"According to NASA's Definition of Life, I'm Not Alive" - Why Nobody Can Define Life | Dr. Kate Adamala

0:00That just shows you that there is really no good definition of life. In your world, life is not clearly defined. I think NASA has a working definition, self-sustaining chemical system capable of Darwinian evolution. That's a fantastic definition, but according to that definition of life, I'm not alive. The fact that we don't fully understand life right now doesn't mean to me that there's something that we're unable to understand. It just means we lack data at this point. You can create spud cells that feed on carbon in the air. Is that what you're talking about? I'm mostly talking about the molecules.

0:36We need a way to make all the molecules that our civilization uses right now. If you put molecules under the right conditions in the right environment, they will start self-assembling, and the emergent property of that assembly is what we call life. It's definitely a milestone, but it's not a mic drop. We're not done. We're showing that you can escape this gravity well of evolution. Could you introduce yourself to listeners, give some of your background? I know that you've been working in this area for a while. You had CINELs, I think they were called, before spud cells. If you could give some of that background and, as I said, state what the goal of this research is.

1:22My name is Kate Haramala. I'm originally from Poland, where I studied chemistry. Then I got my PhD in biophysics in Italy, in Rome. And then I moved to the States for the other half of my PhD in origin of life and biophysics of origin of life. And then I did a brief stint in neurobiology, synthetic neurobiology as a postdoc. That taught me a lot. One of the things that it did teach me is that I don't like neurobiology and I don't want to work on that. So when I started my own lab, I wanted to continue the practical applicability of research that I picked up during my neurobiology work. I wanted to know that my research, the things that I'm doing are actually good for something.

2:12But I was really drawn to go back to my roots to work on something as cool as origin of life, as astrobiology. So I basically wanted to combine that curiosity-driven research that I've done in my early training with the practical, biomedical, economical applicability of the work that I learned during my postdoc. And that led me to synthetic biology and specifically synthetic cell engineering. And the overarching goal of my research is to make biology a general purpose technology. Right now, biology is very specific. We can make a lot of things with biology, but there are very specific things. And general purpose technology is something that you can pick up and apply to a new direction, new application, new purpose without the need to re-engineer the whole chassis.

3:06Like I can take my work computer and start watching dog videos anytime without having to rewrite the operating system. And that's the kind of a philosophy that's the purpose that I want to apply to biology, to moving atoms with biology. And natural biology is not a general purpose technology. And I don't think it, I don't believe it can be because natural biology is very complicated. It has a lot of dependencies that we don't understand. And that led me to synthetic cell engineering. If you want to make a cell that can be reprogrammed on demand, that can be applied to whatever application you want to making medicine, diagnostics, making molecules for bioeconomy, that needs to be a platform that's fully engineerable.

3:55And you do not have full engineerability on a system that you cannot fully describe. And that's what natural cells are. They're amazing, but they're not fully describable. They're not chemically defined. And that's where synthetic cells place themselves. They have some advantages of biology. The fact that we can take energy and take feedstock and turn it into whatever atoms we want, whatever products we want. It replicates, it grows, it can make more of itself, so the costs go down. But we have this engineer ability. We can exactly say, I want to make X pathway with Y product. That's fascinating.

4:38And so it's not, you know, in artificial intelligence, the motivating question for a lot of AI researchers is to understand or how does the brain work? What is consciousness? All of that. So in your work, it's not to understand how life begins. There may be some insight into that, but you're more interested in how can you manipulate biology to produce products. Both, actually. There is a very strong curiosity-driven aspect to my research. I would really like to understand what makes molecules tick, because life to me is a complex behavior of molecules. I don't think life as a phenomena has anything magical to it.

5:38I think life is just a manifestation of properties of molecules, and we don't understand right now how that happened. How did those molecules became what we call alive? So there's definitely this quest to understand how non-living molecules become life. And there's also this quest to understand both healthy and diseased states of life, our own cells. How do we make a full model of a healthy cell, of a sick cell? How do we differentiate between those two? How we understand those processes better? And all of that, I think Richard Feynman said, what I cannot build, I cannot understand. And all of those processes of cells, everything that makes us what we are, cannot be fully understood unless we get our hands on every molecule that makes that possible.

6:31So there definitely is this foundational curiosity driven aspect of my research. But the main big picture motivation is that I want this planet to be habitable in 50 years. I want to leave something, you know, I want our kids to have a world to live on. And that's not, I think that's not possible unless we switch to moving all the molecules with biology instead of dead biology, petrochemicals. So that practical aspect is the main everyday driver. But there definitely is that childhood curiosity of what actually life is, how do molecules make life? And those both are combined in the work we're doing.

7:17Yeah. And just for listeners, you know, as in AI work, understanding and consciousness are very amorphous terms that nobody can really nail down. And in your world, life is not clearly defined. I think NASA has a working definition, self-sustaining chemical system capable of Darwinian evolution. And that definition, that's a fantastic definition. But according to that definition of life, I'm not alive. Yeah. Because, yeah. And that just shows you that there is really no good definition of life. That's right. Although you're talking about the reproducibility, but you're defining a single organism as opposed to a class of organisms.

8:23And certainly reproducibility is a feature of the human race. It may not be of every organism within the human race. spud cells are not self-sustaining and they don't truly evolve. So are you leaning on the absence of a definition to keep the question open and you're looking for a definition, or is that beside the point? To me, it's beside the point because I think there is no clear boundary on the molecular level between life and non-life. I think life and non-life are opposite ends of a continuum. Just a sec, Kate. I'm sorry. Yeah? I'm sorry to bother you, sir, while you're on the pottle, but did I hear you asking for potatoes?

9:20No. Well, regard. I'm recording. I'll talk to you afterward, okay? Of course. Apologies, sir. won't happen again. Carry on. I'm sorry, Kate. Go ahead. To me, it's beside the point because I think there is no clear boundary on the molecular level between life and non-life. I think life and non-life are opposite ends of a continuum. There's a spectrum of complexity, spectrum of organization. And we can see that even Even in our kind of instinctive definition, if you look at a human, if someone cuts off my head, I will immediately become dead by any medical, legal, human definition. But most of my cells will keep going for quite a while.

10:11And the cells in my gut, the cells that make me but are not me, the bacteria in my gut, will keep going happily long after the legal me is dead. And that just shows you life is not a strict, definable phenomena. And I love those discussions for fun over beers, but it's not a scientific problem that we really need to solve, or I don't think it's even solvable. What's fascinating about this work is that you have created, and I'll ask you to, I've read the paper, I've read a lot of the articles about the paper. if you can give us kind of a brief, give listeners a brief for what has happened. You have this mixture of purified chemicals or molecules, and you have this mechanism that forms a lipid bubble around them, and then you can introduce nutrients and ribosomes.

11:22that then can work on the DNA inside the cell, or what you're calling a spud cell. Can you just describe what exactly is going on in simple terms, and those that want a deeper explanation can read the paper? I think you described it really perfectly. So what we have is a mixture of purified defined molecules. And those molecules are derived mostly from living sources. They're bacterial proteins or proteins from other sources that were expressed and purified from bacteria. And they're small molecules that, you know, to me, they come from a shelf, from a bottle on a shelf. And they end up in that bottle because someone purified it from mostly a natural source.

12:16So they're all biologically derived. And we mix those molecules at certain concentrations under the right pH in the right solution. And they start rebuilding some of the elements of life, the translation, protein translation, metabolism restarts when you put those molecules together at the right conditions. and then we take this we load it up with a genome so a dna plasmids that contain the genes that we want our cells to have and then we stuff all of that into a lipid vesicle so a lipid bubble that's very much like the membranes of natural cells and that's what a spud cell is and what happens once we start that experiment is that the spud cell is capable of feeding so it eats nutrient-rich little liposomes or exosomes.

13:11And then it can also divide. It can make daughter cells. It's always daughter cells for some reason. It's never sun cells. It breaks a whole bunch of little daughter cells. Yeah, although that part is an external mechanism that you apply to the spud cells. I mean, I understand it grows with the nutrients. It absorbs these nutrient bubbles. They fuse into the skin of the spud cell, and so the skin of the spud cell grows. Those nutrients are mixed into the inside of the cell and metabolized. But when the cell is grown to a size that you want it to split into two cells, you do that mechanically. That's not a natural process.

14:14No, the part of the paper, the results in the first few figures are with mechanical division. But then the later in the paper, we have genetically encoded division. and I think every figure is labeled which one is mechanical and which one is genetically encoded. There is a genetically encoded division and that's why people get excited about this because this is a first example of not only genetically encoded feeding but also genetically encoded division. There is a mechanism where a spud cell expresses a protein and that protein goes to the membrane and pokes its head out of the membrane. And when that protein is present, it recruits a giant protein from the environment that induces curvature of the membrane and that leads to division.

15:04So even though the giant protein comes from media, their recruitment, which is the key step to bring that protein and start division, comes from the generic activity of the spud cell itself. This is a little bit off topic, But it's something I wondered for a while. You know, Michael Levin, who does the work with the flatworms? Yes. I saw him speak years ago, and he manipulates the electrical field inside and outside and between organelles or whatever they're called inside the flatworm with different solutions, with different pHs. is that happening inside of Spud's Bell? Is there any manipulation of the electrical activity or is this purely something else?

16:11It's purely something else. We do not right now rely on electrical activity. I wish we did. It's one of my wish lists to make spiking synthetic cells that have electrical potential that we can manipulate. But right now, there is no electrical activity that we manipulate across the membranes. Yeah. And from what I understand as well, the two most critical elements this bud cell needs to replicate or to grow are the ribosomes and the transfer RNA. Right. And you introduce those from E. coli bacteria, is that right? That's correct. We purify those from E. coli and then feed it to the spun cell. And what would it take?

17:08Obviously, this must be on your wish list or roadmap. I understand that it's very complicated and that's not the problem you're working on. But how would you create the DNA inside the cell to build its own ribosomes or transfer RNA? That's a great question. that actually is a problem we're working on. This is my personal next big step that I want to work on is how do you make ribosomes? Introducing the DNA that we need to make ribosomes is actually not a big deal. We can express all of the proteins that we need to make a ribosome and make all the tRNAs. The problem is, though, that if we express all those proteins, put them together, they do not form an active ribosome, and nobody knows why.

18:21And that's, to me, right now, if you ask me to name single, there's many, but if I were to pick a single biggest mystery that we need to solve in bioengineering, I would say this process called ribogenesis, how do you make new ribosomes, is absolutely on top of my list. We can put together all the proteins, and that's been done before. And those proteins, unless you really coax them with a lot of external help, they will not even approximate an active ribosome. George Church's lab demonstrated a few years ago now that it's possible to take all of those purified proteins and under very special conditions, once you purified all of them, you can put them back together and some ribosomal activity will be detected.

19:10But it's not a robust process that you can do as you're expressing them inside a synthetic cell. And I think we're missing something. We're missing either some protein, some small molecule, or some environmental factors. And that's what my lab is working on right now. There are two directions you can attack this from you're doing bottom-up, where you take the purified molecules, mix them in very precise quantities or ratios, and they start interacting. I was asking about the electrical potential. I mean, what starts them reacting? Maybe you can talk about that. But the the ribosome the other way to attack this is by taking a living cell and and removing as much as you can and still have it active i think that was done you know in 20 years ago or something that was on my correct yes when you look at a ribosome in a living cell, are there any insights into how that is forming?

20:26Is it being constructed? Are the instructions for building ribosomes in the DNA? They must be. Yes. And the way it happens in living cells is that it happens in stages as those proteins and RNA are being made. So it's not like everything is first made and then staged and then assembled. It's being assembled as you're making it. And I think this is the step we're missing, is we don't know exactly in what order those things need to be assembled, but we do know that you can't just let it all wait there, sit, and then assemble. You have to keep assembling it as it's being made. These are artificially created, do I call them vesicles?

21:20What do I call them? Vesicles is what we call them, yeah. Vesicles, okay, that have nutrients and the ribosomes and tRNA inside them, and that fuses with the outside of the cell, but it fuses locking onto a docking tag that is expressed from within the cell. That fascinated me, that the cell does create the mechanism, I mean, even though you're introducing artificially created vesicles that have this material in it, the fact that it can fuse with the spud cell is a function of the spud cell's DNA or internal mechanisms. Is that right? That's correct. The feeding is induced by basically the cell's ability to express the proteins that ask for food.

22:29So it's like a little bird opening his mouth. That's what the spun cell is doing, essentially. Can you talk about, is this purely atomic bonds being formed between molecules? I mean, what is the spark that starts all of this happening that a lot of people ultimately would call life, that it starts happening spontaneously if you put all of these molecules in the right quantities in close proximity? That's the beauty of it. It starts happening because of the properties of those molecules. And that's why I said earlier that I don't think there's anything special about the phenomena of life. I think it's a property of molecules.

23:23If you put molecules under the right conditions in the right environment, they will start self-assembling and the emergent property of that assembly is what we call life. There is nothing special about a spud cell or really any other cell that does not come from properties of atoms that make it. And the fact that we don't fully understand life right now doesn't mean to me that there's something that we're unable to understand. It just means we lack data at this point, which on one hand, you might think it's kind of a lax, the mystery and romance of what life is. But to me, that is the ultimate romance.

24:05That is the ultimate mystery, because that means the universe is posed to give rise to life. The exact physicochemical conditions of the universe enabled creation of the molecules that eventually are predisposed to become life. That's, to me, that doesn't get more magical than that. How does a real, a living cell feed? Is it through a similar mechanism, these protein tags that capture? It depends. The process that the spud cell uses to feed is very similar to phagocytosis. It's a process where a cell kind of engulfs a piece of food, another cell, or it could be an exosome. And that's one process of feeding.

24:51Most cells, if they're predators, they feed by basically engulfing another cell. Many other cells feed by just uptaking nutrients, molecular nutrients directly from the environment. Spud cell does that too through its membrane pores. There is an optimized growth solution for it. So it's one of the mechanisms. Mutation that you introduced allowed some spud cells to feed more aggressively and grow larger and eventually out-compete the smaller ones. That mutation was introduced to the DNA. Is that right? Yes, by me, artificially. I wanted to clarify, yes, we introduced those mutations artificially, and that's why I claim that spud cell is not capable of evolution right now.

25:46Even though it is capable of selection, if you give it a better mutation, that mutation will win after a few generations, but I have to give it that mutation. It doesn't arise spontaneously. So that's why, to me, that's not evolution. Cells are dividing, whether or not it's mechanically or internally triggered. If they're dividing, I mean, you were saying in the paper that only about 30 % of the divided cells, the daughter cells, carried complete DNA. Is it possible if, I mean, mutations or errors in the replication of DNA, Is it possible if you did this at some monstrous scale that you would start seeing spontaneous mutations?

26:37Absolutely. I think if you scale it up enough, you would start seeing spontaneous mutations arising in that population. It's not a mechanism we can rely on right now just because we don't have infinite resources and infinite volumes. That's why I would like a better evolvability, better evolution mechanism. So the evolution, you're inducing it, but that doesn't mean that it wouldn't happen naturally if this were in a primordial sea, you know. Yes, if you had the scale, if you had the volume of a primordial sea and a few million years, it would totally start evolving. It just, you know, I work on the scale of a length of the PhD, not length of the primary evolution.

27:27So we can't really wait quite that long. Is it possible now with these increasingly powerful models to run this computationally and see what happens after, you know, a few trillion iterations? I think it's absolutely possible. We're working with some collaborators to develop that model. You basically, as you know way better than me, in order to get anything useful out of AI, the quality of your training data has to be really top-notch. You have to have a lot of high-quality, clean training data. So that's what we're working on right now is making models that will help us speed up that iteration and evolution process a little.

28:14You're working toward this goal. Was the spud cell surprising to you, the success of this? Or have there been so many iterations that you finally got to this point and it's incomplete? and there are a lot of things that it doesn't do, but at least you're here and it's just a milestone on a long road. I mean, how do you feel about what you've done with Spud Cells? I feel incredibly proud of our team and I feel very happy that we got it to this point. I feel both surprised and not. I'm surprised because I'm always surprised when experiments work. There's so many failed experiments in this field, like in any experimental science, that when something finally works, it is in a way a pleasant surprise.

29:06But I'm also not surprised because we designed it to work. We built the whole thing from scratch. And it's definitely a milestone, but it's not a mic drop. We're not done. We're not going home yet. It's the big milestone that shows what's possible. all. That's why we're making naming analogies to Sputnik. We're showing that you can escape this gravity well of evolution. You can put molecules together into a system that starts making lifelike noises, but it's not a complete system and it will not hopefully ever be completed. I think this is a platform that will always be changing, that will always be evolving into something better.

29:51You were saying at the beginning that, well, certainly understanding how life arises from molecules is, you know, one of the goals. The other is to create this kind of chassis on which you can, biology as a workhorse in medicine or material science. today the spud cell does a marker protein and you call it a glowing marker protein first of all what does a glowing marker protein mean it's a fluorescent protein so a glowing as in it it it makes fluorescence it's it's a gfp it's that workhorse of synthetic biology a green fluorescent protein originally from jellyfish. So the fact that the spud cell can make this, how big a jump is from that to it making a useful protein drug like insulin?

31:03It's quite a bit of a jump, but less than it seems like. Insulin in particular, it's tricky because you have to post-translationally process it. Other drugs will probably come first, high-value drugs that use different building blocks. There are some antibiotics, there are some protein therapeutics that use amino acids that are not canonical, so not one of those 22 amino acids that all organisms put into proteins. And spud cell right now could be programmed to make most of those. The bottleneck in scaling that up so it actually becomes a useful therapeutic is the spud cell's inefficient replication, because programming it to make something is relatively easy.

31:53Programming it to make something at scale is hard. And you're not going to make money on a drug unless you learn how to do it at scale. And that's why we're pushing for better metabolism for replication now, because that's what really is going to enable applications. And the idea is, so it's not actually expressing the protein inside the cell, it's having the cell reproduce efficiently, and is that right? It's the bottleneck, yes. That's why our discovery is so significant, because we show that it's possible to make those cells grow and replicate. And now the big goal is to make them grow and replicate robustly, enthusiastically, I would say.

32:44And then you could have the plasmids or the DNA inside the cell expressing a drug protein of some sort. That's correct. That's the goal. And then what would you do? You would have a vat of spud cells, and then you would dissolve the liposome exterior, and you'd be left, you know, separate what's left, and you would be able to purify. Wow, that's... Or you could convince, even easier, you could convince the potato to give out its product. The spud cell could be programmed in a way that it would secrete the product into the media, and then it would be much easier to separate it. I see. Yeah. Yeah.

33:35So plastics and materials are very different because there's a whole multi-step enzyme string of interactions that need to take place, not just one gene expression. Has anyone shown that's feasible in a system like that, or is this just conceptual? People have shown that polymers can be made in that cell-free system that's the cytoplasm of spud cell. As far as I know, nothing of economical value has been expressed in a synthetic cell yet, but that might mostly be because we haven't had the thing grow and replicate, so scaling it up is just ridiculously expensive. And what do you think about, I mean, from what you understand so far, what do you think led to life?

34:38I mean, as you said, maybe it's not as mysterious as people have historically thought. Do you think that the conditions were correct on Earth? they're all of these chemicals and you know molecules in in the air and in you know the the the water or whatever medium was covering the earth and that these molecules just spontaneously found each other and started interacting and then grew a membrane as protection or something and that it just it just happened that way rather than there being as I said some I don't know I guess we're all thinking of Frankenstein you know the lightning coming down I mean there's some electricals.

35:36I think definitely electrical impulses helped with the origin of life processes because they gave energy for certain reactions to happen. But that's still a property of matter. The fact that we had an atmosphere that was capable of producing electric discharge is a property of matter. It's a property of that atmosphere. Yeah, yeah. I'm sorry to take the magic out of it. No, that's what I love about this. I mean, as you said, it's still magical. It just is understandable. Your paper argues that a spud cell can't survive outside the lab because it needs chemicals it would never find in the wild.