Floyd E. Romesberg
1,725,019 views • 13:56

All life, every living thing ever, has been built according to the information in DNA. What does that mean? Well, it means that just as the English language is made up of alphabetic letters that, when combined into words, allow me to tell you the story I'm going to tell you today, DNA is made up of genetic letters that, when combined into genes, allow cells to produce proteins, strings of amino acids that fold up into complex structures that perform the functions that allow a cell to do what it does, to tell its stories. The English alphabet has 26 letters, and the genetic alphabet has four. They're pretty famous. Maybe you've heard of them. They are often just referred to as G, C, A and T. But it's remarkable that all the diversity of life is the result of four genetic letters. Imagine what it would be like if the English alphabet had four letters. What sort of stories would you be able to tell? What if the genetic alphabet had more letters? Would life with more letters be able to tell different stories, maybe even more interesting ones?

In 1999, my lab at the Scripps Research Institute in La Jolla, California started working on this question with the goal of creating living organisms with DNA made up of a six-letter genetic alphabet, the four natural letters plus two additional new man-made letters. Such an organism would be the first radically altered form of life ever created. It would be a semisynthetic form of life that stores more information than life ever has before. It would be able to make new proteins, proteins built from more than the 20 normal amino acids that are usually used to build proteins. What sort of stories could that life tell?

With the power of synthetic chemistry and molecular biology and just under 20 years of work, we created bacteria with six-letter DNA. Let me tell you how we did it.

All you have to remember from your high school biology is that the four natural letters pair together to form two base pairs. G pairs with C and A pairs with T, so to create our new letters, we synthesized hundreds of new candidates, new candidate letters, and examined their abilities to selectively pair with each other. And after about 15 years of work, we found two that paired together really well, at least in a test tube. They have complicated names, but let's just call them X and Y.

The next thing we needed to do was find a way to get X and Y into cells, and eventually we found that a protein that does something similar in algae worked in our bacteria. So the final thing that we needed to do was to show that with X and Y provided, cells could grow and divide and hold on to X and Y in their DNA. Everything we had done up to then took longer than I had hoped — I am actually a really impatient person — but this, the most important step, worked faster than I dreamed, basically immediately.

On a weekend in 2014, a graduate student in my lab grew bacteria with six-letter DNA. Let me take the opportunity to introduce you to them right now. This is an actual picture of them. These are the first semisynthetic organisms.

So bacteria with six-letter DNA, that's really cool, right? Well, maybe some of you are still wondering why. So let me tell you a little bit more about some of our motivations, both conceptual and practical. Conceptually, people have thought about life, what it is, what makes it different from things that are not alive, since people have had thoughts. Many have interpreted life as being perfect, and this was taken as evidence of a creator. Living things are different because a god breathed life into them. Others have sought a more scientific explanation, but I think it's fair to say that they still consider the molecules of life to be special. I mean, evolution has been optimizing them for billions of years, right? Whatever perspective you take, it would seem pretty impossible for chemists to come in and build new parts that function within and alongside the natural molecules of life without somehow really screwing everything up. But just how perfectly created or evolved are we? Just how special are the molecules of life? These questions have been impossible to even ask, because we've had nothing to compare life to. Now for the first time, our work suggests that maybe the molecules of life aren't that special. Maybe life as we know it isn't the only way it could be. Maybe we're not the only solution, maybe not even the best solution, just a solution.

These questions address fundamental issues about life, but maybe they seem a little esoteric. So what about practical motivations? Well, we want to explore what sort of new stories life with an expanded vocabulary could tell, and remember, stories here are the proteins that a cell produces and the functions they have. So what sort of new proteins with new types of functions could our semisynthetic organisms make and maybe even use? Well, we have a couple of things in mind.

The first is to get the cells to make proteins for us, for our use. Proteins are being used today for an increasingly broad range of different applications, from materials that protect soldiers from injury to devices that detect dangerous compounds, but at least to me, the most exciting application is protein drugs. Despite being relatively new, protein drugs have already revolutionized medicine, and, for example, insulin is a protein. You've probably heard of it, and it's manufactured as a drug that has completely changed how we treat diabetes. But the problem is that proteins are really hard to make and the only practical way to get them is to get cells to make them for you. So of course, with natural cells, you can only get them to make proteins with the natural amino acids, and so the properties those proteins can have, the applications they could be developed for, must be limited by the nature of those amino acids that the protein's built from. So here they are, the 20 normal amino acids that are strung together to make a protein, and I think you can see, they're not that different-looking. They don't bring that many different functions. They don't make that many different functions available. Compare that with the small molecules that synthetic chemists make as drugs. Now, they're much simpler than proteins, but they're routinely built from a much broader range of diverse things. Don't worry about the molecular details, but I think you can see how different they are. And in fact, it's their differences that make them great drugs to treat different diseases. So it's really provocative to wonder what sort of new protein drugs you could develop if you could build proteins from more diverse things.

So can we get our semisynthetic organism to make proteins that include new and different amino acids, maybe amino acids selected to confer the protein with some desired property or function? For example, many proteins just aren't stable when you inject them into people. They are rapidly degraded or eliminated, and this stops them from being drugs. What if we could make proteins with new amino acids with things attached to them that protect them from their environment, that protect them from being degraded or eliminated, so that they could be better drugs? Could we make proteins with little fingers attached that specifically grab on to other molecules? Many small molecules failed during development as drugs because they just weren't specific enough to find their target in the complex environment of the human body. So could we take those molecules and make them parts of new amino acids that, when incorporated into a protein, are guided by that protein to their target?

I started a biotech company called Synthorx. Synthorx stands for synthetic organism with an X added at the end because that's what you do with biotech companies.


Synthorx is working closely with my lab, and they're interested in a protein that recognizes a certain receptor on the surface of human cells. But the problem is that it also recognizes another receptor on the surface of those same cells, and that makes it toxic. So could we produce a variant of that protein where the part that interacts with that second bad receptor is shielded, blocked by something like a big umbrella so that the protein only interacts with that first good receptor? Doing that would be really difficult or impossible to do with the normal amino acids, but not with amino acids that are specifically designed for that purpose.

So getting our semisynthetic cells to act as little factories to produce better protein drugs isn't the only potentially really interesting application, because remember, it's the proteins that allow cells to do what they do. So if we have cells that make new proteins with new functions, could we get them to do things that natural cells can't do? For example, could we develop semisynthetic organisms that when injected into a person, seek out cancer cells and only when they find them, secrete a toxic protein that kills them? Could we create bacteria that eat different kinds of oil, maybe to clean up an oil spill? These are just a couple of the types of stories that we're going to see if life with an expanded vocabulary can tell.

So, sounds great, right? Injecting semisynthetic organisms into people, dumping millions and millions of gallons of our bacteria into the ocean or out on your favorite beach? Oh, wait a minute, actually it sounds really scary. This dinosaur is really scary. But here's the catch: our semisynthetic organisms in order to survive, need to be fed the chemical precursors of X and Y. X and Y are completely different than anything that exists in nature. Cells just don't have them or the ability to make them. So when we prepare them, when we grow them up in the controlled environment of the lab, we can feed them lots of the unnatural food. Then, when we deploy them in a person or out on a beach where they no longer have access that special food, they can grow for a little bit, they can survive for a little, maybe just long enough to perform some intended function, but then they start to run out of the food. They start to starve. They starve to death and they just disappear. So not only could we get life to tell new stories, we get to tell life when and where to tell those stories.

At the beginning of this talk I told you that we reported in 2014 the creation of semisynthetic organisms that store more information, X and Y, in their DNA. But all the motivations that we just talked about require cells to use X and Y to make proteins, so we started working on that. Within a couple years, we showed that the cells could take DNA with X and Y and copy it into RNA, the working copy of DNA. And late last year, we showed that they could then use X and Y to make proteins. Here they are, the stars of the show, the first fully-functional semisynthetic organisms.


These cells are green because they're making a protein that glows green. It's a pretty famous protein, actually, from jellyfish that a lot of people use in its natural form because it's easy to see that you made it. But within every one of these proteins, there's a new amino acid that natural life can't build proteins with.

Every living cell, every living cell ever, has made every one of its proteins using a four-letter genetic alphabet. These cells are living and growing and making protein with a six-letter alphabet. These are a new form of life. This is a semisynthetic form of life.

So what about the future? My lab is already working on expanding the genetic alphabet of other cells, including human cells, and we're getting ready to start working on more complex organisms. Think semisynthetic worms.

The last thing I want to say to you, the most important thing that I want to say to you, is that the time of semisynthetic life is here.

Thank you.


Chris Anderson: I mean, Floyd, this is so remarkable. I just wanted to ask you, what are the implications of your work for how we should think about the possibilities for life, like, in the universe, elsewhere? It just seems like so much of life, or so much of our assumptions are based on the fact that of course, it's got to be DNA, but is the possibility space of self-replicating molecules much bigger than DNA, even just DNA with six letters?

Floyd Romesberg: Absolutely, I think that's right, and I think what our work has shown, as I mentioned, is that there's been always this prejudice that sort of we're perfect, we're optimal, God created us this way, evolution perfected us this way. We've made molecules that work right alongside the natural ones, and I think that suggests that any molecules that obey the fundamental laws of chemistry and physics and you can optimize them could do the things that the natural molecules of life do. There's nothing magic there. And I think that it suggests that life could evolve many different ways, maybe similar to us with other types of DNA, maybe things without DNA at all.

CA: I mean, in your mind, how big might that possibility space be? Do we even know? Are most things going to look something like a DNA molecule, or something radically different that can still self-reproduce and potentially create living organisms?

FR: My personal opinion is that if we found new life, we might not even recognize it.

CA: So this obsession with the search for Goldilocks planets in exactly the right place with water and whatever, that's a very parochial assumption, perhaps.

FR: Well, if you want to find someone you can talk to, then maybe not, but I think that if you're just looking for any form of life, I think that's right, I think that you're looking for life under the light post.

CA: Thank you for boggling all our minds. Thank so much, Floyd.