it's maybe sacrificing accuracy for speed. Yeah. So it seems like you'd often expect only one end regenerated by these. So do you think it's requiring multiple fork collapses to generate enough ends to do NHEJ? Yes. Any evidence for that? Yeah. Yeah, we just spent a lot of time drawing these things out, yes, and that seems to be— you know, you have all these forks piled up across all the DNA helices, and yes, I think that's the case. And/or also you could get joining across the origin region. And so, you know, one thing we always wanted to do was to try to sequence all the joints to really figure out what was happening. And we tried that a bit, but really, yeah, we couldn't really get examples to follow through on. Yeah? [INAUDIBLE] So Jerry's question is, can you take a polyploid cell and get it to re-enter division? So there are two examples in Drosophila. You may have noticed those SPG glia actually are multinucleate. And what they do is they enter the endocycle, and then they actually receive— restore mitosis and nuclear division, but not cytokinesis, to become multinucleate. And that turns out to be critical for them to get big enough. You can't actually— with these flat cells, they can't actually— we don't think they can get big enough with a single nucleus. So that's an example of restoring mitosis and nuclear division, but not cytokinesis. And then, to my knowledge, there's only one other example example I know of in biology, which is another example in Drosophila. And that is that these cells that make up the rectum actually become— they endocycle to become 8C. And then after they've done that, then they actually restore cell division. And both, you know, they restore cell division, so they're increasing cell number in those The cells are 8C, and you need that to have the proper properties of the gut in terms of salt absorption. And those are the only two examples I know of. And it's a disaster, right? So I mean, when these— you know, in terms of thinking of genome stability, these SPG nuclei, we see so many mitotic defects, and those rectal in fetal cells, also they have so many mitotic defects, right? Because, you know, it's a deal, right, if you do an S/G cycle and you don't completely accurately maintain copy number during that S phase to then reinsert nuclear division or cell division is problematic. But it's part of the normal biology, so it's important.

Thank you, Carrie, again.