NOVA scienceNOWNOVA scienceNOWNOVA scienceNOWComing up
arrow Artificial Life
Really Simple Syndication (RSS)

Francis Collins Interview


Francis Collins isn't one to cast aspersions on highly ambitious scientific goals. But he thinks the chances for concocting life in a lab—from scratch, without cheating—are slim.


Even humble-looking E. coli bacteria are complex organisms, housing a host of different proteins and organelles, which are invisible in this scanning electron micrograph.

Enlarge this image

Red wagon

If a bacterium is a fancy Cadillac, the first synthetic life-form might be a little red wagon with a loose wheel, says Collins.

Robert and Jeremy look in synthesizer

Jeremy Minshull, president of the biotech firm DNA 2.0, gives Robert Krulwich a peek inside a machine that synthesizes DNA from tiny vials of "A's," "Ts," "Cs," and "Gs"—nucleotides derived from salmon sperm.

Enlarge this image

Protein folding

Predicting how a stretch of DNA will translate into the folds and function of a protein is no easy trick. In this 3-D computer model of the protein Acetylcholinesterase, the twisted blue ribbons are strings of amino acids that were encoded by DNA.

Enlarge this image


Does our understanding of life's chemistry—such as the now well-known structure of DNA—profoundly change how we view ourselves?

Enlarge this image

The Stuff of Life

Robert Krulwich: When we talk about going from something that is not alive to something that is alive, we're talking mainly about chemicals. So first of all, do you ever think of yourself as a "bag of chemicals?"

Francis Collins: "Chemicals" has sort of a bad name right now. I'm aware there are certain products that are being advertised—food products—with "no chemicals whatsoever." Well, that would be pretty hard to arrange, since everything around us is made up of atoms and molecules—chemicals—including ourselves. I'd be the first to say, I don't think that fully explains what being human is all about. But I would also be quick to say, yeah, we do have a certain list of constituents of a chemical sort. It's a very long list, and they're arranged in a very special way. So simply melting us down and tallying up the number of compounds of this and that sort would not be very revealing about how the whole thing works.

Krulwich: But these researchers plan to go shopping in some store and take bags off a shelf with, like, inert dust. And the idea is to turn it into life?

Collins: This is sort of the ultimate in reductionism, isn't it? The idea that you could put together something that you could call "living" from basic building blocks—chemicals—assembled in a certain way. It quickly gets you into the dilemma of what's the definition of "living" in this kind of experiment? What would you call success?

Krulwich: What do you think a living thing would require to be alive?

Collins: The word "living" has so many connotations that I'm almost reluctant to try to define it scientifically because it sounds as if I'm then downgrading all the other significances of that word. But if you wanted a short list of the properties of something that we would call living:

It would need to be self-replicating—that is, it could copy itself stably into many other copies. It would need to be stable over time. And it would need to be able to survive without the input of lots of other complicated proteins or other large molecules that aren't generally available, floating around in the atmosphere. Sure, it could have some sort of media that it lived in, some sugars and some salts, and things like that. But let's not require it to need cytochrome P450.

And maybe we ought to require it also to have the ability to evolve over time under the force of some selection process, although that may be an unnecessary added frill, and not everyone would agree that that is needed to define life.

Krulwich: Will there be a problem here if someone says, "Wow! I did it." Will there be an argument over whether "it" has been done?

Collins: I think there will be profound arguments about who has actually crossed this line. If you imagine a single molecule of RNA, for instance, that is somehow able to copy itself, surrounded by a membrane of lipids, is that good enough? It isn't going to be a bright line. This will not be one of those moments where, with a blinding flash, the universe has changed.

Krulwich: That's interesting. Because it feels to me that something huge will have taken place if you can go from nonlife to life.

Collins: I think it's like climbing a hill where you're trying to decide if you're at the top. And by the time you realize you're at the top, you really went past it. I suspect that will be true here, that there will be incremental steps, and people will argue furiously about whether it counts or not. And finally we'll get to the point where everybody agrees, yeah, we probably now have something that meets that definition of a living thing.

Krulwich: It sounds like you find this a plausible quest.

Collins: I think probably scientists will figure out a way to generate some simple self-replicating system. Does that really deserve the definition of life? Or is this kind of a caprice, kind of a nice, interesting academic exercise, but nothing like the way in which we conceive of the abundance and diversity and complexity and wonder of real life on this planet. It might be, really, a very poor representation of anything that we currently look at on the planet and call living.

Frankenstein or a Feebler Life-Form?

Krulwich: Wouldn't the ability to go from inert to living in a small way predict the ability to go from inert to living in a big way?

Collins: Obviously, people will be worried about what comes next. If you can make a small step in this direction and have something that at least, by some definition, is called life, where does this end? Remember, evolution has been at this for a very long time. Something like 3.85 billion years have transpired since the first self-replicating organism appeared on Earth. And that has resulted in very robust self-replicating systems that are remarkable in their ability to adapt to environments and to sustain all kinds of unexpected insults.

I find it very hard to believe that, starting from scratch, we can somehow come up with a better system—one that's going to have much success. So I'm not too worried that somehow, some completely different life-form is going to get out of control. I think the [synthesized] life-forms will be actually very imperfect, very unsatisfying, very, very fragile.

Krulwich: So if it happens at all, it'll be a "ehh, ehh, ehh—whomp!?" It will be such a fragile thing?

Collins: If you want to use a machine analogy—and admittedly, there are lots of problems with this—think of a mammal as an incredibly complicated machine, you know, the latest version of the 767-plus. Even a bacterium on that scale is an incredibly complicated organism, perhaps like a fancy Cadillac, anyway.

Krulwich: Or a Saturn rocket?

Collins: Maybe. Now maybe somebody tinkering in the lab will succeed in building a Radio Flyer wagon, where the wheel will fall off after the first 10 feet, but you will be able to say, "It went." Does that really threaten the larger, more complicated machines? Are they going to be somehow supervened by these clever new developments? I don't think so.

Krulwich: That's interesting.

Collins: If there's really a risk, both philosophically and perhaps in safety terms, I don't think it's in developing life from scratch. It will be that, with extensive knowledge of how life actually works, people will start to redesign organisms, taking what they've learned from what has, in evolution's laboratory, worked quite well, and modifying it.

Rewriting the Book of Life

Krulwich: Well, let's pick up on this redesigning life. Is it easy now to, say, go to the "gene store" and build new life-forms?

Collins: Well, one of the big developments of the last couple of years, and one that I don't think is widely appreciated, is the ability to synthesize any DNA molecule that you want at high accuracy and low cost.

Krulwich: So let's suppose a gene is spelled "A-A-A-A-C-C-C-C-C-T-T-T," okay? Can you go and buy some A at a store, and buy some C and buy T?

Collins: It's getting very much like that. You can set up your synthesizer. You buy your A's, C's, G's, and T's. And you load up your A's in this bottle and your C's in that bottle, your G's are over here, your T's are over there. You sit down in front of your computer, and you program in what you want to make. And presto!

Krulwich: Now let me ask you about the chemical "letters." Once they go into this machine, will they just self-assemble?

Collins: Ha-ha. No, no, no. The way this works is, you're depending upon enzymes, which you are generally borrowing from life itself.

Krulwich: Ah, so you're cheating!

Collins: You're cheating! Your machine is being empowered by what we've learned from biology. Not just principles, but actually proteins that were generated, usually by some bacterium.

Krulwich: So we'll start with inert chemicals; we'll start with some dust. We'll put in a few biological tools stolen from a living thing. And then if we just put in the right order, does God have to breathe on the machine? This is the thing I'm trying to figure out—what happens next?

Collins: Well, remember, what you're making in this machine is not life; you're making DNA. You're making a long strand of DNA with a particular sequence.

Krulwich: But if I put enough DNA together, maybe I'd get a living thing.

Collins: No. You'd get DNA. Remember, DNA is essentially not capable of doing anything; it's just the instruction book. Having printed out the instruction book doesn't make the house come into being, even though it may specify all of the particulars of what that house is supposed to look like.

We now know how to synthesize DNA pretty efficiently. So, not only could we make DNA that's a smallpox virus, pretty soon, we could make the DNA of an entire bacterium. It's not that many letters of the code, perhaps only a couple of million. But would that be life? No, it would not.

The First Synthetic Life?

Krulwich: What about that virus that some guy made? Am I right to say that somebody put in all the right spellings and made a virus-like thing? [In 2002, polio researcher Eckard Wimmer generated a so-called "synthetic" poliovirus. A year later, a team led by J. Craig Venter and Hamilton Smith announced similar work with a virus called PhiX.]

Collins: What they did was synthesize the instruction book. But an instruction book just sits there. It doesn't actually build anything. For that instruction book to cause trouble, that viral DNA had to be put inside a cell, where the cell provided all of the "workers."

Krulwich: So what they made was different from the virus itself?

Collins: A virus is not just DNA; a virus is also packaged up, covered over with a series of proteins in a nice, elegant, well-compacted form. What was made in the machine was just the DNA. And if you put that naked DNA—I mean, really naked DNA—into a cell, you basically bypass the requirements for how the virus would normally get in. Then that DNA, once inside the cell, harnesses the machinery that the cell provides, which is a lot!

Krulwich: Oh, so it did a virus-like thing.

Collins: It did a virus-like thing and made copies of itself. The copies had not just DNA but also proteins, because those were encoded in its instruction book. So these new viruses were capable of infecting the next round of cells.

Krulwich: So now why isn't that an "Oh my gosh!" kind of event?

Collins: I guess we've gotten so familiar with the idea that viruses are just naked DNA with some protein packaging that this was completely a no-brainer. I think a lot of people thought, "Well, what's the big deal? Why is that so exciting? So you synthesized a virus from scratch? Of course!"

Krulwich: Not me, I'm thinking, "Wow!" You went from dust with a little help and some cheating to something that was arguably motile. It did things.

Collins: I think it was way short of that. It wasn't motile, and it didn't do anything. You had to inject it into this wonderful milieu that provided, really, all of the exciting nuts and bolts to make this happen, namely the cell. Without that, it was a pretty boring stretch of DNA with no potential of doing anything without your help. So I would say that was not life; that was a chemical.

Custom-Made Creatures

Krulwich: So there's nothing random about what was done with this virus; you got a recipe. You've decoded a living thing and now you're trying to bake it yourself.

Collins: Yes.

Krulwich: Would you say "Oh, wow!" if some of these people who have more of a business bent create something like a bacterium that is new and very, very valuable to the world? Is that possible?

Collins: There is clearly the potential of being able to build, at will, some kind of organism that has properties that might be beneficial in terms of, say, cleaning up a toxic spill. Again, I'm having trouble seeing that as a quantum leap over what one can do now by stitching together bits of DNA that came from different organisms.

Again, you could argue, evolution has provided us with an incredibly rich array of options here—genes that we know work very well, that could be adapted for an appropriate purpose. And that might be a much more successful way to build new organisms. Take the entire panoply of successful genes from all organisms that we have access to and reshuffle them—which is what recombinant DNA is—instead of imagining that we're so smart that we can come up with something even better.

Krulwich: You don't think there's anything to the argument that we can just synthesize something better?

Collins: We are so woefully ignorant about how biology really works. We still don't understand how a particular DNA sequence—when we just stare at it—codes for a protein that has a particular function. We can't even figure out how that protein would fold—into what kind of three-dimensional shape. And I would defy anybody who is going to tell me that they could, from first principles, predict not only the shape of the protein but also what it does. The idea that we could somehow just step back from what evolution has already figured out and do a better job because we are able to predict those things from theoretical principles, is just so far away from reality as to be laughable.

Getting Back to Life's Origins

Krulwich: So switching away from the folks who want to just write new DNA, there are other guys who are thinking, "Let's try this: let's just create environments that are like, or maybe like, the environment that produced life in the beginning. Let's see if we can find a way to start life here from scratch."

Collins: We've been trying to explore this sort of nebulous question for decades, of course. This is what Stanley Miller did back in the '50s—trying to pass an electric discharge through a bunch of organic compounds and show that he was able to generate amino acids and maybe even some nucleotides. It's still hard to know what you would call success. Again, you better have a very clear definition in mind of what you are willing to say in any way resembles life.

Krulwich: I think they say, "If we can get a little sack full of stuff, that stuff can take the energy of the world and use it somehow, and then reproduce itself, then we have done it! We've actually done it!" And you'd say?

Collins: I'd say, it'll be very interesting to see how that turns out. But I think the probabilities from first principles are probably pretty low that that's going to work. Not zero, but pretty low.

Krulwich: I'm just coming back from meeting some of these guys. They look like they've just swallowed the cat!

Collins: Or drank the Kool-Aid. Well, I'm glad they're excited about it, because you'll never accomplish anything in science if you're not.

I guess what we're really talking about here is what was the origin of life on Earth to begin with? And is that something which, in a scientific experiment, you could get some glimmer of by essentially creating a circumstance where something vaguely resembling life is, under your own observation, coming into being? Isn't that really the point? To try and answer that age-old question, how did it all start?

Darwin himself had a lot of trouble with it and we still do today, because nobody has really come up with a terribly plausible system that would convince the skeptics that it was actually the way it happened.

Now, if you were able in a laboratory situation to create something that was capable of self-replication, that wouldn't prove that's how it happened; it probably would be way off from whatever happened. But it would be a demonstration of the possibility that, from simple compounds in the right circumstances, you might end up with something that starts this process. Some people will find that philosophically very upsetting; I think most people won't, actually.

What About God?

Krulwich: If human beings were able to turn nonlife into life, what would that do to our notions about ourselves?

Collins: I think our notions about ourselves probably can be questioned on many grounds other than this one. With what we've learned about evolution and our relatedness to all living things, it's very clear that—in terms of the mechanical aspects of who we are—that we are derived from many other ancestors down through the last three billion years. Defining how it all started—does that change our view of ourselves profoundly? I'm not sure it does.

Krulwich: If a human being were to create life from nonlife, that would cross—in my view, anyway—a line that I didn't know that we could cross. A mystery would be demystified. It doesn't strike you that way?

Collins: We have been demystifying life for decades. Ever since we figured out "spontaneous generation" is not right, that, in fact, life comes into being from things like spores and eggs, and so forth, we have been learning that life follows certain principles and rules. I think you'd be hard-pressed to find a microbiologist who would say, "There is something unscientific, some vital force about E. coli that allows it to keep going."

Krulwich: So, really, you feel that way?

Collins: Yeah.

Krulwich: A very little, itty-bitty thing gets along because it's just chemicals in the right order.

Collins: Right. Now does that apply to human beings? No, I happen to think human beings fall in a different category. I think we, of all the organisms on the planet, have a spiritual nature which can't be explained by those common elements and "instruction books" and references to machine analogies. We have these remarkable features such as the understanding of what's right and wrong, which I don't think is something that will come out in the study of biology. Nor is it something that I think evolution can explain.

So I am a person of faith, as you might be able to tell from that last remark. But my faith is not in any way resting upon whether or not it's possible in a test tube, at some point, to generate something that looks like it replicated and had the properties of life.

Krulwich: So God wouldn't be a little bit diminished if a human could make a living thing from a nonliving thing? Or humankind wouldn't be inflated if we could make a living thing?

Collins: Well, God would certainly not be diminished. God, if it's the God that I worship, created the universe and all the laws that regulate it, and gave us this incredible gift of an intellect. And I, like Galileo, don't think that he gave us those abilities in order for us to forego their use. And so I think God kind of thinks that science is pretty cool!

So I'm not worried about God. I am worried about humans, because we have a long tradition of assuming greater importance for ourselves than we deserve. And so this does slip into the zone of hubris: "I'm no longer just an ordinary person; I am creating life. That makes me a little closer to God, and maybe a little less in need of Him, after all." If somebody were to wrap themselves in that kind of philosophical mantle, then I think we've actually not upgraded man, we've downgraded him.

“Probably scientists will figure out a way to generate some simple self-replicating system. Does that really deserve the definition of life?”

“I’m not too worried that somehow, some completely different life-form is going to get out of control.”

“If there’s really a risk ... it will be that, with extensive knowledge of how life actually works, people will start to redesign organisms.”

“The idea that we could somehow just step back from what evolution has already figured out and do a better job ... is just so far away from reality as to be laughable.”

“I’m not worried about God. I am worried about humans, because we have a long tradition of assuming greater importance for ourselves than we deserve.”

Interview conducted on August 10, 2005 at the National Human Genome Research Institute by Robert Krulwich, executive editor of NOVA scienceNOW, and edited by Susan K. Lewis, editor of NOVA online

arrow Artificial Life