Scientist Stories: Jack Szostak, The Origin of Life Not as Hard as it Looks

Good evening. Can you hear me?

Welcome to the latest in the series of Eyring Lectures. The series is named in honor of Professor Leroy Eyring, who was an ASU a Regents Professor, and Chairman of the Department, and contributed importantly during his tenure as Chairman to the, the growing reputation and excellence of the department. Today's speaker is Professor Jack Szostak. Jack is a University Professor and Professor of Chemistry at the University of Chicago and an Investigator at the Howard Hughes Medical Institute Prior to this year, Jack was an employee at Harvard University and at Harvard Medical School. He had multiple appointments at Harvard and carried out all of his duties in a distinguished fashion. His training was initially at McGill University for his bachelor's degree, and then he moved to Cornell and worked with Re Wu on his PhD thesis subject. Thereafter, he started his career all in the Boston area, had a number of positions over a period of years at the Farber Cancer Center, Harvard Medical School, and then at Mass General. And while he was there, he carried out a few things that some of you have probably heard about. So he, for example, worked on genetics and biochemistry of DNA recombination, which led to the double-strand break-repair model for mitotic recombination. In parallel, he made fundamental contributions to our understanding of telomere structure and function and the role of telomere maintenance in preventing cellular senescence. For this work and sharing it with two other individuals, he was awarded the Nobel Prize for his work. Most people would be satisfied with that. Jack went on to just as remarkable things in more recent years. I'll abbreviate it for this particular introduction since I have to introduce him again tomorrow, and I'll talk about it in more detail then. What, uh, in the 1990s, uh, Jack and his colleagues developed in vitro selection as a tool for the isolation of functional RNA, DNA, and proteins. Simple concept, but incredibly powerful for finding molecules of special activity, of function. Even though it's a discovery of a few decades now, it's still increasing in its utilization and the extent to which the world relies on that basic technique to find new and highly potent and selective molecules for specific utilizations. Since the year 2000, approximately, his research interests have focused on self-replicating systems and the origin of life, and this leads us immediately to the title of this evening's lecture, namely, The Origin of Life: Not as Hard as It Looks. Please join me in welcoming Jack.

Thank you very much, Sid, for your very kind introduction. Thanks for inviting me to come back to ASU and talk to people again. I've had a great day talking with lots of different people about all kinds of exciting new things. So in today's lecture, I'm going to try to give not a comprehensive overview of the origin of life, but sort of focus on some of the things that I find particularly interesting and and maybe try to give some idea of where we've made progress and what kinds of things have held us back from, say, maybe moving as quickly as we would like. So what you see on the slide here is just a schematic of our conception of what a really simple primordial cell might have looked like just a kind of a lipid envelope with some small bits of genetic information on the inside. And this looks simple. I mean, it's obviously a stripped-down version of a modern cell. It's got the important things except for proteins. But, you know, there's an enormous number of questions that arise if you try to think about how something like this could have emerged from the chemistry of the early Earth. And then even if you get to this stage, then what? How did it lead subsequently to the evolution of more and more complex forms of life? So I was talking to a few people before giving this lecture, and they said that it's always interesting for people to hear how you got into a field. So I'll try to bring that out a bit. I thought about different ways of beginning this talk. And I could have— everything, of course, worked before while we were setting up. But now I can't advance the slides. Okay, I can use this. Good. All right. Okay. So I thought about beginning by talking about exoplanets, because after all, that's been one of the just most amazing advances in science in the last two decades. We have a catalog now of over 5,000 planets orbiting different stars. There must, by extrapolation, be millions of very Earth-like planets out there in our galaxy alone, and of course we'd all like to know if there's life on any of those planets, or is it so hard to get to life that maybe we're the only place in the galaxy or even the universe where there's life? But I'm not going to talk more about that. I'm going to instead— I also thought, well, you know, I could talk about life being a far-from-equilibrium phenomenon, right? And it's— there's so many interesting questions. How does that generate local order of the kind that you see in life, like these incredibly complicated metabolic pathways? And everything depends on the input of energy into the system in the right way. But you have Paul Davies and Sarah Walker here who can tell you all about that from their work. So instead, I decided to talk about life from the point of view of information, because that's really how I got into this field and got interested and obsessed with it, really. So when we talk about information, it can be talked about in a lot of different ways. So this obviously goes back sorry, I'm having trouble with this— to, at the beginning, from Shannon, classical information theory. How much information does it take to specify a given string of symbols? And then that led to advances in compression technology, which I think mostly stemmed from Kolmogorov algorithms. Which is how much— what's the shortest algorithm or computer program that can actually regenerate a given string? But in biology, we don't really care so much about the actual sequence of symbols in a string. I mean, we all have about 3 billion base pairs of information in our DNA, and it's not really that much. Put it on your smartphones, right? But what's important about it is the parts of it that actually code for something useful. And even those parts, the exact sequence is not usually what's important, right? Because there are many positions in the sequence that don't matter, that can be anything. There are others that covary. And so the actual amount of information to specify say, a given functional tRNA or given functional protein is less. And even that is not exactly what you need or should be thinking about. In a biological system, and just like in a conversation between two people, what matters is the meaning of the string, right? And physical complexity doesn't quite get to that because there are many, many different ways of building different structures that do the same thing. And so that means, in effect, it takes less information to specify a given meaning in a biological system. Okay, so, okay, not to beat a dead horse, but there's this obvious parallel between the information stored in DNA and in long-term storage on computers and In biology, we have this sort of more transient form of RNA, which, by the way, also does all kinds of functions just like proteins and so on. And so it's very parallel to what we see in all these sort of built devices. And so I've been interested for a long time in how did the first organisms begin to accumulate information about their environment, about themselves, and about how they function in the environment? Where did that information come from, and how did it grow to the extent that we see in biology today? Okay, and so the way that I got into that was essentially by doing Darwinian evolutionary experiments on molecules instead of on collections of organisms. Darwinian evolution has been operating for billions of years and has generated all the diversity of life that we see. Humans have been doing directed evolution for thousands, probably maybe tens of thousands of years in animal and plant breeding. All that it really took to sort of extrapolate that from organisms to molecules was the technology of PCR. And so the experiments that we started doing were to begin with completely— a library of completely random sequences, a very large collection of sequences. And what we wanted to know is, how likely is it that a given random sequence can actually do something interesting? And as an experimenter, of course, you can define the task at will. It could be binding to a target molecule, could be catalyzing a reaction. I want to know how much information does it take to specify a functional molecule. And one of the reasons that I decided to to do this experiment was that by talking to different people, I got vastly different answers. Some people thought that, oh, you can get an RNA molecule to bind the target. It'll be like maybe 1 in 10 to the minus 5th random sequences. Others said, no, building a structure is like really hard, and you have so many factors to contribute. It'll be like 1 in 10 to the minus 50th. You'll never find one experimentally. So with 45 orders of magnitude of uncertainty, it seemed like worth doing an experiment to see if we could narrow it down a bit. So what we were able to do pretty easily was build libraries of on the order of 10 to the 15th different random sequences made in DNA, transcribed into RNA, and then take that set of sequences and subject it to a selection, so enriching for the ones that do what we want, and throwing away the ones that don't, and then amplifying those survivors with or without adding a little bit more variation, and going around and around this cycle, going around and around that cycle until the population is taken over by molecules that do what we want. Okay, and so the answer was that we could actually— this actually worked, and we could get molecules like the one shown here. So this is a sort of surface view of the three-dimensional structure of a short bit of RNA that folds up, and it makes this beautiful three-dimensional shape that has a little cleft on the surface that's complementary in shape and electrostatics. To ATP. So it combined ATP at a concentration that's biologically relevant, and this was at about 1 in 10^-10 of the random sequences that we started with. So actually not that hard to get functional sequences out of nothing, right? So then I got interested in taking this one step further and saying, Okay, this molecule binds with a certain affinity and specificity. How much harder is it to bind something more tightly? Intuitively, it's to do any task more efficiently, it's harder. And so it should take more information. The question is how much? And so to get at that, we embarked on a series of really painful and tedious and laborious experiments that were done by James Carruthers when he was a graduate student in my lab. And what James did was to select for RNA sequences out of this random pool that could bind GTP as a target. And he got a whole bunch of different solutions. And what I'm showing you on this slide are the simplest-looking solutions. They're just stem loops. The red bases have to be what they are. The blue ones can be anything. And so there's a certain amount of information in the sense of the Chrysodomy type of physical complexity that is required to specify each of these structures. But if you select— for tighter and tighter and tighter binding, then you end up getting more complicated structures. So these are a little bit more complicated. And if you keep going, you get these ones, which bind 100 times tighter. And they're much more complicated in their secondary structures and in the number of bases that have to be what they are. And so you can— actually, you can do the same kind of experiment selecting for catalysis. So this came out of an experiment that was done by Dave Bartell when he was a graduate student in the lab, selected for ribozymes, RNA enzymes, that could catalyze a ligation reaction. And again, he got more than one answer to that problem, and I'm showing you here one ribozyme that's fairly small and simple that does this reaction with a certain rate, and this much more complicated reaction that does a much better job. It's a better catalyst. And it's more complicated. You can see it takes more information to specify a sequence that does something better. All right, now can you quantify that? And well, sort of, right? So if you calculate the information required to specify all of those structures and plot it against either the affinity or the catalytic rate, there is a lot of scatter about this line. But you'd expect a lot of scatter. And the basic lesson is that it takes about 10 more bits of information to specify about a 10-fold increase in activity. And so this is a very old result. I've always wanted to improve on this, but the experiments back then, 20 to 30 years ago, were so hard that no one's ever wanted to go back and do this again. But I'm still fascinated by the idea of trying to quantify the relationship between information and meaning or function. We can also do these kinds of things with DNA and with proteins, and this is just a pretty picture of how you can use the ribosomal machinery to— as the ribosome is reading a messenger RNA and generating a nascent peptide chain, you can trick it into linking these things covalently. This allows you to select for function on the basis of what the protein's doing. But it drags along its coding messenger RNA, so you can tell what you got. And unfortunately, so this one lesson from this is, again, it's not that hard to find functional molecules, which is a good thing, because otherwise I think it would be very hard for life to evolve new functions, to become more complex, to adapt better to its environment. In a small-scale lab experiment, we were able to make new proteins that bind, for example, ATP, and other people have evolved catalysts this way. So I think these kind of give you some insight into how the first living systems got the initial information. The way we think about it is that protocells, like in that schematic I showed in the very first slide, could have been seeded with just random sequence bits of, for example, RNA or whatever the primordial genetic material was, probably RNA. And if you have enough of those, some of those sequences carry meaning intrinsically and could do something that would help that cell survive better or replicate better, take over the population. And that's the beginnings of Darwinian evolution. Okay, but to do that requires a few more things, right? It's all in the details. Like, okay, to do this kind of acquisition of information, you have to be able to replicate that genetic material, and that has to happen with errors so that new variation is possible. And it has to happen in a way that selection can occur. So how can all of those things happen? And so if you look at a modern cell, it's very hard to think about how that could ever arise because we have DNA storing information in an archival form, RNA as an intermediate, generating proteins that do all the functions in the cell. And the problem is that all of these things depend on each other, right? So you need RNA and proteins and metabolites to replicate DNA. You need DNA to encode the RNA. You can't make proteins without the RNA and the code and the archival information and the metabolites. So trying to think in a rational way about how a system like this could have emerged, I think, held back progress in understanding the origin of life for decades because people were generating all kinds of crazy theories about how everything could just emerge at once, which is basically not possible. And there's a clue in the structure of this diagram that maybe RNA, molecule in the middle is the answer to that sort of paradox. And that was recognized in the '60s by a few smart people like Francis Crick and Leslie Orgel and Carl Woese. More recently, the structure of the ribosome was solved by several groups, and I'm showing you here a picture of the large subunit of the ribosome. From work by Tom Steitz. And this is a huge and complicated molecular machine whose job is to make all of the proteins in our body. And if you look at this sort of face-on view of the large subunit of the ribosome, which is the one that actually catalyzes peptide bond formation, what you see in the very middle where this little green squiggle is, this is a— transition state analog that is occupying the catalytic center of the ribosome. All around that is only RNA, which are these gray squiggles. The proteins of the ribosome are these gold structures, but it's clear from the structure that the ribosome is a ribozyme that makes proteins. So all the proteins in all the organisms in the world are made by catalysis by RNA molecules, the large subunit of the ribosome. So logically, the only— the conclusion we draw from that is that RNA came first and that early life was a lot simpler. So we imagine primitive cells looking kind of like this— still cells with a membrane boundary. Inside, RNA as the genetic material and coding for ribosomes that carry out different kinds of functions, including the replication of the RNA. So now our problem of how to get to this point has become a lot simpler. All we have to do is figure out how to go from the chemistry of the early Earth to RNA molecules, and then figure out how they could replicate and develop to this state. So we can deal with the other complexities of modern life— DNA, proteins, metabolism— all of that you can think of as phenomenon that arose through the process of Darwinian evolution. They're evolved features. They don't have to be there from the very beginning. All right, so when could this have happened? So this is adapted from a review by Jerry Joyce. I think you can take all the numbers on the timeline with a grain of salt. Obviously, the Earth had to cool down from its molten lava surface to the point where you could have water on the surface, liquid water on the surface. There's evidence from the chemistry of zircon crystals that that was actually pretty early, maybe as little as 100 million years after the Earth was actually formed or after the Moon-forming impact. And then somewhere between that period and the earliest convincing evidence of life on the Earth, there's a period of about 700 or 800 million years. So somewhere in there, life got started. We don't know when, but there must have been chemistry going on on the surface of the planet that gave rise to all the building blocks of biology—nucleotides, lipids, sugars, amino acids. All made by just chemistry happening in different environments on the surface of the young planet. And then somehow that gave rise to RNA, which gave rise to replicating RNA, and eventually to the first cells, to the RNA world. And then the RNA world evolved the ability to make all kinds of complexity and adaptation. Okay, so then the problem is now simpler. I wouldn't say simple, but it's simpler. We can ask, how did this RNA-based type of life emerge from the chemistry of the early Earth? And you can talk actually to more than a few synthetic organic chemists, and they'll say that is just completely ridiculous. And I think that preconception that doing complicated multi-step pathways that give rise to a complicated and delicate product— the preconception or the bias that that's impossible has also held back the field for a long time. And it's only in the last, I'd say, 10 or 20 years that we're starting to actually address that problem head-on and think about, well, how can you overcome the problems, right? The idea that life emerged from this primordial soup where everything was mixed together and sort of magically assembled into cells is completely ludicrous, and people now are developing ideas, doing experiments, and showing that there are ways to break down long pathways into shorter subsets. There are ways of concentrating materials that you need, crystallizing intermediates. There's this buzzword of systems chemistry, which means thinking about things in a bit— a little bit of a different way. Instead of, you know, mixing A and B and trying to get only C, you say, well, what else had to be around, and can you make use of other things that were present to actually give you the products that you want. So what I would say is, so you've probably all heard about the famous Miller-Urey experiment, right? So Stanley Miller as a graduate student with Harold Urey passed an electric discharge through a reducing atmosphere, what they thought the atmosphere of the early Earth looked like, and found precursors to amino acids were generated. And so at the time, it was revolutionary. It said that making at least this subset of building blocks of biology is not that hard. And I think there was a period where people thought, well, maybe everything will be that easy. And there were a few examples that raised that hope again. So for example, Wanner Roh showed that you could take hydrogen cyanide and just reflux it in water, just boil it in water, and make adenine. Adenine is just a pentamer of hydrogen cyanide. And so people started to think, well, maybe all of the steps in making all the building blocks will be easy. And then things kind of stalled because people realized, well, in these reactions, in the Miller-type reactions, in this Oro-type of reaction, what you actually make is a mess, right? There are often thousands of products, and they're all present at parts per million or parts per billion. The more sensitive your techniques are, the more you see there. But the way to build biology is to make a lot of just a few compounds, just the ones we need. And so that's the real problem, and that's where progress has been made in prebiotic chemistry. In recent years. Now, this idea of cyanide as the ideal building block is something I really love, you know, partly because there's a certain irony in cyanide giving rise to life, but I kind of like that. There are lots of different ways to make cyanide in the atmosphere. We don't really know what is most important quantitatively. But it's pretty clear that you can make HCN, and it can rain out onto the surface. And my colleague John Sutherland and others have figured out very elegant ways to make use of cyanide to efficiently generate the right product molecules. But there is a problem. In the cyanide that's generated in the atmosphere can dissolve in raindrops and rain to the surface, but it's going to be very dilute. Very dilute cyanide doesn't do anything interesting. It just sits there and hydrolyzes to formamide and ammonium formate, and then you have nothing. So how can you actually make use of the cyanide that that would be generated in the atmosphere. And so John came up with this very elegant idea that cyanide could rain out into surface lakes, and if you're in a region where there's hydrothermal circulation of the water through fractured rock, for example, volcanic regions like, like you see in Yellowstone today, or impact craters, what can happen is this lake water will circulate, it will percolate down, it will be heated by the underlying magma reservoir, and then rise back up to the surface carrying with it ions that have been leached from the rock. And prime among those will be iron, iron-II, and iron-II reacts very rapidly and avidly with cyanide to make ferrocyanide. And the idea was that ferrocyanide salts could build up in concentration. They're a stable, very stable form of cyanide, could build up and precipitate out. And you could build up over maybe thousands of years layers of sediment highly enriched in ferrocyanide salts. And this idea has been refined very nicely by recent work from David Catlin. At the University of Washington in Seattle, who has shown that this actually can work very nicely in alkaline carbonate lakes, which are, again, a likely prebiotic type of environment. So I think this is kind of a major conceptual breakthrough. You start with something that's very dilute that you can't use. You figure out how to build up concentrated reservoir of material that at later times can be thermally processed and then can go on and do subsequent reaction. But going from cyanide to nucleotides, the building blocks of RNA, that's a complicated process, right? There's a lot of steps to go from something as simple as cyanide to something as complicated as a nucleotide. So how could that happen? And again, I think one of the real conceptual breakthroughs has been in realizing that some of the intermediates on these pathways tend to selectively crystallize out. And this is a particularly beautiful example. So this is a crystal from a paper by Donna Blackmond. This compound is ribose aminooxazoline, which is one of the key intermediates on the way to making all of the nucleotides. So the idea is that the reaction that generates this molecule generates the one we want plus a bunch of side products. But the one we want crystallizes out. We can build up a reservoir. We can wash away all the impurities. And then at some later time, when conditions change, you're ready to go on to the next step in the pathway. So I'm not going to go into any more detail on the prebiotic pathways because I want to talk about what came next. So back to how to make RNA, which we really have to understand because this is going to be the basis of primordial heredity and evolution. So we're going from cyanide. Through a pathway to making nucleotides. Nucleotides, when their phosphate is chemically activated, something I'll talk about a little bit later and more tomorrow, you can make short chains of RNA. And then making the complementary strands, you have a duplex, a double-stranded RNA, is a little bit more complicated, but that's a step that we've worked on in my lab and have made, I think, a fair bit of progress on recently. So can't really— uh-oh. Come on. Really? This is a really nice movie. I don't know. Any idea how to start the movie? I think one of the mysteries of science is you can test everything and it always works perfectly, and then it comes to the real thing and nothing works. Okay, well, we can skip this one, but the next couple I really need, so hopefully we can figure this out. Yeah. Okay, well, I can't show you the structure. It's a beautiful, elegant structure, and I think it's really incredible that you can actually assemble it just using chemistry and with no fancy evolved enzymes. So One of the huge advances, again, a conceptual advance in how to do that, came from the late Leslie Orgel, who's really one of the founding fathers of the whole field of prebiotic chemistry. And what Leslie realized was that thinking about how biology does this can be misleading. And so in biology, the molecules that are used by enzymes to make RNA and DNA are nucleoside triphosphates. So they all have this chain of 3 phosphates, and these are great substrates when you have an enzyme that's a really good catalyst, right? If you don't have an enzyme, if you're trying to let things happen spontaneously, these molecules just sit there and they don't do anything. And what Leslie realized is you need to make things that are more reactive. And she explored a lot of possibilities, and by the early 1980s had settled on these molecules like what you see here. These are phosphoramidazolides, and they're chemically reactive enough to start polymerizing all by themselves without an enzyme, and to some extent they can polymerize on a template and build up a copy of that. Now, We've had some advances in the chemistry recently that make this work better, and I'll talk more about that a little bit later and tomorrow. But first, I wanted to just mention another kind of preconception that I think held back progress for a long time. And again, it came from thinking about how does all this template copying, synthesis of long polymers? How does it work in biology? So in biology, we always have a primer that will bind by Watson-Crick base pairing to a template strand. And the enzyme comes along and uses these monomers as substrates and incorporates them one at a time to grow the chain, the complementary sequence. And so for a long time, everybody who was trying to do this kind of copying chemistry without enzymes figured it had to work like this. You have a primer, you extend it one base at a time. We thought, well, this is maybe too simplistic for prebiotic system or a primordial system. We thought of, well, maybe you can make it a little bit messier by having a whole bunch of primers scattered along the sequence and then extend them a bit, join them all together. That seems like a move in the right direction. But it turns out it's not enough. So, okay, here's another— really? I cannot seem to play any of the movies, which is really sad. Huh. Okay, do I— how do I do that? Actually, yeah, we can just turn off the laser pointer. Maybe we can— sorry about this. Oh, all right. Yay. OK, so this is how we were thinking about things when I started to work on this. So you imagine a little bit of RNA floating around in a really chemically rich environment, and activated monomers coming in, finding their partner by base pairing, and building up a complementary strand. So it seems easy and beautiful when you see these animations that were made by Janet Duasa, but when you try to do the chemistry, it turns out to be not so great. And there's a whole list of problems here with trying to get that to work experimentally. And what I can say at this point is that we've solved most of these. We're still working on optimizing a few, but I think we're pretty good now at doing copying chemistry. And so I'm just going to show you one thing about that that illustrates how a bias, a preconception, can get in the way. So this is the way copying a template actually works using these reactive building blocks. So we thought at the beginning we would have this— so this is a primer that's bound to a template. We want to copy this part of the template. We thought monomers would come in one at a time, sit down on the template, react with the primer, which would grow by one unit, and then we repeat that process until we built up a full copy. So it turns out it doesn't work that way. And so for the chemists in the audience, you can look at the right-hand side here. For the rest, you can get the schematic view here. What happens is that two monomers react with each other. They make this bizarre dinucleotide, this dimeric product. This binds to the template by two Watson-Crick base pairs. And what's really amazing is that the structure of the bridge, the linkage between the two monomers, is perfectly set up in three-dimensional space so that the reaction goes really quickly. And so it really took us a while to overcome this idea that copying had to work like it does in biology. But apparently, we now think that primitive copying worked in this chemically distinct way way. And only later, when enzymes, ribozymes or enzymes evolved, did we get to the modern way of doing things. Okay, now, that's good for copying a template. But what we need in order to do evolution is to actually replicate that we got to make— copy the template, then copy the copies, copy those copies indefinitely, right? And that turns out to be really hard for a lot of different reasons. And so one of the things that happened over the course of the pandemic was my lab was shut down for a while, like everybody's. And so I had nothing to do except stay home and think. And came up with this model that you see here called the virtual circular genome model, which we think is a way of actually replicating RNA sequences that overcomes all of the problems that had blocked us for so long. And so the idea is the genome of a primitive cell would be this green circular sequence. So this is an RNA sequence just shown as a line, but there's no actual circular molecules. The idea is you have a whole bunch of little pieces, little fragments, that map to the circular sequence. And the result is they can come together in all kinds of different configurations. And whenever they come together like this, you can extend one using the other as a template. So the difference is that now the act of replication has become distributed across the whole circular sequence. Working like it does in biology, where you start at the beginning, you go through the middle, and you get to the end. Here, you do a little bit of copying all over the place. It's everything everywhere all at once. And we're starting to test it. And I think it's looking pretty exciting. And figuring out exactly how to get this to work and under what conditions is one of the big things that we're— working on now. So if we can get that to work, then we should be much closer to being able to actually build protocells like you see here. So little membrane sacs that include bits of RNA, but with the chemistry to actually drive replication. So I haven't said anything— the membrane— that encloses us, and that also has to grow and divide. We want to have a primitive cell system. It's got to be able to grow and divide and do that forever also. And so we started working on that about 20 years ago, actually. And I was pretty nervous about getting into this field because I'd never worked on lipids, and the technologies for working with lipids are not as easy or how should I say, effective as a technologist for working on nucleic acids. But it's turned out to be a fascinating venture, and I've learned a lot of biophysics. And one of the nice things is you use microscopy to get beautiful images. So like what you see here is when a vesicle made of a simple fatty acid, oleic acid, you just shake it up in water under the right conditions, and it spontaneously makes these beautiful membrane structures, and here you can see smaller vesicles trapped inside a big one. They're really, really quite beautiful, but they also have the amazing property that they can grow in very interesting ways. And so I'm going to show you— here's another movie. This is one of these vesicles. It's encapsulating a fluorescent dye, which is what you can actually see. We add food, which is more fatty acids, and it grows like in this really unexpected way. I almost fell off my chair when I saw this the first time. This is a whole series of experiments that were done by Ting Zhu when he was a graduate student in the lab. So it turns out that if you have these vesicles that are kind of complex and they have multiple membrane layers, you can throw in more of the building blocks of the membrane, fatty acids in this case, they'll get incorporated into the membrane, which will grow and will grow in this unusual way into filaments, which we kind of understand but not completely. These are very fragile, so they can break down to smaller daughter vesicles, which can grow and the cycle can repeat. But then in more recent years, we came up with a a completely different way that this can happen. And this is work that was done by Stephanie Zhang, a graduate student, and Anna Wong when she was a postdoc in the lab. And here we're making much bigger vesicles, but they only have a single membrane. We start to feed them with more fatty acids, and what you'll see you look carefully, is that this— the spherical structures start to fluctuate, and they start to spontaneously divide. So it turns out we have two completely different ways now of driving growth and division that works basically very— with very simple physics, and with— it's just soap and water. There's nothing elaborate or fancy about this. It's the same thing. Okay, so some of the things that we thought were like super hard, I had no idea how we were going to get growth and division, turned out at least in some cases to be really easy. And whereas some of the other things like doing RNA synthesis and replication, we're still working on. Okay, so I think putting those things together is getting us closer and closer to being able to build in the lab a system that can grow and divide and evolve, which is what we really want to see. And if we can do that, that will be, I think, giving us some understanding, not necessarily of how life actually evolved, but at least how it could have evolved. So I'm almost done. I just want to end by saying a few words about the origin of coded protein synthesis. So a lot of people think that protein synthesis is universal in life as we know it now. And so it's obviously central, and it's important. So it would be nice to understand how a system that complicated could have arisen spontaneously by evolution. And the way I think of it is that there are kind of three major mysteries about this process. The one that almost everybody likes to think about is the origin of the genetic code, right? How did certain amino acids become associated with certain 3-letter RNA sequences? I'm not going to talk about that. I have no idea how that arose. But I think there are 2 other equally important questions. So the ribosome, that machine that builds proteins, uses very special substrates, right? It builds proteins using RNA molecules which have an amino acid attached to them. You could not evolve the ribosome unless molecules like that were already around. Why on earth would amino-isolated RNAs have been around? And so we think maybe they were doing something else. Before they were involved in protein synthesis. And that led to the emergence of perhaps ribozymes that made them more efficiently and with specificity so that a given amino acid would be attached to a given RNA. That would set the stage for the evolution of protein synthesis. The other thing we're starting to think about is the other requirement for evolution you couldn't evolve the ribosome again unless the thing that it made is useful. It has to contribute to the survival or the growth of the cell that it's in. And what could that function have been? We really have no idea. I mean, there are a few possibilities, but it's completely up in the air. Okay, so let me just— so it's just a diagram showing how the way the ribosome works, you know, you have RNA RNAs, tRNAs that have an amino acid attached, they come in, the ribosome hooks things together, builds the growing chain, kicks out the empty tRNA when it's done. So why would you have molecules like this around? Okay, so we think that these amino-isolated RNAs might have been doing something else. What could that have been? Well, it turns out that the amine of the amino acid, if it's not protonated like it's drawn here— sorry— is a good nucleophile. It's more reactive. It can attack activated phosphates. And it turns out that it can facilitate the assembly of little bits of RNA into more complicated structures, which can be ribozymes. So we think there could have been an early role for these molecules in building up structures, and then that kind of set the stage by having a selective pressure for making more of these amino-isolated RNAs. And then that chemistry was co-opted during the evolution of the ribosome to actually make proteins. OK. All right. So just to summarize, I think what we've actually learned about the process is that a lot of things that looked really hard, if you overcome your preconceptions, you can figure out ways that they happen easily. And almost all the big advances in this field have come from overcoming some kind of bias or preconception. Personally, I feel that this is a great field to be in because there have been so many surprises. And there's just been this enormous wealth of totally surprising and unanticipated chemical and physical phenomena that we never would have stumbled across if we weren't trying to figure out how life got started. And I think that we're close to maybe having a coherent pathway that gives us a picture of how things could have happened step by step where none of the steps are just incredibly contrived, for example. So that's at least the goal. I think we're getting there. So let me just end by— I've tried to mention some of the key people. My lab has had so many brilliant graduate students and postdocs over the years. Here are a few of the ones that have contributed recently to some of the advances that I talked about. So I'd like to thank all the people who've worked in my lab, contributed to this, all the people who've helped us, all of our collaborators and people who've supported the work. And they deserve a lot of thanks. And thank you for listening. I'm happy to take questions. Oh, please, if you'd like to ask a question, which is strongly encouraged, please acquire a microphone. Hi, hello. I'm an astrobiologist. My name is Jonathan. I was— you mentioned the isolated amino acid RNA and how they form. What was the— origin of that, or is that also a mystery? Like, of the— Okay, how did that amino acid end up? Yeah, because— To the end of the RNA? Yeah. Yeah. No, that's a great question. There are some simple kinds of chemistry that you can use to do that. I'm not convinced yet that we have something that the right one, the one that looks prebiotically realistic. There are a lot of different ways to do it. It's not a hard reaction, but we're still searching for something that looks more likely to have happened. And you also mentioned you were— in the many ways that broke the primordial soup paradigm, that one of them was the cyanide diluting in the water with the hydrothermal vents mixing, creating life. Is that like your main belief, or is that like— That's just one step, right? That's— it's the important thing is the idea that you can build up a reservoir of material and accumulate it and then use it later on. And that happens or can happen at many steps in all of these pathways. I only gave that as one example. But there are a lot of other examples. Thank you very much. Yeah. Where do you imagine the chirality coming in? Because we have the chirality of amino acids and the sugars and the lipids and so on, and they all had to become chirally pure at some point. So that could be a whole lecture. A few, like 20 years ago, it was this huge mystery that how could you end up with all the molecules of a certain type having the same handedness? And they're just, there was this, the Soai reaction, a sort of autocatalytic way of doing it. There were, no one ever figured out how to do that in a way that would get you the right molecules. There were other, many other ideas about how it could happen. Then I forget, maybe 5, 6, 7 years ago, Cristóbal Viedma, a Spanish geochemist, came up with a process called, at the beginning called crystal grinding, later chiral amnesia, now Viedma ripening. And the idea that if you have something that crystallizes to make separate right and left-handed crystals. And you put energy into the system by grinding them down. They get ground down to a finer powder that dissolves, recrystallizes on the bigger crystals, and there's a cycle. And that leads the crystal, the solid form, to diverge spontaneously to be all right-handed or all left-handed. Simple physical process. No one believed it. So counterintuitive, but then it was quickly replicated. Then, turns out, again, no, we can't figure out how to get that into the chemistry at the right stage. And now there's a whole lot of exciting work that's just coming out about how to use spin-spin interactions, magnetic field, to essentially to drive enantioselective nucleation of crystallization. And this works beautifully on exactly the right molecule. So maybe that's the answer. Yeah, it's— that whole field's been transformed multiple times. Just, um, thank you for your wonderful talk. So RNA is very negatively charged. Yes. So has that worked to its advantage or disadvantage? I think it's a really— one of the most important things about a genetic material is that it should be fairly uniform. You want to be able to have an arbitrary sequence because you need different sequences to do different things. And so to copy those sequences, you want to have the structure, the surface, and the chemistry be fairly uniform, independent of the sequence. And so having a highly negatively charged backbone, I think, helps with that. If you don't have that, you make molecules that have an uncharged backbone, they tend to just glom up into little three-dimensional blobs of goo. Which you can't replicate. So I think that polyanionic backbone is an essential aspect of the chemistry of RNA. There's an online question. Would you care to comment on the concept of the lipid world, the idea that the earliest organized structures were lipid micelles, and that the catalytic, catalytic entities were different arrangements of the hydrophilic lipid head groups? Yeah, so the way I think about it at least is that, yeah, the lipids are critical. They could be very simple, maybe just fatty acids, maybe more complicated. I am a little bit skeptical, but I could be totally wrong, about the idea of whether interactions between the head groups contribute to specific interesting catalysis. I think they're more important for making the bilayer membrane structure that encloses all this other stuff of the cell, and does it in a dynamic way that allows for growth and division, and as well as permeability to substrates and exit of waste materials. So I think it's those physical properties that are important, but what's missing so far is a really effective pathway for generating things as simple as fatty acids. It's one of the least explored areas of prebiotic chemistry. Hi. In one of your slides, you had a plot of the amount of bits it takes to describe a certain arrangement and the amount of effective like work it can do, or— Right. Yeah. I was wondering how you were— how you go about looking at a structure and giving a number for the amount of bits it takes to describe it. Sure. Good, good question. So what we do is we, by in vitro selection, we'll find a structure and then we can take that structure and make a new library in which we've doped every position with all the other possibilities. So if there's an A, we put in mostly A, but a lot of also some GC and U. Do that at every position. So now we have another huge library. Then we again, we select out the variants that work. And then we align all those sequences. And we can immediately see, OK, at this position, you have to be an A. At this position, you have to be a C. At this position, you can be any— at these two positions, you have to be able to make a Watson-Crick base pair. And so from that information, we can calculate the information content required to specify that structure. What is the difference between RNA and DNA? Yeah, so the difference is a small chemical difference. It's just one oxygen atom at a particular place on the sugar. It has very important consequences. And so DNA is very stable. It doesn't degrade very quickly just in water, whereas RNA is— it tends to fall apart. And you can ask, is that a good thing or a bad thing? Why didn't everything begin with DNA instead of RNA? There are some prebiotic pathways that might give us the building blocks of DNA at the beginning. So this is actually something people are still debating. Oh, hello. I have a question. So is there any biological evidence for your— proposed pathways? Like, is this special nucleotide exist in some ancient bacteria and still take part in some biological pathways? Some biological evidence like that? No, all of this early chemistry is kind of gone. I mean, it's only laboratory chemistry. Nothing like this happens in biology. Despite rather intensive searches people have made for the so-called shadow biosphere, small remnants of this earlier RNA world, nothing like that's been seen. And I think the answer, the reason for that is just that early life wasn't very efficient. It wasn't very well adapted to its environment like modern life is. And it was just— heaten up. So, and the conditions of the chemical conditions of the modern Earth are very different from the chemistry of the early Earth. For example, there was no oxygen, no free oxygen around on the early Earth. So the kinds of chemistry that could have happened would be very different. And none of this could happen now. So we just have to reconstruct what's possible from experiments in the lab. So thank you for your brilliant talk. I'm curious that when you were introducing the central dogma of life, which is the relationship between DNA, RNA, and protein, so you link DNA and RNA with one arrow. However, you linked RNA and protein with three arrows. So for that three arrows, are you Do you have a specific meaning? Yes, yes, actually the meaning is that the messenger RNA codes for the protein sequence, the ribosomal RNA makes it, and the tRNAs bring the amino acids to the ribosome. So RNAs have 3 different roles in building proteins. Another clue that life began with RNA. Thank you for your talk. I have a very general question about something that I never fully understood over here. What is the driving force for making things more and more complex over time? Like, usually we talk about systems that strive towards equilibrium. They want to be as boring as possible. But once we have an energy gradient of some sort, degenerate rat race, things want to be more complex and more efficient. How does this work? Yeah, well, there are many, many examples in which far-from-equilibrium systems generate local order, and that energy is dissipated through locally ordered structures. And biology is just one example of that. So all of this depends on different inputs of energy, chemical energy in the form of reactive molecules that we're still figuring out. We have ideas for how to use— thermal energy to drive some parts of the cycle, mechanical energy can drive division. There are many ways in which energy comes into the system. So all of this is a collection of far-from-equilibrium phenomenon that conspire together to generate order. Is it possible to describe it as a driving force for complexity? Is Is that some way to formalize it in some way? That's above my pay grade. I mean, I know that you need energy to drive the assembly of these systems. But yeah, advanced non-equilibrium thermodynamics is complicated. Thank you very much. I have a question. This is a good lecture to make us rethink where we come from. So my question here— Yeah, so it seems there's assumption most of the— this chemistry, like involving RNA happens in solution. So is it possible that it happens with assistance of solid surfaces, or kind of a closed compartment and hopping from surfaces? Yeah, yeah, that's a very, actually very interesting idea. So long time ago, Jim Ferris, and I think initially with Leslie Orgel, showed that you could assemble— that the surfaces of certain clays basically could catalyze the assembly of nucleotides into RNA strands. And unfortunately, it hasn't really gone beyond that. I mean, the idea persists, but experiments that would show something more interesting haven't worked out or been done. I think part of the problem is if you have an RNA molecule absorbed on a surface, it won't be in the right geometry that you need to copy it into a double helical product. And so personally, a little skeptical that minerals would play a direct role in, for example, replication chemistry, but they certainly could do all kinds of other interesting things. It is something we think about a lot. Where do amino acids come from? Cyanide. In the beginning, early on in the early Earth's history, there are very well-worked-out very high-yielding pathways to go from cyanide to not all, but most of the biologically important amino acids. Nowadays, they're made in a totally different way. They were made by metabolic pathways inside cells by enzymes. So how that transition was made from a purely chemical approach before there was life to this metabolic approach that we see now, that is, again, one of the big mysteries in this whole area. There are very few cases where we can see how something went from a prebiotic chemical reaction to a metabolic reaction.

Hi, I had a quick question. So when you were showing— oh, hi.

Sorry, the vesicle growth with the introduction of micelles, it had this weird elongated morphology. And I was wondering what drove that instead of keeping like a more uniform circular morphology. Yeah, we've wondered about that for a long time. I mean, I could tell you the theory that we have. I have no idea if this is correct, but the idea is if you start off with a vesicle that has multiple layers of lipid, and then you feed it with more lipid coming in from the solution, the outermost membrane layer is going to be the one that grows first. But there's not much volume in between it and the next layer on the inside. And so the result is the extra layer gets sort of And then over time, everything equilibrates to a multilayered filament. We can sort of see that happening in some of our confocal microscopy images, but I still think that's an area that needs a lot more study to really understand the biophysics. Thank you. So a lot of biologists I've met, they have varying estimates on how long they think it takes life to emerge. Some of them say a 1,000-year process, 2 million-year process, et cetera. So I was wondering, your proposed model here where RNA developed this way, how long do you think that takes and why? Yeah. OK, so the question of time scales, again, I think it's very interesting. And I think you have to— the time scales that are relevant depend on the process you're thinking about. So in that image that I showed, we're thinking maybe on the order of 100 million years for the Earth to cool down enough to have liquid water. So that's 100 million-year timescale. Once you've got liquid water on the surface, Okay, then maybe you need the right environment in terms of hydrothermal circulation, a volcanic region, impact craters. Maybe getting the right geophysical environment could have happened on the scale of millions of years or tens of millions. Once you've got the right geophysical environment, then the chemistry can start to happen. Maybe it takes thousands of years to build up a reservoir of cyanide in the form of ferrocyanide salts. Once that gets processed by heat, the derivatives that are made are much more reactive, and now the chemistry is going to start happening on a time scale of hours to days to weeks. Once you start building up nucleotides that can assemble and can polymerize, now you're talking about reactions that happen on a timescale of minutes to hours. When we get to an entire protocell system that can grow and divide, I think the timescale will be hours to days because there's a lot of different parts to that. So I think once everything was in place, things started happening really quickly, but it may have taken a long time to get everything just set up perfectly. What role do you envision for the extraterrestrial delivery of organic molecules and things like carbonaceous chondrites and so on, as opposed to the in situ synthesis of some of these building blocks on Earth? Yeah, yeah. So there's no doubt that a lot of carbon and a huge variety of molecules were delivered to the surface of the early Earth by the impacts of stuff from space, comets, asteroids, whatever. The problem is that those collections of molecules are mostly things we don't want, very little of the right molecules at very, very low concentrations. So I don't think they played a major role in the actual origin of life. I think what was much more important for that were the in situ, the chemistry actually happening in local environments that could give you a lot of just the right things. Thank you so much for your talk. Really fantastic. On the slide where you're talking about the various forms of energy that drive the production of cyanide, wondering specifically about UV radiation. And is there a sense of an optimal level of UV radiation to drive the production of cyanide and production of amino acids from there? Yeah, I mean, it's— so people have worked pretty extensively on photochemical processes that lead to cyanide production. It depends extremely on the chemical composition of the atmosphere and the wavelength of the light. So this is high, high up, and since there's no ozone, right, you have a lot of very short wavelength UV which can basically rip, rip things apart into atoms that can then come back together to make this triple bond species that are very high in energy. But whether that's— so the nice thing about that is it could be a— it's a kind of global process, whereas the cyanide that's made an impact is localized. So there's still a lot of debate about what's the best way, what would be the most efficient way to get a lot of cyanide in the right place at the right time. Yeah. Hi, so I'm wondering if after studying this for all these years, if you're convinced or have an idea if life is possible on other planets, whether or not you think that that's something that is out there. Yeah. So I used to say, we have no idea. That's why we're doing this. We're trying to understand if it's easy or hard to get from chemistry to life. And having worked on this for a long time, we still don't really know, but there's a lot of steps in the pathway that we used to think looked really hard. Right? We have no idea. And now it turns out they're easy. So it's kind of making me think that, well, maybe— I mean, there are gaps that remain to be understood, but maybe the whole pathway is actually not that hard. And if that's really the case, there ought to be life on a lot of exoplanets out there. But then that's a little scary, right? Because Nobody has contacted us. Either getting to intelligent life is really hard, which could be, or maybe the most likely thing is that life frequently does get to the point of intelligence and then it self-destructs, which we're very close to now. So we don't know. We'll see. Sorry to end on a down note. I have a follow-up to my last question. So if the origin of life, when conditions are right, is something you think happens on a very short timescale, why is it that we only, as far as we know, see one lineage of life here on Earth? Ah, okay. Yes, that's a different question. So everything that we see now traces back to a common ancestor. But that common ancestor, LUCA, the last universal common ancestor, was essentially like a modern bacterium. It had everything that modern biology has. And so we don't see, with very few exceptions, anything from before that point. So it looks like a single unique origin, but it didn't have to be that way. It could have been that life was popping up all over the planet, and, you know, pops up here, gets hit by a comet, you know, it's killed off, pops up again over here. It's possible that there were many independent origins and that, you know, cell fusion events resulted in the exchange of genetic information, horizontal gene transfer, right? Primordial sex, if you want to look at it that way, that would have speeded up evolution. But all of that period of history has been erased. So all we can really do is speculate and think about possible mechanisms. I was wondering about the primordial Earth and how I'm over here. And so you were saying it was all localized, right, based on different environments. So I was wondering, A, is there any evidence to see that ancient nucleotides, ancient amino acids, are they different on different parts of the Earth? And then second of all, what was kind of the driving force to have the standardized nucleotides we see today and the amino acids we see today? Right. OK, so two very interesting questions. So first, if you look for old rocks, the older the rock, the less there is of it. Because the Earth has plate tectonics, right? And so we're constantly losing continental landmass to subduction. And so there are only a couple of, 2 or 3 places on the Earth where you have rocks that date back to 3.8 or more billion years ago. And those rocks have by and large been fairly heavily processed. So there's no record of the early organic chemistry of the Earth. That is what is making Mars look so interesting, right? Mars never developed plate tectonics, and so there's a lot of really old surface on Mars, and maybe or maybe not there might be some residue of early— either early chemistry or early life, if there ever was early life on Mars. It's one of the main things that's driving the exploration of Mars at this point. And then, sorry, the second question was? Oh, yes, right. Yeah. Yeah. Yeah. So that's also a very interesting and somewhat controversial question. So there are some people who think that the chemistry just intrinsically goes towards the standard building blocks of RNA. Now, is that because people have spent the most effort trying to understand how to make those building blocks and they've neglected other possibilities? To some extent, maybe. But we have looked at a lot of alternatives, alternative chemistries. And this is the subject of my talk tomorrow. And basically, the summary is that RNA always wins. RNA just looks better than anything else that could have been made. And I think that's pretty remarkable. It's maybe too early to state that too strongly, but it really looks like the chemistry is just funneled towards making RNA. Maybe if we ever get to exoplanets, some other star, we'll find forms of life that have the same basic underlying chemistry. We don't know. So I'm also very curious about the third mystery. What is the function of the first peptide? The third question you had. What is the function of the first peptides that were made by coded synthesis? So, okay, so I think that you have to go back a step. And so there are many ways of making peptides.

Peptides.

It's not hard. What's hard is to make them in a coded way. So the question, at least the way we're thinking about it, is could there have been some simple peptides that were made in a non-coded way? Maybe just because there are some amino acids around more than others, they come together, they make kind of random sequence strands that are biased to contain certain amino acids? Could those do something useful? And if so, if there's something, any, a little bit useful there, then it would be worthwhile making more of it, right? And then you would have a selective advantage to start to build in coding for the system. Now what that primordial function is, we don't really know. We're thinking about peptide aggregation, phase separation, forming hydrogels. Other people are thinking about peptides that interact with RNA, stabilize structures. There are a lot of possibilities, but we don't actually know what the most likely thing is. I was just thinking in biomineralization, for example, for formation of crystalline— like you have IDPs, intrinsically disordered proteins, which allows, for example, not to crystallize. So you said how crystals are important, you need certain crystals, maybe not. Do you think that can be a function, like you have just tiny peptide just creating an environment, like this environment just modulates crystal formation or non-crystal formation for RNA? Could be, could be either, either of those kinds of things. We're also thinking about, you know, peptides like to interact with with membranes, and maybe primitive membranes had not only lipids but also peptides adhered to them, inserted. So yeah, what we need are experiments to actually show, like, okay, this is a way that a primitive peptide actually could be useful, and that's what's missing.

Additional questions? If not, Jack, thanks again.

We're absolutely grateful. Okay, definitely. Okay. Maybe they can edit out the parts where I couldn't get the movie to play. That would be, that would be good. All right, let's see. You agreeable? Yeah, sure. Yeah. Oh, I've seen Yeah, yeah, I was very happy about that.