Life’s most poignant hallmark is inescapable, inexorable change. Each of us is born, grows old, and dies. Each species arises from prior species, fills its own niche for a time, and eventually becomes extinct. In such a sweeping, undirected evolutionary drama, the human species might be seen as but one small, insignificant player. Little wonder that Darwin’s theory of evolution by natural selection has met with so much hostility.

The process of evolution by natural selection rests on two incontrovertible facts. First, every population is genetically diverse, possessing a range of traits. Second, many more individuals are born than can hope to survive. Consequently, over time those individuals with more advantageous traits are more likely to survive and pass on their genetic characteristics to the next generation. This selection process drives evolution.

Even today, almost a century and a half after the publication of The Origin of Species, a vocal minority of Americans views the theory of evolution as a dangerous and subversive doctrine that substitutes materialism for faith. Efforts to expunge Darwin from textbooks and to augment curricula with thinly veiled religious beliefs in the guise of “scientific creationism” or “intelligent design” continue unabated in many states.

But natural selection is not a sinister development in life’s emer-

gence. On the contrary, evolution is the natural and necessary sorting-out process that led to the origin of life.

Imagine yourself back to the primitive Earth more than 4 billion years ago, to a time before life had emerged. Oceans and shorelines must have held a bewildering, chaotic diversity of organic molecules from many sources. Somehow that confused chemical mess had to be sorted out. Two connected processes—molecular selection and molecular evolution—winnowed and modified the prebiotic mélange.

Molecular selection, by which a few molecules earned starring roles in life’s origin, proceeded on many fronts. Some molecules were inherently unstable or unusually reactive and so they quickly disappeared from the scene. Other molecules proved to be soluble in the oceans and so were removed from contention. Still other molecular species sequestered themselves by bonding strongly to surfaces of chemically unhelpful minerals or clumped together into gooey masses of little use for the emergence of life.

Earth’s many cycles amplified these emergent selection processes. Tidal pool cycles of wetting and evaporation concentrated molecules, while the Sun’s ultraviolet radiation fragmented the least stable molecular species. Pulses of hydrothermal seawater delivered new supplies of chemicals to deep-ocean vents, where differential adsorption and detachment of molecules on reactive mineral surfaces concentrated a select subset of molecular species. Day and night, hot and cold, sun and rain, high tide and low—these and other periodic phenomena refined the chemical mix.

In every geochemical environment, each kind of organic molecule had its reliable sources and its inevitable sinks. For a time, perhaps for hundreds of millions of years, a kind of molecular equilibrium persisted, as the new supply of each molecular species was balanced by its loss. Such an equilibrium assemblage features nonstop reactions among molecules, to be sure, but the system does not necessarily evolve.

At some point, by processes as yet poorly understood, a self-replicating cycle of molecules emerged, thereby changing the character of molecular selection. Even a relatively unstable collection of molecules could persist in significant concentrations if it made copies of

itself. Molecules within that first self-replicating collection would have thrived at the expenses of other molecular species.

Theorists imagine a more interesting possibility—an ancient environment in which two or more self-replicating cycles of molecules competed for atoms and energy. Such competition inevitably arose as new molecular species offered alternative chemical pathways, or perhaps as changes in environment triggered slight variations in a cycle. Dueling molecular networks would have vied for resources, mimicking life’s unceasing struggle for survival. In such a competitive environment, increasingly efficient cycles emerged and flourished at the expense of less efficient variants, slowly shifting the molecular balance. Molecular evolution had begun on Earth.

The dynamic, competitive tussle of molecular evolution differs fundamentally from the more passive process of molecular selection. Competition among self-replicating cycles drives evolutionary change, fostering efficiency and introducing novelty. That’s why many origin experts draw the arbitrary line that separates living from nonliving systems at the emergence of a self-replicating chemical cycle that began to evolve by this powerful process of natural selection.

Replication of a cycle, in and of itself, is not enough to claim life. The cycle must also possess a sufficient degree of variability so that when it competes for resources, more efficient variants win out. Over time the system changes, becoming more adept at gathering atoms and energy. Under these conditions, the emergence of increasing molecular complexity is inevitable, as new chemical pathways overlay the old. So it is that life has continued to evolve over the past 4 billion years of Earth history.

A theory, no matter how plausible, requires experimental testing, but how can one test molecular evolution in a laboratory? That challenge seems almost beyond imagining, yet that is exactly what a few teams of biochemists have done.

Laboratory studies of molecular evolution began in the 1960s at the University of Illinois, where Professor Sol Spiegelman and his colleagues investigated a tiny virus called Qβ (pronounced “que-beta”) that attacks the common bacterium E. coli. Qβ is nothing more than a protein shell surrounding a small loop of RNA. Its entire existence is

devoted to infecting E. coli cells and inducing those cells to make more copies of the Qβ virus. To do that, one of Qβ’s proteins, Qβ replicase, has to make copies of Qβ RNA.

This close relationship between Qβ replicase and Qβ RNA suggests a promising experiment. If you put some Qβ replicase and a Qβ RNA strand into a test tube along with lots of small RNA building blocks, you’ll wind up with countless new Qβ RNA strands. One caveat: Qβ replicase makes lots of mistakes, so each new RNA strand is likely to vary slightly from the original.

Here’s what Spiegelman and co-workers did to make their molecules evolve. They allowed the chemical mix to make copies for just 20 minutes. During that short period, some RNA sequences bound strongly to Qβ replicase and duplicated rapidly, while others interacted poorly with Qβ replicase and were inefficiently copied. After 20 minutes, when the solution had become enriched in RNA strands that easily replicate, they transferred a small amount of that liquid into a new beaker with fresh Qβ replicase and RNA building blocks. Then they ran the experiment for another 20 minutes and transferred a small amount of the second batch of RNA copies into a third beaker with fresh Qβ replicase.

Seventy-four times they repeated the process, gradually reducing the time allowed for the reaction to proceed. Each step preferentially selected those RNA strands that were copied most efficiently. By the end of the experiment, the length of the most efficient RNA strands had been shortened to a sixth of the original size, and these evolved RNA sequences were being copied 15 times faster than before. Similar experiments generated strands of RNA that were unusually resistant to high temperature or to the effects of damaging chemicals. Under such severe selection pressure (time, temperature, or some other stress), the Qβ RNA evolved.

“Spiegelman monsters,” as these systems came to be known, are not alive. Only by the most overt manipulation (cycling through a succession of beakers, for example) does the Qβ RNA evolve. But the repetitive selection process developed by Spiegelman’s group proved that molecular systems under competitive pressure can be induced to evolve—a result that suggested to many researchers a brave new world

of artificially evolved molecules designed to accomplish specific chemical tasks.

Harvard Medical School biologist Jack Szostak has set his sights even higher. Indeed, Szostak’s overriding ambition is nothing less than to design an evolving life-form in his lab. “Our ultimate goal is to create life,” he admits.

Jack impresses you as someone who is supremely happy in his profession. Unassuming and quiet by nature, he often wears a half-smile, as if he were amused by a private joke. Despite his 50-odd years, his slender build and the bright eyes behind round-framed glasses make him look like a scientific Harry Potter—and he is in fact a bit of a wizard.

For Jack, RNA is the key. He began his scientific career studying yeast chromosomes—an important mainstream task in molecular biology that sheds light on how DNA operates in humans. Chromosomes are elongated structures that divide and separate during cell division; they carry all of yeast’s DNA. Szostak and his student, Andrew Murray, synthesized one from scratch. “Even then,” he recalls, “my guiding principle was that the best way to show that you really understand how something works is to try to build it from pieces and then see if it works as expected.” [Plate 8]

The study of yeast DNA was a crowded field, and Szostak itched to try his hand at something different. The discovery in the early 1980s of ribozymes—RNA that behaves as a catalyst—was too seductive to pass up, and he soon changed research directions.

In their first series of ribozyme experiments, Szostak and a group of his students focused on engineering chains of RNA that can copy other RNA molecules. Such an “RNA replicase” represents a first key step in designing a self-replicating chemical system. Success came in 1989, when he and Jennifer Doudna (now a professor at Berkeley) made an RNA molecule that copies short RNA sequences, albeit rather sloppily. Other high-profile papers soon followed, as their ability to manipulate RNA improved. These were spectacular advances, but still a long, long way from a reliable self-replicating RNA strand. They’d have to do better, but how? Jack decided to let molecular evolution work for him.

Szostak’s objective was to evolve an RNA molecule that would attach strongly and selectively to a “target molecule” of distinctive shape. His team tackled RNA evolution by first generating a solution with

more than 10 trillion random RNA sequences, each about 120 “letters” long. They poured this RNA-rich solution into a beaker whose glass sides had been coated with the target molecule. After sitting for a few minutes, the vast majority of RNA chains had done nothing, but a few RNA strands, by chance, were able to grab onto the target molecule and thus remained firmly attached to the beaker. When they flushed the RNA solution out of the beaker, those relatively few RNA sequences that bonded to the target molecule remained behind.

At first, only the tiniest fraction of the random RNA sequences attached to the target, but Szostak’s team collected those precocious strands and made trillions of approximate copies—similar sequences but with lots of random mutations thrown in. Then they repeated the experiment and picked out a second generation of RNA sequences that did the job better than those from the first. Again and again, they cycled the RNA, each time copying the best sequences and thus improving the speed and accuracy with which the RNA latched onto the glass-bound target molecules. After several dozen cycles, the surviving RNA strands had evolved to the point where they were perfectly adapted to the assigned task. This elegant evolutionary process has now been extended to the design of numerous new RNA sequences with a wide range of specialized functions, from locking onto viruses to splicing DNA.

With their new procedures, Jack Szostak and his colleagues have turned their attention to self-replicating RNA. In 2001, David Bartel, a former Szostak graduate student now at MIT and the Whitehead Institute, managed to produce an impressive RNA sequence that can grab onto a shorter piece of RNA up to 14 letters long and copy it. Optimistic researchers are convinced that it’s only a matter of time before self-replicating RNA strands more than a hundred letters long will be commonplace. If so, then synthetic life may not be far behind.

Self-replicating RNA is not alive, but it’s getting closer. It’s a big molecule that carries genetic information, catalyzes its own reproduction, and mutates and evolves to boot. But no plausible geochemical environment could feed such an unbound molecule, nor would it have survived long under most natural chemical conditions.

One key to survival is protection, and that’s where a lipid mem-

brane comes in. Building on David Deamer’s discoveries, Szostak protégés Martin Hanczyc and Shelley Fujikawa have experimented with ways that membranes might have encapsulated RNA strands. Among their findings: Lipid vesicles that are squeezed through tiny pores stretch out, divide, and start to grow larger—a process that mimics cell division. What’s more, the process is greatly accelerated by the addition of fine-grained clay minerals, some of which end up inside the vesicles. Recall that Jim Ferris at RPI demonstrated that clays can attract and help to assemble RNA strands. So, in the spirit of Pier Luigi Luisi’s Lipid World, it might be possible to make cell-like structures that spontaneously incorporate RNA-bound clay particles.

This behavior of lipids and RNA has led Szostak and his students to propose a remarkable scenario for the first life-form to evolve by natural selection. Imagine a lipid vesicle that contains self-replicating RNA. Previous authorities have suggested that RNA must have played many roles in such a protocell (roles that DNA and proteins play today)—manufacturing new membrane molecules, controlling cell shape and size, copying itself, and more.

Szostak’s team realized that RNA could drive cell growth by the much simpler process of internal pressure. RNA pushes out on the membrane, which in turn presses on neighboring protocells. They speculated that this contact would promote the transfer of lipids from cells with less internal pressure (hence, less RNA) to those with more. Thus the competition for space would lead to a natural-selection process. Protocells with more RNA would be more successful.

To test their ideas, they first prepared one set of vesicles filled with a solution of the sugar sucrose and another set filled with pure water. When mixed together and confined, the sucrose-ladened vesicles grew larger by drawing lipids from their sugarless neighbors. Repeating the experiment with RNA strands yielded the same results. Vesicles swollen with RNA grew, as the adjacent empty vesicles shrank.

Previous workers had assumed that RNA would have to learn how to accomplish several tasks—lipid synthesis, self-replication, metabolic functions, and more—before a protocell could evolve by natural selection. Szostak’s latest results suggest a much simpler scenario, in which the only essential task for protocell competition is RNA self-replication. “If we can get self-replicating RNAs,” Szostak suggests, “then we can put them into these simple membrane compartments and hope to actually see this competitive process of growth.” The more

RNA a vesicle captures and copies, the more successfully it will compete with its neighbors—a sort of molecular the-rich-get-richer scheme. In such a world, the most efficient RNA replicators would enjoy a tremendous advantage. With the emergence of competition, Darwinian evolution could take center stage.

Every year, Jack Szostak moves closer to his goal of creating a self-replicating, encapsulated, evolving chemical system in the lab. If and when he or his successors accomplish this, it will be a historic achievement. Synthetic life will also trigger a new flood of ethical questions about the potential dangers of scientific research, as well as philosophical questions about the meaning of life. But will synthetic life tell us how life emerged on Earth?

A synthetic RNA organism will certainly give credibility to the RNA World hypothesis that a strand of RNA (or some precursor genetic molecule) formed the basis of the first evolving, self-replicating chemical system. But laboratory-created life will not have emerged spontaneously from chemical reactions among the simple molecular building blocks of the prebiotic Earth. Researchers still stack the deck by supplying a steady source of RNA nucleotides and vesicle-forming lipids. And so, for the time being, deep mysteries remain.