In their third experiment, Venter’s crew changed one bacterial species into another one. They did this by taking the genome from one species and transferring it into members of the second species, which then turned themselves into members of the first. In this case the researchers used a natural (as opposed to a synthetic) genome, however, and the species in question were two different types of Mycoplasma: M. mycoides and M. capricolum. “These species are more convenient experimental organisms than M. genitalium because of their faster growth rate,” the researchers wrote in their report on the project, which was published in Science in 2009. (M. genitalium has an extremely slow growth rate.)
Although the procedure was technically complicated, it was simple enough conceptually, since what the researchers did was to isolate an M. mycoides genome and transplant it into wild-type M. capricolum recipient cells. For a while there were two different genomes residing in the same cell. Eventually the new DNA was recognized and taken up by the recipient cell, which thereupon transformed itself into an M. mycoides bacterium.
“Changing the software completely eliminated the old organism and created a new one,” Venter said of the experiment. This might at first glance sound like a magical changeover, but the invading genome was merely acting like a virus, taking over and transforming the cell into which it had been placed. Just like Venter’s genome, a virus is software that completely eliminates the old organism and creates new ones.
Still, Venter’s capping and culminating experiment was yet to come. This was to design, digitize, and then chemically assemble a 1.08-million base pair M. mycoides genome and boot it up inside a cell. They called this synthetic genome M. mycoides JCVI-syn1.0. Then they did with it exactly what they had done with the natural M. mycoides genome of the earlier experiment: transplant it into an M. capricolum recipient cell. The results were the same: the new (synthetic) genome took over the old M. capricolum cell and turned it into an M. mycoides cell.
As the researchers told the story in Science: “There was a complete replacement of the M. capricolum genome by our synthetic genome during the transplant process. . . . The cells with only the synthetic genome are self-replicating and capable of logarithmic growth.”
These developments created a minor sensation in the scientific world and a major sensation in the general media (“Scientists Create Artificial Life”). There were news reports saying that President Barack Obama had expressed unspecified “genuine concerns” about this work.
With one notable exception, however, Venter and his colleagues were quite restrained in their claims. In their report on the project, the researchers drew two general conclusions from what they had done. First, “the demonstration that our synthetic genome gives rise to transplants with the characteristics of M. mycoides cells implies that the DNA sequence upon which it is based is accurate enough to specify a living cell with the appropriate properties.” In other words, there are no mystic features, holdovers, or leftovers from vitalism pertaining to DNA molecules. Whether they were “natural” or “synthetic” genomes, they still controlled a cell.
Second, “this work provides a proof of principle for producing cells based on genome sequences designed in the computer. DNA sequencing of a cellular genome allows storage of the genetic instructions of life as a digital file.” The reduction of genetic instructions to a digital file delivered a knockout second blow to vitalism.
But then the scientists advanced a third claim: “We refer to such a cell controlled by a genome assembled from chemically synthesized pieces of DNA as a ‘synthetic cell,’ even though the cytoplasm of the recipient cell is not synthetic.” They made it sound as if they had created an artificial life form even though a nonsynthetic, natural cell had actually given rise to the new organism.
The genome constitutes only about 1 percent of the dry weight of a cell, which means that only a tiny proportion of the cell is actually synthetic. The rest of the organism was as natural as any other ordinary cell. Indeed, Venter’s synthetic genome depended on the rest of the recipient cell’s natural and native apparatus for its expression: it depended on the cell’s molecular machinery of transcription, translation, and replication, its ribosomes, metabolic pathways, its energy supplies, and so on. (Although Venter was fond of saying that “the DNA software builds its own hardware,” it would be more accurate to say that the recipient cell builds whatever hardware the DNA software codes for—and only if the existing hardware is pretty close to the target hardware already.)
Building a living cell that is genuinely synthetic is one of the goals of synthetic biology. By separating what’s essential to living systems from what’s not, such a cell would advance our understanding of what constitutes the necessary and sufficient conditions for being alive. In addition, a synthetic cell, provided that it is also a minimal cell, is considered by some to be a beguiling platform for genomic engineering since its lack of extraneous or inhibitory components might improve its efficiency at turning out desired end products such as biofuels, medicines, vaccines, or green chemicals, although others say that a larger genome is better.
Further, discovering or creating a minimal organism would establish the limits of what’s possible in the miniaturization of living systems. Biological minimalism can exist on two different levels. First, there is the minimal genome: the smallest genome that is sufficient to create, maintain, and replicate itself. Possibly such a genome could be as small as two 3-mers (three-part molecules) that come together to form a 6-mer (a 6-part molecule). Increasingly interesting genomes (of 187, 2587 and 113,000 base pairs) will be introduced soon (explained below).
On the second level there is the minimal cell, composed of the fewest components that can jointly carry out all the normal processes of life, including metabolism, reproduction, and evolution. As the genome grows larger, it gets steadily harder to separate the full length of the two strands in order to make new copies. The solution is to separate out only a little at a time (say a dozen base pairs out of millions) and synthesize a new strand of DNA or RNA a few base pairs at a time—with a long copy emerging from the intact double helix. This argues for a separation of information storage and factory functionalities. These functions are typically encoded in RNA and in proteins that fold up into complex machines instead of being long rods. The ability to fold gives access to vast capabilities, and in principle this could be done simply with RNA genomes and with RNA as folded machines. But RNA has only four closely related functional groups (A, C, G, U), so the coding of another class of polymers (proteins) with vast diversity (20+ amino acids) was, perhaps, inevitable.
Back to what it’s like for a cell to be a cell. Animations help us visualize how polymers are made from monomers. Often these are depicted as happening in an orderly fashion, similarly to how workers on an assembly line might pass car parts down the line “just in time” for the next sequential production step. The process, however, is hardly so orderly. In reality, the four nucleotides or twenty amino acids are randomly tried out before a single correct one is accepted, and each of these with considerably more jostling and false moves than is typically shown in animations.
Seemingly, a minimal genome would automatically produce a minimal organism, but this is by no means obvious. Some protozoa, for example, have genomes that are over one hundred times larger than the human genome, which means there is a big mismatch between the size of the genome and the size of the corresponding organism. This is largely due to the fact that large stretches of the protozoa’s genome may consist of noncoding regions (sometimes called junk DNA), meaning dispensable under some circumstances. A minimal genome, however, would exclude such sequences by design and intention. Still, a cell produced by such a genome might nevertheless contain extraneous, redundant, or other inessential components. Whether a minimal genome will in fact produce a minimal cell is something that can be decided only after the fact, by experiment, not in advance, by theory. (But if “theory” means, as it often does, going from vast numbers of experiments to
a new best guess, then the minimal genome will likely come from theory.)
Attempts to build a synthetic cell have not been entirely successful. In 1969, for example, three biologists at the State University of New York–Buffalo, K. W. Jeon, I. J. Lorch, and J. F. Danielli, decided to create a synthetic living organism. “After participating in a symposium on the experimental synthesis of living cells,” they wrote in their report on the project in Science, “we decided that we had the means to carry out the reassembly of Amoeba proteus from its major components: namely, nucleus, cytoplasm, and cell membrane.”
Amoeba proteus is a comparatively large (0.4 mm) aquatic organism that is easy to work with using tools such as micropipettes and other micromanipulators. And so the experimenters took the nucleus from one amoeba, the cytoplasm from a second, and put them together inside the evacuated cell membrane of a third. Eighty percent of the time, the new composite organism lived. “The techniques of cell reassembly appear to be sufficiently adequate so that any desired combination of cytoplasm, nucleus, and membrane can be assembled into living cells,” the researchers concluded.
Unfortunately this cobbled-together organism was not really a synthetic cell, for all the parts used were natural; only their locations had changed. It was more like reshuffling the deck than providing the players with a new set of cards.
A truly synthetic cell is one that we create ourselves, from the ground up. This could be a new form of living matter fabricated out of pure ingredients. Such a cell might tell us something about the original cells that arose at the dawn of life on earth. Arguably, life originated when a group of molecules and molecular structures first organized themselves into living systems, but precisely what those molecules were and how they arranged themselves so that life emerged from the mix is an open, possibly unanswerable question. Nevertheless, successfully creating a synthetic cell would represent a key advance in the understanding of living processes. For life, like a machine, cannot be understood simply by studying it and its parts; life, to be understood, must also be put together from its parts.
How then do we create a truly minimal living cell that is also genuinely synthetic? In 2006 a colleague and I advanced a proposal for the creation of a minimal cell that in addition would be a substantially synthetic living organism. My colleague is Anthony Forster, then of Vanderbilt University Medical Center, now of the University of Uppsala, and our object was to build, from the molecules up, a chemical system capable of replication and evolution. As opposed to Venter’s reductionist biology, this would be an example of constructive biology, the putting together of a living entity from its constituent parts.
The work would proceed by starting with the smallest molecular components and arranging them into subsystems, then having those subsystems self-assemble into larger units, and so on.
We designed our genome by looking for genes that have homologues, closely related sequences, in the genomes of several groups of organisms. The rationale is that genes that appear in many groups of organisms are somehow “required” for those species. This method, along with others from genetics and biochemistry, suggested a genome that consists of only 151 genes and is only 113,000 base pairs long. Our plan is to construct the genome, place it inside a lipid-bilayer membrane-sphere filled with the macromolecular enzymes encoded by the 151 genes and a minimal inventory of small molecules needed for life. The entire system could finally be bootstrapped into existence by the addition of synthetic ribosomes, translation factors, and other structures inspired by similar components existing in natural cells such as E. coli.
Eventually this approach will produce a synthetic, self-replicating, and self-sustaining minimal cell. To prevent the cell from replicating outside the laboratory, we are building into it a deliberate dependence on nutrients not found outside the lab environment.
Unlike the synthetic bacterial genome, where many genes are of unknown structure or function, the end result would be a functionally and structurally well-understood self-replicating biosystem, synthetic but nevertheless alive and well. If and when it happens, it will be a major milestone in the history of biology: civilizing, taming, and domesticating the basic processes of life. This synthetic minigenome presents a clear path to the mirror world introduced in Chapter 1. The synthetic minimal cell would enable the production of materials too large or otherwise incompatible with the more elaborate functioning systems of a complex cell. It also represents our best shot at a general nanotech assembler, the dream of Eric Drexler and many nanotechnology enthusiasts since he first described it in his 1986 book Engines of Creation. We could then harness these synthetic minimal cells and put them to use in drug, vaccine, chemical, and biofuel development.
CHAPTER 3
-500 MYR, CAMBRIAN
The Mirror World and the Explosion of Diversity. How Fast Can Evolution Go and How Diverse Can It Be?
Life, the Genetic Code, the Mirror World, and the Generation of Diversity
Mirror life, including mirror humans, may sound like something out of science fiction, an outlandish concept unrealizable in practice even if conceivable in theory.
But mirror life is a real possibility, not just a flight of fancy. To convince you of this I’d like to show you how it can be created. But first we need a deeper understanding of life itself, in all of its complexity.
What, then, is life?
Erwin Schrödinger’s short 1944 book What Is Life? inspired physicists such as Maurice Wilkins and others to establish the field of molecular biology. Schrödinger championed the idea of life as based on an “aperiodic solid,” a suggestion that anticipated DNA as the sequenced biopolymer in which the genetic information is encoded.
Many people think that life is an all-or-none, black-or-white, on-or-off, matter-antimatter phenomenon, with no in between. However, let’s consider the possibility that life is a continuous, scalable, and measurable property. Similarly, many thinkers are tempted to argue that life is “the pinnacle of complexity.” But let’s consider replacing complexity with the notion of replicated complexity (which can be shortened to “replexity”), or mutual information. Two images composed of scads of random ink dots seem equally complex, equally likely or unlikely. Similarly, the atomic arrangement of two stones may seem equally complex. But if we see an image of a complex set of dots with a mirror duplication (like a Rorschach inkblot) or a “living stone” (Lithopsjulii), we experience the feeling that so much information is unlikely to be duplicated or transformed in a predictable manner from page to page, leaf to leaf, or cell to cell within those leaves. Two complex random patterns seem unremarkable consequences of an inorganic, lifeless world, but two complex patterns that look precisely alike are a hallmark of life.
From an information theory standpoint, we need fewer bits by far to convey an image of a grain of salt and an ice crystal than a movie of a liquid mixture of the two. The chaotically changing solution is more complex than either of the two crystals. Complexity increases with chemical randomness or entropy. Chemical systems tend to spontaneously go toward random, complex mixtures. In information science, entropy refers to the number of bits needed to transmit a complex message. Both from a chemical and an information entropy viewpoint, a living system is less complex than the solution but more complex than the crystals. It would have higher replexity than both because the living system requires more bits to describe its repeating motifs (e.g., its nearly identical DNA, proteins, and cells) than the strict and simple repeating motifs of the crystals or the disordered and weakly repeating structure of the solution. So too the chemical imperative for randomness is satisfied in the solution.
The notion of replicated complexity directly addresses seven counter-examples that have challenged previous definitions of life.
(1) Life as a growing and replicating system. Fire, then, may seem to be a living thing, as it can grow and replicate itself. It can even “evolve” to have new properties; for example, while devouring a succession of materials with different flash points (gasoline
at -40 C, ethanol at 13 C, diesel at 62 C, vegetable oil at 327 C, Mg metal at 634 C). Indeed, a flame can replicate in a wide variety of environments while many endangered species can’t survive even with tender loving care in a zoo. However, a flame has little long-range order and so would have a low replexity value.
(2) Replexity also removes difficulties with apparently living things such as mules (the sterile offspring of a male donkey and a female horse) that normally cannot reproduce and thus lack a defining characteristic of life: the ability to self-replicate. But mules have reasonably high replexity for three reasons. First, they have a non-zero probability of whole-organism replication. Second, their cells and subcellular components replicate. And third, they are closely related to organisms that do replicate (i.e., their parents).
(3) Life as a steadily increasing reservoir of replexity. A lone herbivore could destroy a complex plant ecosystem and then die, resulting in a net decrease in replexity. This is an average property, not a guarantee. This is analogous to the laws of thermodynamics, which describe the average behavior of vast numbers of molecules, not the instantaneous motion of a few molecules.
(4) Should life display motion? “What about a frozen animal?” you might ask. “That doesn’t seem very alive.” Many complex organisms survive a freeze-thaw protocol (e.g., human IVF embryos and Chironomous larvae). In principle, a frozen organism could be assembled one molecule at a time with small tweezers such as those of an atomic force microscope. (This has not been done yet, but I addressed the idea of assembly from parts among the five grand challenges to vitalism discussed in Chapter 1 and in the context of booting up minimal genomes in Chapter 2.) If we wanted to email a description to specify that organism reliably, we could determine the minimum number of bits needed to transmit the specifications of the organism’s structure so that it would be in a living state on construction and warming. The replexity of frozen animals reflects both their origin from a replicating entity as well as probability of replicating again on thawing.
Regenesis Page 6