 All right, hello everyone. I'm Mayor Sadu. I'm an investigator in NHGRI. I'm honored to be introducing our guest today, Dr. Harmeet Malik. Harmeet is at the Fred Hutch in Seattle, Washington. For those of you who attended his walls talk yesterday, welcome to part two. As you're saying, it's very impressive. Harmeet is talking about two different topics in two days. And I'm still seeing if we can schedule you for another day tomorrow, maybe. You can give a third talk. So I wanted to give some quick background. Harmeet did his PhD at the University of Rochester in Tom Ikebush's lab, working on rector transpos on evolution. From there, he went to the Fred Hutch to do his postdoc with Steve Heneckoff, where he did the work that sort of seeded the stuff he'll be telling us about today, looking at the rapid evolution of centromeric histone proteins. I wanted to quickly note how crazy that is. Centromeric histone proteins are incredibly conserved, and they're found across all eukaryotes. And yet, apparently, they're rapidly evolving, which means changing quickly. And I think this sort of speaks to a theme in Harmeet's work. He finds these topics that are incredibly important on the one hand, and also incredibly cool on the other hand. And it's the marriage of the two that I think makes Harmeet's work so fascinating. Harmeet has won numerous awards and titles. I wanted to quickly mention he's a HHMI investigator and member of the National Academy of Sciences, and last year was the president of the Society of Molecular Biology and Evolution. I think the real testament to Harmeet's impact on the field can be seen in how his trainees, so many of them, have gone on to, in their independent careers, make lots of important contributions to the field of genetics and evolution, which I think goes to show how Harmeet is really a thought leader at the forefront of how we think about genetics and evolution research. And I think it also goes to show Harmeet's personal warmth as a person, and how he really nurtures and is dedicated to his trainees and to young scientists in the field. All right, so with that, I'd like to invite Harmeet Malakup. We're very excited about your talk and about your visit. Thank you, Mary, for that very warm introduction, and thanks to those of you who are braving a second afternoon worth of talks with this. I hope this is going to be worth your while. So I'm going to actually start my talk, and it might look very familiar for the first two slides, but I want to assure you it is a different talk, and you're going to be in like very different space. But actually, if you were here yesterday at the talk yesterday, I think you'll be able to see like why these two arcs fit into the same lab, even though at first glance they might seem like totally different projects. Like I mentioned yesterday, my lab is very interested in the idea of genetic conflicts, and this idea actually owes its roots to the fictional character, the Red Queen, that I've cartooned here, who we were introduced to by Louis Carroll in his books about Alice's adventures in Wonderland. At one point, in response to Alice's complaint that they've been walking for a very long time and really haven't made any progress, the Red Queen responds that in Wonderland it takes all the running you can do to keep in the same place. If you really want to go somewhere, you really need to run doubly fast, and like I pointed out yesterday, this is a perfect metaphor for what it means like to run a lab. I wanted to also sort of point out that this metaphor was actually seized upon as this very important theorem that led to the Red Queen hypothesis being formulated, which was basically pointing out that when systems are in conflict with each other, it's not sufficient for them to respond to their abiotic environments because a large degree of the fitness space is dictated by all of the other species in the environment that are competing for that ecological niche. So in a predator interaction, we cannot actually decide what's going to happen to the snow hare population without taking into account the adaptive strategies of the snow leopard population. And going back and forth between these gives us this beautiful cadence between these two species that are distinct species, but they are linked because of their fitness effects on each other. And that analogy can be very nicely extended to the analogy between host antiviral proteins and viral proteins, like I discussed yesterday, where the viral proteins upon being recognized by the antiviral proteins are spurring a wave of adaptation that allows them to escape the antiviral protein, and that forces the antiviral proteins to respond to the newly evolved version of the virus. And this basically is an engine for forward evolution, and essentially that's what my lab does is we study this kind of relentless rapid evolution both as a way to understand how these battles get waged in biochemical terms, but also to discover new battles that we might not even have been aware of in the lab. And we are primarily going to focus on protein coding genes. There's a very small subset of very brave people in the lab who are venturing outside protein coding genes in the lab, but that's not going to be something that I discussed, but I'm happy to discuss in person or in the Q&A section. But in protein coding genes, one of the primary reasons we focus here is because we have this beautiful internal control to distinguish the effects of mutation from the effects of natural selection, because most nucleotide substitutions will either alter the amino acid being encoded, leading to these non-synonymous or replacement changes versus those that do not alter the amino acid being encoded, silent or synonymous changes. And we make the somewhat like false assumption that silent or synonymous changes are essentially hidden from natural selection because they didn't alter the amino acid being encoded. The reason this works, even though this is a false assumption, is because on average an average silent change is completely benign compared to an average replacement change over the course of evolution, which means that if you're interested in what is conserved in evolution, right, catalytic residues of enzymes, et cetera, you would identify what are the amino acids in a protein alignment that are completely conserved despite millions of years of divergence. This is an indication of selection acting in a purifying capacity because it has purged all of these presumed deleterious non-synonymous changes that have occurred during the course of evolution. In other words, we use this metric to deduce what is highly conserved because what is highly conserved in evolution must by definition be very important. But we sometimes make the mistaken assumption that things that are not conserved must by definition also not be important. In fact, that's not really true. There are many parts of the genome that are also under meaningful selection but in the opposite direction, where they're actually accelerating the rate of fixation of amino acid altering mutations. And we use this sort of signature of constant, relentless innovation to identify where these innovations are occurring, what are the cellular processes that are subject to this? What's really nice about this very simple metric, and there are much more sophisticated metrics that take advantage of multiple alignments, et cetera, but even with this very simple metric, we can very quickly cast an eye onto our genome and the genome of aclosus relatives and identify what are the protein-coding genes that show a signature of slow evolution, much slower than what you'd expect based on that dotted line, which is the speed limit that is potentially imposed by selection, and what is actually under rapid evolution, much faster than what you'd expect based on chance alone. And we're very interested in identifying why these fast-evolving genes are fast-evolving. That's basically the essence of what my lab is actually interested in. In particular, we're very interested in gene categories that show up in this fast-evolving class no matter whether we look at primates or more charismatic organisms, like the drosophila species that we study in the lab, where the same genes often and often the same genetic categories show up repeatedly. And I'm just going to highlight some of the categories, which might be of interest to some of the people in the room, because I hope you can appreciate that, whereas you can begin to make a story about things like BRCA1, BRCA2, DNA repair genes, transposon defense, these are, after all, some sort of arms race, that sort of intuition starts to really break down when you realize that there are many genes in this category that are in the rapidly-evolving class that really have no business in this rapidly-evolving class, given that they're completely essential for the viability and fertility of the organism. And in particular, I'm going to focus on mitosis and meiosis proteins and chromatin proteins in general to highlight this, but I'm very happy to talk to you again about all of these gene categories, because in some sense we are interested in all of them. And this sort of foray has forced us to confront the possibility of actually studying these in a variety of different model systems all the way from ciliated protozoa to different species of fungi to butterflies to C. elegans and even very grudgingly to mouse, which is like a very expensive four years in my lab, which I'm very glad is in the past. So this is essentially all telling us that we have this view of the cell that has emerged from many, many molecular biology studies, which is completely true, that we can think about the cell and in some sense the genome as this Swiss watch-like mechanism. My wife pointed out to me that the more I give the stock, the fewer people in the audience are going to know what this is actually a picture of. But this is something that we used to wear on our wrist before like Apple watch totally redefined what that meant like. So I should probably replace this with circuits now. But anyway, let's go with this, because in the Swiss watch mechanism, even though the outer workings of the Swiss watch has changed repeatedly because of design, the actual inner workings of the watch have basically stayed stable despite decades of innovation of the other fronts. And that's partly because it's really difficult to make wholesale changes to pieces of the gears because they're so closely interlocked with other gears. And you could quite easily break the mechanism simply by basically swapping in pieces from other Swiss watch companies, for example. And yet this picture from the genome is also telling us that the genome is right with these pre-predator-like interactions that we are basically discovering by virtue of these evolutionary analyses. How could both of these be true at the same time? And the facts are that they're both true at the same time. So what we are trying to do is reconcile how could we explain these two facts simultaneously. And so my job today is to actually give you two vignettes of stories where we start off in this very perplexing space and we arrive at what might be satisfying sort of explanations for this unexpected rapid evolution. So I'm actually going to start with the final answer to a final Jeopardy question that you might face if you're ever on Jeopardy, which is what is the fastest evolving protein-coding gene in the human genome? And you might think that it's actually an immunity gene or something involving xenobiotics, but it's not. It's this gene called Protamine 1, which is absolutely essential for male fertility. And I'll explain what Protamine does in a second. This is the sequence of Protamine 1 on the first line from humans and the amino acid sequence of the Protamine 1 from chimpanzees. There's 13 amino acid differences here and no synonymous changes. This is by far the fastest DNDS signature we have in the human genome. Protamine 2, if those of you are interested, is also essential for male fertility is number three on this list, right? So the two protamines, both essential for male fertility, very, very rapidly evolving. What are protamines? So most of you are aware of what histones are. Histones are, of course, the defining unit of chromatin in eukaryotes. And most of our cells in our genome are packaged in histones or variants of histones, and that's kind of how they're defined. But in many animal species, during the course of spermatogenesis, that histone-rich genome is basically replaced by a protamine-rich genome, and you do that because you need to package the genome even more tightly into space so that it can actually fit inside the sperm head. Because ultimately, that's dictating what the fertilization success of sperm might be. So you go from essentially a histone-rich environment to sperm which have very few histones, not completely replaced, but very few histones. And then after fertilization, you strip off the protamines and put back the histones because you need to go and re-enter the cell cycle. These protamines are themselves quite mysterious. So for example, there are frog species that lack protamines and frog species that have protamines. And you can see that they make a dramatic difference in the shape of the sperm head because they essentially change the streamlining, if you will. You can see that the sperm on the left would be like a me trying to run a 100 meters race with the Usain Bolt sperm on the right. Because streamlining means it's going to be much faster in terms of swimming and potentially fertilization success. I should also point out that there are very charismatic organisms, like zebrafish, I have to say that because Sean sitting in the first row, that never undergo a protamine transition. And yet you're perfectly good and sexually mating. So the reason this has been quite interesting, but only sort of mildly so, is because we've really had a very good hypothesis for a very long time to explain why these proteins that are completely sensual male fertility are nonetheless very rapidly evolving. Because it really maybe has to do with sperm competition. After all, the nature of the protamines and how they package this sperm genome really maybe dictated to their fertilization success. Even then you sort of realize that, okay, you might read some optimum and stay there, but the fact that you're relentlessly evolving suggests that there is no optimum to do that. There are some chinks that begin to appear in this armor. The first chink that appears is zebrafish, right? Because zebrafish has multiple mating and yet doesn't need to do protamines. We also have species that are completely monogamous. And yet they have protamines, even though they've lost all of the other genes related to sperm competition, and their protamines continue to rapidly evolve. Almost suggesting that the competition that is happening is maybe not at the level of sperm competition, but perhaps even earlier. That was sort of like a first kind of signal. The second problem is actually a much more severe flaw in the hypothesis, which is those of you who've been paying attention would realize a lot of these are transitions two or away from arginine. And the arginine content of protamines is directly related to their ability to package sperm genomes. But these arginines are almost all CPG codon bearing arginines, which means that they're also highly hyper mutable, given that the CPG dinucleotides are also subject to high mutation rates because of methylation, which means that what we might have here is a strange mutation selection balance where you need a lot of arginines to package DNA. But those arginines are actually very labelled because of their nature of mutation and maybe that's what's leading to this acceleration. And the field cannot falsify this null hypothesis, which is an unusual but new null hypothesis. So we were stuck here for a while partly because there's no way to divorce the two until we thought we are going to take a look at Drosophila species because Drosophila species lack DNA methylation completely, but they also have protamines. So we can begin to assess do their protamines rapidly evolve and what might be the forces that drive their rapid evolution. I should also point out that even though we refer to all of these as protamines, that really refers to the class of sperm chromatin proteins. Their evolutionary origins are completely different. So Drosophila protamines are derived from these HMG box containing transcription factors which is completely distinct from what vertebrate protamines are derived from. But they go through the same stage where you start with a histone rich environment and over the course of sperm genesis replace almost all those histones with protamines. So I should also give like an extra kudos and I do that every slide not just because I'm speaking to the long read genomics crowd here that a lot of the work that we are doing would not actually have been possible five years ago because a lot of these genes and genetic elements I'm going to talk about reside in elements that are in highly repetitive regions of the genome, which we had some access to through short read sequencing, but that's a little bit like assembling a jigsaw puzzle with very small pieces and you've all done that and realize how hard that can be. And then you suddenly finish the puzzle and you realize that there's completely puzzle pieces missing like your three-year-old has gotten at the puzzle box and like things are like disappeared or you have weird things that shouldn't fit in the puzzle because you've got maybe some puzzle pieces that shouldn't belong there, but long read sequencing really helps us kind of solve this problem and in the Drosophila community there's been this considered effort from this group that I've highlighted here that has really made long read sequencing possible for nearly a hundred species of Drosophila at various levels of completion and that really sort of allowed us to do our analysis to a much better extent than we were able to previously. So first answer is let's look at all of the protein-coding genes in the Drosophila genome. This is primarily based on a melanogaster comparison. How do the protomines evolve? It actually turns out that most of them are among the fastest-evolving genes in Drosophila genomes just like they are in mammalian genomes. So the tendency for protomines to be rapidly evolving appears to be true and this has actually been borne out in a couple of other species as well. But this analysis also reveals something quite unexpected whereas all mammals except one encode two protomines that have remained constant even though they're rapidly evolving, Drosophila protomines appear to be under very strong rapid turnover which means that amazingly no two Drosophila species have exactly the same complement of protomine genes. So the top gene is our focal species Drosophila melanogaster. You'll notice that Drosophila melanogaster is about 15 protomine genes but they are not completely identical across different species where the black boxes refer to intact open reading frames, the X's refer to pseudo genes and if there's no box or it's missing that means there's no evidence that that gene ever existed in that species and probably has been lost. So you can actually see that there are very few black boxes that are extending across the Drosophila phylogeny. Most of them appear to have been gained or lost frequently over the course of evolution. In particular I'd like you to draw attention to the fact that some of these genes are fertility essential which means you knock them out and single gene knockouts those males are sterile. They're not capable of basically fertilizing eggs correctly and some of them are clearly fertility non-essential. Some others are actually uncharacterized and we are hard at work to try to characterize them but already there's like some amazing sort of things to conclude here which is genes that are fertility essential in Drosophila melanogaster are not the genes that you'd expect which are completely conserved of the course of evolution. In fact they are the young genes that have recently arisen that have become fertility essential and some of the genes that are completely conserved through all Drosophila evolution. In fact there's only one it's right at the right-hand side of this diagram we knock it out it has no measurable effect at all on male fertility. So this paradigm that we sort of all assume which is what is deeply conserved and retained at evolution is probably more important really does not appear to hold true at least in relation to protomines. So age is not really a good surrogate for importance you know age or retention. Yeah they're all randomly distributed I should also point out that all of these are autosomal but they're completely distinct no two genes are together if they're together that would appear with that number in the box like that two in Ficusfila. But there are arrays and I'm coming to arrays which are very interesting because what Ching Ho who did this work also discovered was that there were these amazing examples where protomines which are ancestrally always on autosomes had jumped on to sex chromosomes and when they did so had undergone this rapid amplification often to copy numbers of like 12, 13, 50 whereas the original gene was present in only one copy on the autosome. And so this is just showing what that happens so you can see in the Drosophila melanogaster we have one gene that jumped on to the Y chromosome and expanded out to 13 copies in simulans a different gene jumped on to the X chromosome shown here in red and expanded to 37 copies collating all the data from all of the genome sequences we had we found 19 independent cases of this protamine expansion always on the sex chromosome but very few of these were shared with more than one species. So they're occurring recurrently but they're not very long-lived when they occur in the genome. So all of these things began to make less and less sense with our way of sort of thinking about fertilization success and we realized that the way we have been thinking about protomines as sort of really this protein that is essential for genome packaging into the sperm head and the primary way we think about it is maybe not the driving force for protamine evolution maybe the alternate hypothesis is when protomines package genomes they also pretty much silenced the genomes sperm genomes are basically transcriptionally silent and maybe that's the more important function because when they do so they might actually prevent selfish genetic elements from like hijacking the process of meiosis and jumping around. So Ching Ho sort of actually arrived to the lab with this hypothesis that he wanted to really test so what are the types of selfish genetic elements that could thrive of course transposeable elements are like first in mind but male meiosis actually provides also another opportunity for selfish genetic elements in which the elements can manifest following meiosis where both the red and the purple chromosomes that should have an equal opportunity to become sperm do not because the red chromosome might harbor a gene a toxin gene that prevents the purple chromosome from undergoing sperm eugenesis correctly and when they do so that basically gives that red chromosome a huge advantage because even that this male is heterozygous it is only going to transmit red chromosomes through the next generation so this kind of selfish behavior is really good for the red chromosome really bad for the purple chromosome as you can imagine but actually it's really bad for the rest of the genome because the rest of the genome is paying a 50% cost in male fertility because of the selfish action of just one chromosome this is the definitional sort of you know this is the by definition what selfish genetic elements do which means that the rest of the genome is going to try to evolve suppressors of this toxin to try to restore myotic parity this might seem like science fiction but this happens all the time in drosophila melanogaster we have three separate myotic drive systems this happens in mice we estimate that there's probably lots of these discoveries to be made in populations especially where we can discover them with large effective litter sizes this is a lot more severe on sex chromosomes you can imagine because where as myotic drivers can actually even drive to fixation on autosomes they can never drive this fixation on the sex chromosomes they'll always be this dichotomy between the recombinationally isolated X and Y chromosomes and that also has an immediate effect that can be seen by natural selection because if you have a sperm killer gene that arises on the X chromosome to kill the Y chromosome this will instantly alter the sex ratios of the offspring now regardless of your like sociological views on the topic it is actually genetically favorable to have equal numbers of males and females for the proper genetic transmission of all alleles in the population which means the rest of the genome including the autosomes are now going to suffer the consequences of having this female bias offspring and they're going to try to do their best to try to suppress this sperm killing behavior and there been a number of these kind of sperm killers and our suppressor the sperm killers that have been discovered and we were really sort of surprised and encouraged because we realized that through two different independent studies that there was one instance where a protamine derived gene had itself become a sperm killer and a different protamine derived gene had become a suppressor of a sperm killer so you would think what's different here like what is why is the same gene capable of being both a sperm killer as well as a suppressor and it all comes down to location if you're paying attention you'll notice if you're on the sex chromosome it is in your evolutionary interest to actually become a sperm killer because you can out compete if you're on the autosome it is in your best evolutionary interest to become a suppressor because you want to suppress and that basically entirely explains this dichotomy between these so if you were playing God and you had this conflict between X and the Y chromosomes what is the kind of genetic experiment you would love to do well one experiment I would love to do is let's just fuse the X and the Y chromosomes or most of them because if we did that anything that the X is encoding that kills the Y or vice versa is going to be a suicidal gene and will be culled out by natural selection amazingly of course this happens very rarely in evolution in fact prior to our study that has only been one instance of an XY fusion in an island population of wolves and you can see why that might be because there's all this hidden variation for X killing the Y and that is probably instantly quite deleterious but if this were to happen not only did we lose the sperm killers but we would also potentially lose the suppressors of the sperm killers because there's really no reason to keep them around anymore since there's no sperm killing happening because of this sex fusion so those are the two predictions that we were sort of motivated by that we would make the driver suicidal leading through their loss but then we would make the suppressors of the drive obsolete why am I telling you this because those of you with very keen eyesight have realized that there's actually one species of Drosophila we found which had actually lost most of the ancestrally conserved protamine genes this gene called this species called Drosophila kikovai and motivated by this we actually went to the other members of the montium group that includes Drosophila kikovai right here at the bottom and we realized that in fact they have repeatedly and ancestrally lost their protamine genes and the reason we think they did that is because most genes that are present on the Y chromosome which is paternally inherited are now actually maternally inherited so that's happened because the X and the Y chromosomes have actually fused in the common ancestor of the montium group the only the second ever instance detected in animals where we think that XY fusion happened rendering the drivers obsolete and then rendering the suppressors of the drivers obsolete basically making a very strong case that protamine genes which are otherwise highly conserved can be completely dispensable if you simply fuse the sex chromosomes arguing that it's really this conflict between the sex chromosomes that's keeping these protamines in the genome in the first place so if you if you're sort of paying attention here what we are telling you is that there's a radically new way to think about what is potentially driving the presence proliferation and turnover of protamine genes which is they are suppressors of this inherent competition that goes on between sex chromosomes all the time if these genes jump on to the sex chromosomes they actually completely switch sides and become drivers themselves even though they ancestrally started off as being suppressors and if the sex chromosomes fuse like they do in the montium group you can actually lose the protamines without any detriment because natural selection has no reason to keep them around anymore so this is basically the explanation that can provide both an explanation for why protamine genes turn over because you're constantly dealing with new driver systems but also why you can actually completely lose them in in one fell swoop I will also point out that the same genes that tend to jump onto the sex chromosomes and drive are the genes that got lost in the sex chromosome fusion so we think that there are two classes of protamine genes one that is involved in genome packaging the canonical thought-off function but some that are actually involved in competition between the sex chromosomes so I hope your disbelief quotient has like risen to high levels at this point because of course we were very excited about this but also very dubious so we really want to ask the question how do we go about testing this hypothesis for a radically new function for protamine genes and so to do that we actually turn back to one very cool tool we have in Drosophila is that we do have these essential newly risen protamine genes that are rapidly evolving so I'm going to tell you about just one of these genes even though we've done that for three genes now this gene is called MST 77F this simply refers to the 77F is the cytogenetic location in the Drosophila genome only fly people would like appreciate this but there is one gene that is present on the autosome and there are 13 copies present of the Y chromosome so we simply replaced this gene which is by the way completely fertility essential with the Y chromosomal equivalent right asking well can the Y chromosome complement this male fertility essential function and can the genes from the other related species also compensate for this male fertility essential so in other words ask the very simple question how has rapid evolution changed the function of the MST 77 so what we found was mixed when we did the knockout of course the males completely lose fertility so that shows that they are not capable of producing progeny it turns out that the Ananasia MST 77 is too divergent to recover fertility we were never able to recover fertility interestingly the Y chromosomal gene which is genetically most closely related to the Drosophila melanogaster gene is also not capable of restoring fertility but the glass half full is we have rescued trans genes from melanogaster simulans in Yacuba so you know at least it partially worked and when we tested the sex ratios of these progeny we found something quite dramatic which is that in the melanogaster replacement case we got a 50 50 male female transition ratio but as you get more and more divergent like with Yacuba you now have 66 to 33 male to female transition which would be actually very deleterious if this would actually a trait that occurs in the in the population and Yukiko Yamashita's lab actually ended up doing the final experiment in a separate paper where they over expressed the Y chromosomal trans gene in a wild type background and what they found was that also skews towards a maleness because of course this gene is on the Y and probably favors the transmission of the Y chromosome. So what we have here is the ability for these trans genes to rescue fertility but not for them to maintain the optimal sex ratios that are needed in these species arguing that this mismatch that is occurring has been related to the sex ratio transmission. Psychologically what's going on so these are sperm nuclei that are at the final stage of intervalization all of these are brother sperm this is one cyst all of these sperm were produced from one kind of meiosis so they're all produced exactly at the same time these are kind of beautiful needle shaped sperm that basically show this equal levels of condensation which is exactly what you'd expect because they were produced in one developmental step this is what we see with the male rescue trans gene but with the Yakuba rescue trans gene we see something quite different now we have cysts that are producing two types of sperm at the end of spermogenesis some sperm that have completely finished apparently this process of condensation and some sperm that are one step prior at a level of condensation that is basically not sufficient to be packaged into the sperm head and we think that these are cellular signatures of myodic drive where in some cases the Y chromosome bearing sperm have basically finished and are ready to go whereas the X chromosome bearing sperm are basically stuck at that decondensed state and not capable of fertilization and this is exactly what we would predict for a condensation defect associated mutation. So I'm going to in the interest of time actually switch to another story which just again highlights the coolness of using sort of genetic conflicts as a potential explanation for what might be going on in the genome and this one actually has some is something that is very near and dear to my heart as Meru pointed out I've been actually working on this for like 20 years and and sometimes the retort is why haven't you published anything for a while so I will just tell you this is one of those things that keeps kind of giving you more and more fruit. This is involving a process of chromosome segregation which is a process that's absolutely critical in most eukaryotes in multicellular organisms like us we go through a million such mitosis on a daily basis it's really important that all of them get policed correctly so we do not have aneuploidy that occurs during the chromosome segregation process and in most cases the process is heavily dependent on chromosomal sites that are basically associated with microtubule attachment called centromeres and centromere associated proteins so what we discovered in this analysis that centromere associated proteins are among the fastest evolving genes in the genome completely in contrast with their completely essential functions in chromosome segregation. I'm going to spend all my time telling you about one of these centromeric proteins called SenPA so whereas most of the genome is actually packaged in canonical core histones shown here in green at the centromeres the histone H3 molecule is actually replaced with a histone H3 variant called centromeric histones or SenPA and the reason I'm forcing this sort of dichotomy onto you is that whereas core histones are actually among the most slowly evolving proteins in the genome so the amoeba histone H3 and the human H3 are only two amino acids different despite a billion years of divergence the centromeric histone H3 variants are among the fastest evolving protein-coding genes in many different taxonomic groups including primates including Drosophila including plants so what's going on what are the processes that could be driving this so we considered a number of possible examples we of course first considered mitosis the one process that we are all thinking of when we think of chromosome segregation although this is a very very important cell biological process from an evolutionary standpoint this is largely a very boring process I hate to admit it because you're basically going to end up with daughter cells that are genetically twins of each other so discriminating one from the other is actually going to be very difficult at least that's what I thought for a very long time until a very enterprising graduate student came to my lab and pointed out that there are these things called extra chromosomal plasmids that do rely on plasmid segregation and we do need to think about conflicts and in fact that has you know started a whole different research program on two micron plasmid segregation in yeast but I won't talk about that today because this does not actually affect the centromeres themselves because it affects a different form of partitioning we've already talked about male meiosis and male meiosis could be a very good explanation for centromere competition and evolution except all of this competition is actually happening after meiosis is already complete you're actually distinguishing between the products of meiosis to decide who lives and who dies so this is not going to be a very good explanation of the rapid evolution of centrometic DNA and the rapid evolution of centrometic proteins which brings us to my favorite process which is female meiosis which by the way has arisen independently seven different times during the course of eukaryotic evolution where once again you end up with four haploid products of meiosis but you don't have to invent an elaborate killing mechanism to get rid of the competition because only one of these products is going to be selected to be chosen as the oocyte nucleus and transmitted to the next generation the other three are going to be discarded as polar bodies so you go through that whole process you make four meiosis products but you only going to choose one to transmit so now the entire game is not about eliminating the competition it's about outracing the competition to whatever that preferred position is that gives you the maximal chance of propagating to the next generation so we imagined that if the purple chromosome were able to recruit more microtubules for example via their centromeres they might be able to orient themselves to the innermost position of this tetrad and therefore end up into the oocyte nucleus at a higher than expected ratio based on Mendelian frequency now what's really insidious about this process is unlike in male meiosis where you drop 50 percent of the sperm you are going to produce exactly the same number of eggs here you're simply genetically skewing the composition of the eggs to transmit yourself which means it's actually very hard to detect female meiosis unless you're actively looking for it and you don't need to do a 100 to 0 ratio even a slight deviation or Mendelian frequency is enough to lead to a basically fixation over just a few generations so what's really cool about this is this is cheating that is actively occurring during chromosome segregation you're actively taking advantage of chromosome segregation to a cheat which is why we think that this is a process that could drive this nonetheless it still leads to the question what the hell is going on like this is a fundamental process of chromosome segregation why would you allow yourself to be subject to this kind of drive and this leads us to point out that unlike in budding yeast the definition of what becomes the centrometic region shown here in these cartoon diamond examples i apologies to adam and all the other people from his lab you know these are think of these as alpha satellites or whatever your favorite centrometic satellite is they could even be identical across the centrometic microtubule attachment region versus the surrounding pericentric heterochromatin but what dictates where the heterochromatin sits down versus the centromere is really dictated by where the heterochromatin proteins and the centrometic proteins are basically so this is very much an epigenetic definition and can be influenced by mass action which means you can over express heterochromatin proteins artificially and bump into the centrometic domain encroach into the space which was previously centromeric and you can actually do the reverse which often happens naturally in some cancer cells where you over express centrometic proteins and encroach into heterochromatin and potentially even lead to dysentrics so in this dynamic sort of centromere heterochromatin compartment steve hannikoff and i propose that any kind of change at the underlying centrometic dna level either by mutation or by recombinational expansion that allows you to recruit more centromeric proteins to yourself is going to give you an advantage in female meiosis basically all you need to do is take this highly repetitive satellite array and arrive at a configuration that makes you a much more attractive landing site for centromeric proteins relative to heterochromatin proteins it's also very important that once you do that you have a selfish advantage over the other chromosomes in the population you are never going to revert back to that and selfish days biology doesn't work like that because you become toast in this competition so the only way to restore the status quo is for the proteins to themselves alter their dna binding preferences such they are no longer capable of being hijacked by this sort of potentially selfish centromeric dna so what's kind of interesting is that the first cartoon and the third cartoon are identical in terms of their centromere heterochromatin boundaries and you might be tempted to conclude that nothing interesting happened here because after all the centromere is exactly in the place it should be but you would have missed an intense episode of adaptation first at the dna level and a second episode of counter-adaptation at the protein level that restores the paradox but this is the definition of a red queen conflict you're running hard just to maintain the status quo of what the centromere boundary is you can in fact make the analogy between this system and a parasite that invades the host population by learning to hijack a host protein and the only way for the host protein to essentially regain fitness is by moving away or evasion from this binding affinity in this case the parasite happens to be centromeric dna which by the way is completely essential and the immune equivalent happens to be the centromeric protein which is also essential and is a collaborator with centromeric dna in the process of chromosome segregation yet the asymmetry of female meiosis provides an opportunity for these previous essential collaborators to become combatants in this genetic conflict so what's kind of interesting even though this also sounds a little bit like science fiction is that all of the first steps of the model have now been beautifully and elegantly proven by work actually done by Takeshi Akira who I don't know if people know he's at the NIH now in Mike Lamson's lab who showed that indeed expansions of centromeric dna satellites in mouse robot sony chromosomes can actually lead to over recruitment of centromeric proteins and lead to transmission bias favoring yourself but what has remained uncertain is okay if you do that what is the deleterious consequence of that why do you need to restore the status quo at all why do the centromeric proteins need to evolve in spite of the fact that they're completely essential and so that's the process that we became very interested in asking is why what is the sort of driving force for this centromeric protein evolution and so to do this we basically turn to my favorite model organism which is Drosophila and this is work that has been carried out by Ida de la Cruz my lab manager and a former technician Emily Colo now keep in mind I've been telling you all along these are very rapidly evolving proteins but I hope you all appreciate it I mean that not on a virus infection timescale but on a geological timescale which means that you know I'll have to wait like several graduate thesis equivalents to wait for the next amino acid mutation to occur which is very much frowned upon in the at least in the fred hutch in terms of how long this would take so we are not going to wait for the next mutation to occur instead we decided that we are going to revert the current day version of centromeric histone in Drosophila to an ancestral version hoping that that will unleash the deleterious effects and reveal what drove the rapid evolution in the first place right so you are basically in this two-step dance between DNA and protein we are forcing the protein backwards hoping that that will actually tell us what is the deleterious consequence that occurred here so this is a series of experiments we call reversing SID evolution where SID stands for the Drosophila ortholog of centromeric histones the reason it works quite nicely is because we have all of these sequences from Drosophila so we can recode them all to be beautifully encoded and translated in Drosophila that's what the big R stands for and by taking advantage of all of these sequences we can even reconstruct mathematically what the ancestors of these sequences were so for example two and a half million years ago before the melanogaster and simulans lineage is split they must have had a SID gene and we can completely tell you what that SID gene was like because we can reconstruct that from present day sequences so that's the first step we can completely mathematically tell you what has happened with SID evolution the second step came from the magic of CRISPR-Cas9 engineering where we can basically bring in all of these evolutionary related orthologs just like with msd 77 that i showed you before and bring them into the endogenous SID locus replacing the SID locus remember SID is completely essential so if you knock it out and don't rescue function you're going to be dead so we replace them with three versions the melanogaster version which is identical at the amino acid level the simulans version which is 21 amino acids different that's 10% divergence by the way in just two million years just giving you a sense of how rapid evolution might be and the ancestor which is about halfway so what could possibly go wrong like you know you could think about well it could be mitosis that goes wrong it could be meiosis perhaps male meiosis or it could be sentiment identity inheritance through sperm or egg that gets screwed up and we need all of these things to work in order to get fertile and viable offspring so to do this first sanity check we basically crossed flies so my pointer is not working so i'll just point out that these are egg cartoons of flies which have basically got a cytoplasm which is chock full of proteins that you need for cell division because essentially you need all of the supplies you need for early rounds of cell division because you're not going to get any transcription and the early embryo and then the sperm actually has very little centromeric histone because that's on the sperm centromeres right so there are three sources of centromere proteins sperm chromatin oocyte chromatin and oocyte cytoplasm and when we do this replacement with melanogaster we're able to show that you can get perfectly good offspring that are in Mendelian expectations so our re-engineering whatever we did it didn't screw up said function in a big way but the first cool experiment comes when we take heterozygous individuals and cross them to each other so now we have three types of progeny possible melanogaster flies with only melanogaster seed melanogaster flies that are like parents melanogaster and ancestral seed or melanogaster flies that are incapable of melanogaster seed only produce ancestral seed and amazingly we got all three flies but with the melanogaster we got them in mendelian ratios and we got them in lower and lower ratios the more divergent we got so remember these flies are completely identical except at the seed gene except at these positions that are rapidly evolve and what we're seeing is the fitness consequences at the organismal level of mismatches with the seed gene this is for the fly aficionados i don't know if there are any perhaps there are some listening online what we do here is actually leading to a delayed development so this is actually having a profound consequence on the development of these flies even in the heterozygous state so much so that if you try growing these at 18 degrees which by the way delays development even more these flies are dead so the the combination of delayed development because of seed incompatibility and the combination of delayed development because the temperature becomes a temperature lethal what's going on we realize that actually the defects were visible very early in embryogenesis and what we can take advantage of is that in drosophila early embryogenesis occurs in a common cytoplasm right basically everybody is in a syncytium all of these white dots in the periphery represent individual nuclei that will go through this completely synchronous rounds of division and when they do so they kind of stay there right they basically continue and that until they're produced enough nuclei to progress to the next stage i hope you notice that there were very few nuclei that left the periphery and ended up kind of in the middle now if you look at the right embryo which is again identical except for one of its copies is the ancestor you again go through these synchronous rounds of division but now you reach a stage when checkpoints kick in and suddenly it's like somebody announced there's pizza outside and all of you leave the door that's basically what's happening because all of these nuclei were deemed to have some mitotic defect that basically got marked and then once that was marked they basically released you don't block mitotic divisions but you mark the nuclei that are destined to not participate in development here's another way with a completely different marker on the left again we've got the melanogaster said again going through these synchronous rounds of division when you move in and out and i hope you can see with the simulans every time you go through this you're creating this trigger pressure that is releasing nuclei kind of into the middle because you're losing the anchoring to the outside of the embryo and losing that kind of thing the result is you reach here and thanks Sophie for where you basically reach a stage where with the melanogaster you have this beautiful ordered array of red marked nuclei but with the ancestor or simulans said you have these gaping holes where there should be nuclei but there aren't because you've not made enough nuclei that have survived this quality control checkpoint the result is on the left you have these beautiful embryos with the melanogaster i know that only fly people appreciate that but these are beautiful denticle patterns mouth parts and on the right you might not know why they're screwed up developmentally but i hope you can appreciate the screwed up developmentally they'll be dead within another few hours of imaging they weren't even able to crawl out of the vitilin envelope so huge viability defect that we can see here but this is glass half full again we can test what happens to their fertility right we have now adult flies that somehow magically have made it so we can ask what happens if you take male flies with completely the wrong sit and cross them to females that was my favorite hypothesis for nearly a decade you take sperm which have completely the wrong sit my prediction was that this male would be sterile because of some incompatibility and that would have been awesome if that was true it was completely not true it was one of those irritating things where you do the experiment over and over because it is exactly not what you predicted these males are perfectly fertile so we said ah it's got to be the other way around where you take female flies with the wrong sit and cross them to males that has to be wrong because now the 10000 fold excess protein is of the wrong type and these are also fine so like my two favorite hypothesis and in one series of lab meetings were completely destroyed and this was like life was pretty depressing for a while because these are males and females appear to be fertile but we can now cross those perfectly fertile males and females to each other which we haven't up till now and when we do that we recover no progeny so instead of the hundreds of progeny that we get with the male rescue trans genes we basically get less than a dozen with the ancestor and zero with the simulant so this is our most penetrant phenotype that at least one of the two parents needs to supply the correctly matched version of the melanogaster sit protein this is actually the very best evidence of the incompatibility even more dramatic than the delayed development and I should also point out we can completely rescue this if we have the right rescue trans genes so all of this is because of the incompatibility with the sit gene so why the incompatibility we think that this is both an incompatibility at the centromere specification effect which is why this acts at early embryos but it is also an ongoing chronic incompatibility in centromere function which is why we had that delayed development effect is the incompatibility at the centromere DNA that's what we favor but we've not ruled out the possibility that it might be just a protein protein interaction incompatibility between the centromere histones and other proteins of the kinetochore I'll leave you with this one kind of really mind-blowing result now given everything I've told you remember this cross with the wrong sit parents completely fails but this cross with the only parent that has the sit protein is the male parent and it only provides in a diminishingly small amount of the sit protein is perfectly able to rescue early development which is really a dramatic example of whatever that is happening at the epigenetic level because remember there's no transcription happening at the early embryos is nonetheless able to rescue this so all of after all of that work I have shown you it turns out that my least favorite hypothesis turns out to be correct which means that the incompatibility that is probably driving the rapid evolution of centrometic proteins is not actually occurring in myosas at all it is occurring subsequently in the next generation in early embryonic mitotic divisions that occur in the absence of heterochromatin in the answers of checkpoints at very rapid speeds and that's where those incompatibilities show up now I'll point out that the male miotic drive that I talked about before the instant deleterious effect of male miotic drive is male fertility so you want to rescue it there with female miotic drive actually there's no drop in female fertility it actually turns out that what you really need to rescue is the subsequent generation mitosis so you're actually one generation removed from the deleterious consequences of the miotic drive itself so I think I will just leave you back with this cartoon just reminding you that there are definitely two views to the cell and the genome there's lots of interesting stories that remain to be told where both the swiss watch and the pre predator interactions appear to be alive and well thank you very much I've been told that if people use the mic that people online can hear them if there are any questions I have to ask this of course no where I'm from you see any sign of haploid nuclei in the heterozygous crosses you so you may yeah so excellent question Daniel do we see any signs of haploid nuclei which would indicate like a gross effect we haven't gotten that far but we are looking for why chromosome loss which is very easy to assess and that's what they I will say that we have done some cut and tag to map where these proteins are going and for the most part they seem to be going to the same spot except on the fourth chromosome so we think that the incompatibility might be related to the dot chromosome and it makes sense because the fourth chromosome centromeres are actually very divergent from each other across these two species as well yeah the smoking gun evidence is to see lagging chromosomes we are working very hard at that but we haven't gotten that data yet sharp great sort of a half-baked idea but is it possible that you allow these arms wars to go on because it's actually creates opportunities for speciation and that's actually advantageous in terms of long-term survival to create subspecies that can then further adapt yeah so this is an idea we've had a long time ago that incompatibilities can drive speciation we don't necessarily think about speciation as an adaptive phenomenon in fact Darwin himself I'll just show the quote here he was actually mystified as to why speciation occurs at all because essentially what you're doing is you're actually reducing the reproductive partners that you can actually have in the population so even though you can exploit niches you might be able to do that by assertive meeting much more efficiently than reproductive barriers but actually just to change a question slightly we do think that these reproductive barriers are an inevitable consequence of these arms races that you're going to if you're fundamentally altering these chromatin proteins you're going to create incompatible consequences and we and others have actually shown that in drosophila speciation genes there's a heavy preponderance of satellite DNA binding proteins that show up repeatedly as the incompatibility here so yes go ahead hi I have just I was just curious that whether the other centromeric proteins are highly evolving or is it just a centromeric histone variant so maybe the other proteins are not rapidly evolving just to tolerate you know the changes in only the centromeric histone variant yeah that's a very good question it totally depends on which species you're looking at so in drosophila centromeric protein centromeric histone is the one that is the most rapidly evolving sem-c is not and cal-1 the chaperone is not but that's not true in mammals or in plants where sem-c is the most rapidly so it actually more is a function of whichever protein gets the right mutation to like that's off to the races and that's where the adaptive evolution is likely to occur as you know that there is a sort of a semi-equal relationship between senpa and sem-c deposition with cal-1 and drosophila which is distinct from the more hierarchical deposition and so even the chaperone age herb is actually rapidly evolving in mammals and so we think that it's not restricted to just one protein it just happens to be whichever protein bears the brunt of what the first adaptive change and that's off to the races there yeah and my second question is that do you think that this evolution will happen during the lifespan of a certain individual or an organism like because as the heterochromatic organization is changing do you think that this even you've got your aging hat on that's a very good question i mean certainly the mutations arise de novo and they're like a germline mutation-driven process so it's possible that even within the lifetime of the organism you can explore mutation but in a given organism you're probably exploiting very very small subset of mutational space necessary at a population level certainly it's actually happening and the more intense the deleterious consequence the more intense the fitness of recovery as well thank you fantastic talk um i had a question about natural variation so um in the like the pertamine copy number expansion that you see on the y-cremism i think if you look at different individuals within the species are there any individuals that have more or less pertamine copies and then does that affect the transmission and then relatedly for the sent bay you said it would take several graduate students these days to wait for the next mutation to arise but actually if you look across the population and do you find any naturally occurring variants yeah both excellent questions so the first question i think it's i think we are currently at the point where technically it's much more addressable than it was in the past but we haven't done that these are all based on reference genome sort of assemblies but it's a fantastic question um and certainly worth looking into for the ones that are up to 50 copies like seeing a difference of plus or minus five copies may be more challenging but the ones that are 13 that'll be much more possible um with SID and melanogaster that has actually undergone an adaptive sweep so there's very little standing variation in simulans there's more standing variation but since our focal species of melanogaster it leaves us in this kind of experimentally weird ground where all of the interesting standing variation might be in the other species so we're trying to sort of graft some of those variants into melanogaster right now but we don't have any expectation that there'd be a functional recovery there yeah okay is there somebody monitoring the online michael is there any online question no okay thank you very much thanks