Figure 9.1: Map of Greenland, showing location of Isua.
… if (and oh! what a big if!) we could conceive [of] some warm little pond, with all sorts of ammonia and phosphoric salts, light, heat, electricity etc. present, that a protein compound was chemically formed ready to undergo still more complex changes …
CHARLES DARWIN, LETTER TO JOSEPH HOOKER, 1871
GREENLAND is not particularly green. Sometime around AD 982, a Danish Viking known as Erik the Red fled a charge of murder by sailing westward from Iceland and discovered the island. He wasn’t the first: it had already been discovered several times by stone age people who arrived from eastern Canada as early as 2500 BCE. But Greenland’s is a harsh and unforgiving environment, and those earlier cultures vanished, leaving only faint traces. Erik hoped to fare better, arriving during the so-called Medieval Warm Period when conditions were more clement; he therefore gave the island its current name, trusting that its promise of verdant pastures would lure his fellow countrymen westward. The ploy evidently worked, because a colony of several thousand was soon established and, initially at least, appeared to thrive. But as the warm period waned Greenland returned to climatic conditions more typical of the North Atlantic, and the central icecap grew to cover 80 percent of the island’s landmass. With the weather turning increasingly fierce, the islanders struggled to sustain their Scandinavian farming system in the shallow soil of the thin coastal strip, and both crop yields and livestock dwindled.
Ironically, at about the same time as the Viking colony was failing, another wave of immigrants, the Inuit (Eskimos), were making a living in the north of the island with a sophisticated fishing and hunting technology that was well adapted to the local conditions. The Vikings could have been saved if they had borrowed survival strategies from the Inuit, but the only record we have of contact between the two peoples is the remark from a Viking settler that the Inuit bleed a lot when stabbed—an observation that hardly indicates a willingness to learn from their northern neighbors. The result was that some time in the late fifteenth century the Viking colony collapsed, the last few inhabitants having apparently resorted to cannibalism.
However, the Danes never forgot about their western outpost and in the early eighteenth century an expedition was sent out to renew ties with the settlers. They found only abandoned homesteads and graveyards, but the visit did lead to the establishment of a more successful colony that, along with the native Inuit, eventually became the modern state of Greenland. The economy of Greenland today has grown from its Inuit roots, depending largely on fishing, but the potential mineral wealth of the island has been increasingly recognized. In the 1960s, the Danish Geological Survey of Greenland hired a young New Zealand–born geologist named Vic McGregor to conduct a geological study of the southwest corner of the island near its capital, Godthaab (now renamed Nuuk).
McGregor spent several years traveling through the fjord-riven region in a tiny, partly open boat, just big enough for himself, two local crew and the occasional guest, all crammed in among camping, hunting and fishing tools—not dissimilar from the kit of those early Inuit colonists—and geological equipment. Using standard techniques of stratigraphy, he concluded that the rocks in the area had been laid down in ten successive layers, of which the oldest and deepest was likely to be “very old indeed”—perhaps even more than three billion years old.
In the early 1970s, McGregor sent a sample of his ancient rock to the Oxford laboratory of Stephen Moorbath, a scientist who had established a reputation for radiometric dating of rocks. The method depends on measuring the ratio of radioactive isotopes and their decay products. For example, uranium 238 decays with a half-life of 4.5 billion years (through a chain of nuclides, eventually into a stable isotope of lead); so, as the earth is about four billion years old, the concentration of natural uranium in a rock will take the entire age of the earth to drop by half. By measuring the ratio of these isotopes in any sample of rock, scientists can therefore calculate how long it is since those rocks were laid down; and it was these techniques that Stephen Moorbath used in 1970 to analyze a sample of a type of rock called gneiss, which McGregor had chipped out of the southwest Greenland coastal region known as Amîtsoq. Amazingly, he discovered that the gneiss contained proportionally more lead than any terrestrial ore or rock ever reported. The finding of very high levels of lead meant the Amîtsoq gneiss rock was, as McGregor had guessed, “very old indeed”; at least 3.7 billion years old, older than any rock previously found on earth.
Figure 9.2: A modern mud volcano in Trinidad. Could the first life on earth have bubbled out of a similar mud volcano, leaving its traces in Isua greenstone? Photo: Michael C. Rygel via Wikipedia Commons.
Moorbath was so struck by the discovery that he then joined McGregor on several expeditions to Greenland. In 1971, the two of them decided to visit the remote and virtually unexplored Isua region on the edge of the inland ice sheet (see figure 9.1). They first had to sail in McGregor’s tiny boat up to the head of the iceberg-packed Godthaab Fjord, where the Viking settlers had eked out their precarious living in the middle ages. They were then picked up by a helicopter belonging to a local mining company that was also interested in the region, which aerial magnetic surveys had suggested was potentially rich in iron ore. The scientists discovered that within the local Isua greenstone there were many pillow-shaped masses of rock, known as basaltic pillow lavas and serpentine rocks formed by mud and gas extruded directly into seawater: so-called mud volcanoes. These rocks were again dated back to at least 3.7 billion years ago. The finding clearly demonstrated that the earth had liquid warm oceans not long after its formation,*1 with mud volcanoes (figure 9.2) bubbling out of hydrothermal vents at the bottom of a shallow sea.
However, the real surprise came when Minik Rosing, a researcher from the Geological Museum in Copenhagen, measured the ratio of carbon isotopes in the Isua greenstone. The rocks contain about 0.4 percent carbon, and when the respective ratios of the two isotopes 13C and 12C were measured, it was found that the amount of the heavier and rarer 13C in the rocks was much lower than expected. Inorganic sources of carbon, such as atmospheric carbon dioxide, have about 1 percent 13C, but photosynthesis prefers to incorporate the lighter 12C isotope into plant and microbial biomass, so a low level of 13C is generally an indicator of the presence of organic material. These results suggested that within the warm waters surrounding the Isua mud volcanoes 3.7 billion years ago there lived organisms that, like modern-day plants, were capturing carbon from carbon dioxide, either from the atmosphere or dissolved in water, and using it to construct all the carbon-based compounds that made up their cells.
The Isua rocks theory remains controversial, and many scientists are not convinced that the low levels of 13C found there necessarily imply so early a presence of living organisms. Much of the skepticism derives from the fact that 3.8 billion years ago the earth was in the throes of what is known as the “Late Heavy Bombardment,” suffering regular impacts from asteroids and comets with energies sufficient to vaporize any surface water and presumably also to sterilize the oceans. Discovery of fossils of any such ancient photosynthesizing organisms would of course clinch the case, but the Isua rocks have been severely deformed over the millennia and any such fossils would be unrecognizable. We have to skip forward at least several hundred million years before proof of the existence of life is clearly present in the form of recognizable fossils of ancient microbes.
Notwithstanding the lack of conclusive evidence, many believe that the Isua isotope data provide the earliest indications of life on earth; and the Isua mud volcanoes would certainly have provided an ideal environment for the emergence of life, with their warm alkaline waters spouting from thermal vents. They would have been rich in dissolved inorganic carbonates, and the extruded snake-like serpentine rocks, which are highly porous, would have been riddled with billions of tiny cavities, each of which could have been a microenvironment capable of concentrating and stabilizing tiny amounts of organic compounds. Perhaps life really did first become green in the mud of Greenland. The question is: How?
The three greatest mysteries in science are generally reckoned to be the origin of the universe, the origin of life and the origin of consciousness. Quantum mechanics is intimately involved in the first, and we have already discussed its possible connection to the third; as we will soon discover, it may also help to account for the second mystery. But we should first examine whether nonquantum explanations are able to provide a complete account of the origin of life.
The scientists, philosophers and theologians who have for centuries pondered the origin of life have come up with a rich variety of theories to explain it, ranging from divine creation to the seeding of our planet from space in the so-called panspermia theory. A more rigorously scientific approach was initiated in the nineteenth century by scientists such as Charles Darwin who proposed that chemical processes taking place in some “warm little pond” may have led to the creation of living material. The formal scientific theory that built upon Darwin’s speculations was put forward separately and independently by a Russian, Alexander Oparin, and an Englishman, J. B. S. Haldane, at the beginning of the twentieth century and is now generally known as the Oparin–Haldane hypothesis. Both proposed that the atmosphere of the early earth was rich in hydrogen, methane and water vapor that, when exposed to lightning, solar radiation or volcanic heat, combined to form a mixture of simple organic compounds. They proposed that these compounds then accumulated in the primordial ocean to form a warm, dilute organic soup, which swilled around in the water for millions of years, perhaps flowing over the Isua mud volcanoes, until some chance combination of its constituents eventually yielded a new molecule with an extraordinary property: the ability to replicate itself.
Haldane and Oparin proposed that the emergence of this primordial replicator was the key event that led to the origin of life as we know it. Its subsequent success would still have been subject to Darwinian natural selection. As a very simple entity, the replicator would have generated many errors or mutations in its replication. These mutant replicators would have then competed with nonmutated forms for the chemical materials from which to build more replicators. Those that were most successful would then have left the greatest number of descendants, and a molecular process of Darwinian natural selection would have taken hold to drive the swarm of replicators toward greater efficiency and greater complexity. Replicators that captured accessory molecules, such as peptides, that enzymatically catalyzed their replication would have gained an advantage, and some may even have become enclosed within vesicles (tiny fluid- or air-filled sacs) bounded by fatty membranes, as today’s living cells are, that protected them from the vagaries of their exterior environment. Once enclosed, the interior of the cell would then be able to support biochemical transformations—its metabolism—to make its own biomolecules and prevent them from leaking out. With the ability to maintain and sustain its internal state while keeping it separated from its environment, the first living cell would have been born.
The Oparin–Haldane hypothesis provided a scientific framework within which to understand how life could have originated on earth. Yet for several decades the theory went untested—until two American chemists took an interest.
By the 1950s, Harold Urey was a distinguished but controversial scientist. He had been awarded the Nobel Prize in chemistry in 1934 for discovering deuterium, the isotope of hydrogen that, as you may remember from chapter 3, was used to study the kinetic isotope effect in enzymes and thereby demonstrate that their activity involves quantum tunneling. Urey’s expertise in the purification of isotopes led to his appointment in 1941 as head of the uranium enrichment part of the Manhattan Project, which was attempting to develop the atomic bomb. However, Urey became disillusioned with the Manhattan Project’s aims and the secrecy in which it operated, and later attempted to dissuade the US president, Harry S. Truman, from dropping the bomb on Japan. After Hiroshima and Nagasaki, Urey wrote an article for the popular Collier’s magazine entitled “I’m a Frightened Man,” warning of the dangers posed by atomic weapons. From his post at the University of Chicago he also actively opposed McCarthy’s anticommunist “witch hunts” of the 1950s, writing letters to President Truman in support of Julius and Ethel Rosenberg, who were tried for espionage and eventually executed for passing atomic secrets to the Soviets.
Stanley Miller, the other American chemist involved in testing the Oparin–Haldane hypothesis, joined the University of Chicago as a PhD student in 1951, working initially on the problem of the nucleosynthesis of elements inside stars, under the guidance of the scientist known as the “father of the hydrogen bomb,” Edward Teller. Miller’s life changed when in October 1951 he attended a lecture given by Harold Urey on the origin of life, in which Urey discussed the feasibility of the Oparin–Haldane scenario and suggested that someone should do the experiments. Fascinated, Miller transferred from Teller’s to Urey’s lab and set about persuading Urey to become his PhD mentor and to allow him to carry out the experiments. Urey was initially skeptical about his enthusiastic student’s plans to put the Oparin–Haldane theory to the test: it might, he reckoned, take millions of years for inorganic chemical reactions to generate a sufficient number of organic molecules to be detected, while Miller had just three years to get his PhD! Nevertheless, Urey was prepared to give him the space and resources he needed for six months to a year. That way, if the experiments were not going anywhere, Miller would have time to switch to a safer research project.
In his attempt to replicate the conditions in which life originated on the early earth, Miller simulated the primordial atmosphere by simply filling a bottle with water, to represent the ocean, topped up with the gases that he thought would have been present in the atmosphere: methane, hydrogen, ammonia and water vapor. He then simulated lightning by igniting the mixture with electric sparks. To Miller’s surprise, and to the general astonishment of the scientific world, he discovered that after only a week of sparking his primordial atmosphere the bottle contained significant quantities of amino acids, the building blocks of proteins. The paper describing this experiment was published in the journal Science in 19531—with Miller as sole author, Harold Urey having adopted the highly unusual position of insisting that his PhD student gain full credit for the discovery.
The Miller–Urey experiment—as it is generally known today despite Urey’s unselfish gesture—was hailed as the first step in the creation of life in the laboratory, and remains a landmark in biology. Although no self-replicating molecules were generated, it was generally believed that Miller’s “primordial” soup of amino acids would have polymerized to form peptides and complex proteins and, given enough time and a sufficiently large ocean, eventually yield the Oparin–Haldane replicators.
Since the 1950s the Miller–Urey experiment has been repeated in many different ways by scores of scientists using different mixtures of chemicals, gases and energy sources to generate not only amino acids, but sugars and even small quantities of nucleic acids. And yet here we are, more than half a century later, with no laboratory-created primordial soup having yet yielded an Oparin–Haldane primordial replicator. To understand why, we need to look more closely at Miller’s experiments.
The first issue is the complexity of the chemical mixture that Miller generated. Much of the organic material produced was in the form of a complex tar, of the kind familiar to organic chemists who often see such substances whenever their complex chemical synthesis procedures are not strictly controlled and so lots of wrong products are made. In fact, it is easy to produce a similar tar in the comfort of your own kitchen just by burning the dinner: that blackish-brown gunk that is so hard to remove from the bottom of your pan is rather similar in composition to Miller’s tar. The problem with such chemical mixtures is that it is notoriously difficult to produce anything more than this tar-like gunk from them. In chemical terms, they are not what is called “productive,” because they are so complex that any specific chemical, such as an amino acid, tends to react with so many other different compounds that it then gets lost in a forest of inconsequential chemical reactions. Millions of cooks, and thousands of undergraduate chemistry students, have been producing such organic gunk for centuries, resulting in little more than a tough washing-up task.
Imagine trying to make a primordial soup by scraping all the gunk off the bottom of all the burnt pots in the entire world and then dissolving all those trillions of complex organic molecules into an ocean-sized volume of water. Now add a few Greenland mud volcanoes as your source of energy, and perhaps the spark of lightning, and stir. How long would you have to stir your soup before you created life? A million years? A hundred million years? A hundred billion years?
Even the simplest life is, much like this chemical gunk, extraordinarily complex. Unlike gunk, however, it is also highly organized. The problem with using gunk as the starting material for generating organized life is that the random thermodynamic forces that were available in the primordial earth—the billiard-ball-like molecular motions that we discussed in chapter 2—tend to destroy order rather than create it. You throw a chicken into the pot, heat it up and stir it, and make chicken soup. No one has ever poured a can of soup into a pot and made a chicken.
Of course, life didn’t start with chickens (or eggs). The most basic self-replicating organisms alive today are bacteria, which are far simpler than any bird.*2 The simplest is called a mycoplasma (the bacterium that was the subject of Craig Venter’s synthetic life experiment); but even these creatures are extremely complex life forms. Their genome encodes nearly five hundred genes, which produce a similar number of highly complex proteins that, as enzymes, make lipids, sugars, DNA, RNA, the cell membrane, its chromosome and a thousand other structures, each far more intricate than your car engine. And, in reality, mycoplasma is actually a bit of a bacterial wimp as it cannot survive on its own and must obtain many of its biomolecules from its host: it is a parasite and, as such, would be unable to survive in any realistic primordial soup. A more likely candidate would be another single-celled organism called a cyanobacterium that is able to photosynthesize to make all its own biochemicals. If present on the early earth, these cyanobacteria would have been a potential source of those low levels of 13C detected in the 3.7-billion-year-old Isua rocks in Greenland. But this bacterium is much more complex than a mycoplasma, with a genome encoding nearly two thousand genes. How long would you have to stir your ocean of primordial soup to make a cyanobacterium?
The British astronomer who coined the term “Big Bang,” Sir Fred Hoyle, had an interest in the origins of life that lasted throughout his own lifetime. The probability of random chemical processes coming together to generate life, he said, was as likely as a tornado blowing through a junkyard and assembling a jumbo jet by chance. The point he was making so vividly was that cellular life, as we know it today, is just too complex and organized to have arisen by chance alone; it must have been preceded by simpler self-replicators.
So what were those early self-replicators like? And how did they work? As none survive today, presumably because they have been out-competed into extinction by their more successful descendants, their nature is mostly educated guesswork. One approach is to extrapolate backward from the simplest life forms alive today to imagine a much simpler self-replicator, a kind of stripped-down bacterium that may have been the precursor, billions of years ago, of all life on earth.
The problem is that it’s not possible to dissect simpler self-replicators out of living cells because none of the components of cells are capable of self-replication by themselves. DNA genes don’t replicate themselves; that is the job of the DNA polymerase enzymes. In turn, those enzymes don’t replicate themselves, for they need to be first encoded within DNA and RNA strands.
RNA will play an important role in this chapter, so it may be useful to recall what it is and what it does. RNA is DNA’s simpler chemical cousin, and it comes as a single-stranded helix compared with DNA’s double helix. Despite this difference, RNA has more or less the same genetic information coding capacity as its more famous cousin—it just doesn’t have the complementary copy of that information. And, just like DNA, its genetic information is written in four different genetic letters, so genes can be encoded in RNA just as they can be in DNA. Indeed, many viruses, such as the influenza virus, possess RNA genomes, rather than DNA genomes. But in living cells such as bacteria, animal or plant cells, RNA performs a role distinct from DNA: the genetic information written into DNA is first copied into RNA in the gene-reading process that we discussed in chapter 7. And since, unlike the relatively massive and immobile DNA chromosome, shorter RNA strings are free to move around the cell, they can carry the genetic message of genes from the chromosome to the protein synthesis machinery. Here the RNA sequence is read and translated into the sequences of amino acids that go into proteins, such as enzymes. So, in modern cells at least, RNA is a key intermediary between the genetic code written in DNA and the proteins that go on to make all the other components of our cells.
Returning then to our origin-of-life problem, although a living cell as a whole is a self-replicating entity, its individual components are not; just as a woman is a self-replicator (with a little “help”), but her heart or liver is not. This creates a problem when trying to extrapolate backward from today’s complex cellular life to its much simpler noncellular ancestor. If you put it another way, the question becomes: Which came first: the DNA gene, the RNA, or the enzyme? If DNA or RNA came first, then what made them? If the enzyme came first, then how was it encoded?
One possible solution was provided by the American biochemist Thomas Cech, who discovered in 1982 that as well as encoding genetic information, some RNA molecules could take on the job of enzymes to catalyze reactions (work for which he shared the 1989 Nobel Prize in chemistry with Sidney Altman). The first examples of these ribozymes, as they are known, were found in the genes of tiny single-celled organisms called Tetrahymena, which is a type of protozoan found in freshwater ponds; but ribozymes have since been found to play a role in all living cells. Their discovery was quickly seized upon as a possible way out of the chicken-and-egg origin of life conundrum. The RNA world hypothesis, as it came to be known, proposes that primordial chemical synthesis resulted in the generation of an RNA molecule that could act as both gene and enzyme, and thus could both encode its own structure (like DNA) and make copies of itself (like enzymes) out of the biochemicals available in the primordial soup. This copying process would initially have been very hit-and-miss, giving rise to lots of mutant versions that would have competed against one another in the molecular Darwinian competition envisaged earlier. Over the course of time, those RNA replicators would have recruited proteins to improve their replication efficiency, leading to DNA and eventually the first living cell.
The idea that a world of self-replicating RNA molecules preceded the emergence of DNA and cells is now almost dogma in origin-of-life research. Ribozymes have been shown to be able to perform all the key reactions expected of any self-replicating molecule. For example, one class of ribozymes can join two RNA molecules together, whereas another can break them apart. Yet another form of ribozyme can make copies of short strings (just a handful of bases long) of RNA bases. From these simple activities we can imagine a more complex ribozyme able to catalyse the complete set of reactions necessary for self-replication. Once self-replication kicks in, then so too does natural selection; so the RNA world would have been set on a competitive path that led eventually, or so it is argued, to the first living cell.
There are, however, several problems with this scenario. Although simple biochemical reactions may be catalyzed by ribozymes, self-replication of a ribozyme is a far more complex process involving recognition by the ribozyme of the sequence of its own bases, identification of identical chemicals in the ribozyme’s environment, and assembly of those chemicals in the correct sequence to make a replica of itself. This is a tall order even for proteins having the luxury of living within cells packed full of the correct biochemicals, so it is even harder to see how ribozymes surviving in the messy and gunky primordial soup could achieve this feat. To date, no one has discovered or succeeded in making a ribozyme that can undertake such a complex task, even in the laboratory.
There is also the more fundamental problem of how to make the RNA molecules themselves in the primordial soup. The molecule is made of three pieces: the RNA base that encodes its genetic information (just as DNA bases encode the DNA’s genetic information), a phosphate group and a sugar called ribose. Although some success has been achieved in devising plausible chemical reactions that might have made the RNA bases and phosphate components in the primordial soup, the most credible reaction that yields ribose also produces a plethora of other sugars. There is no known nonbiological mechanism by which the ribose sugar can be generated on its own. And even if the ribose sugar were made, putting all three components together correctly is itself a formidable task. When plausible forms of the three components of RNA are brought together, they just combine in arbitrary ways to form the inevitable primordial gunk. Chemists avoid this problem by using special forms of bases whose chemical groups have been modified to avoid those unwanted side reactions—but this is cheating; and, in any case, the “activated” bases are even more unlikely to have been formed in primordial conditions than the original RNA bases.
However, chemists are able to synthesize the RNA bases from simple chemicals by going through a very complex series of carefully controlled reactions in which each desired product from one reaction is isolated and purified before taking it on to the next reaction. The Scottish chemist Graham Cairns-Smith estimated that there are about 140 steps necessary for the synthesis of an RNA base from simple organic compounds likely to have been present in the primordial soup.2 For each step there is a minimum of about six alternative reactions that need to be avoided. This makes the chemical synthesis easy to visualize, for you can conceive of each molecule as a kind of molecular die, with each step corresponding to a throw where the number six represents generating the correct product and any other number indicates that the wrong product has been made. So, the odds of any starting molecule eventually being converted into RNA is equivalent to throwing a six 140 times in a row.
Of course, chemists improve these stupendous odds by carefully controlling each step, but the prebiotic world would have had to rely on chance alone. Perhaps the sun came out at just the right time to evaporate a little pool of chemicals surrounding a mud volcano? Or perhaps the mud volcano erupted to add water and a little sulphur to create another set of compounds? Perhaps a lightning storm stirred up the mix and accelerated a few more chemical changes with the input of electrical energy? The questions could go on and on; but it’s easy enough to estimate the probability that, relying on chance alone, each of the 140 necessary steps would have yielded the right one of six possible products: it is one in 6140(roughly, 10109). To have a statistical chance of making RNA by purely random processes you would need at least this number of starting molecules in your primordial soup. But 10109 is a far bigger number than even the number of fundamental particles in the entire visible universe (about 1080). The earth simply did not have enough molecules, or sufficient time, to make significant quantities of RNA in those millions of years between its formation and the emergence of life at the time suggested by the Isua rocks.
Nevertheless, imagine that the synthesis of significant quantities of RNA did happen, through some as yet undiscovered chemical process. We now have to overcome the equally daunting problem of stringing the four different RNA bases (equivalent, you’ll remember, to those four letters of the DNA code, A, G, C and T) together in just the right sequence to make a ribozyme capable of self-replication. Most ribozymes are RNA strings at least a hundred bases long. At each position in the string one of the four bases must be present, so there are 4100 (or 1060) different ways to put together a string of RNA 100 bases long. How likely is it that the random jumbling together of RNA bases will generate just the right sequence along the length of the string to make a self-replicating ribozyme?
Since we seem to be having such fun with big numbers, we can work it out. It turns out that 4100 individual strings of RNA 100 bases long would have a combined mass of 1050 kilograms. So this is how much we would need, in order to have a single copy of most strings and therefore a reasonable chance that one of them would have all its bases arranged correctly to be a self-replicator. However, the entire mass of the Milky Way galaxy is estimated to be approximately 1042 kilograms.
Clearly, we cannot rely on pure chance alone.
Of course, there may not be just one arrangement among the 4100 possible 100 base-long RNA strings that would act as a self-replicator. There may be many more. There could even be trillions of possible replicators that can be formed out of RNA strings 100 bases long. Perhaps self-replicating RNA is actually quite common, and we only need a million molecules to have some chance of forming a self-replicator. The problem with this argument is that it is just that: an argument. Despite many attempts, no one has ever made a single self-replicating RNA (or DNA, or protein), or observed one in nature. This is not so surprising when you consider what a challenging job self-replication is. In today’s world it takes an entire living cell to achieve this feat. Could it have been done with a far simpler system billions of years ago? Surely it must have, or we wouldn’t be here contemplating the problem today. But how this was achieved before cells evolved is far from clear.
Given the difficulties of identifying biological self-replicators, we might gain insight by asking a more general question: How easy is self-replication in any system? Modern technology has provided us with lots of machines that can replicate stuff, from photocopying machines to electronic computers to 3-D printers. Can any of these devices make a copy of itself? Probably the closest is a 3-D printer such as one of the RepRap (short for Replicating Rapid prototyper) printers that are the brainchild of Adrian Bowyer at the University of Bath in the United Kingdom. These machines can print their own components, which can then be assembled to make another RepRap 3-D printer.
Well, not quite. The machine only prints in plastic, but its own frame is made of metal, as are most of its electrical components. So it is only the plastic parts that it can replicate; and these have to be manually assembled with additional parts to make a new printer. The vision of the designers is to make self-replicating RepRap printers (there are several alternative designs) freely available for the benefit of everyone. But at the time of writing we are a long way from building a truly self-replicating machine.
So, if looking for self-replicating machines doesn’t really help us in our quest to discover how easy or difficult self-replication is, can we eschew the material world entirely and explore this question within a computer, where those messy and hard-to-make chemicals can be replaced by the simple building blocks of the digital world: namely, the bits that can only have a value of either 1 or 0? A “byte” of data, consisting of eight bits, represents a single character of text in a computer code and can be roughly equated with the unit of genetic code: a DNA or RNA base. We can now ask the question: Among all the possible strings of bytes, how common are those that can replicate themselves in a computer?
Here we have a huge advantage, because self-replicating strings of bytes are actually quite common: we know them as computer viruses. These are relatively short computer programs that can infect our computers by persuading their CPU to make loads of copies. These computer viruses then hop into our e-mails to infect the computers of our friends and colleagues. So if we consider the computer memory as a kind of digital primordial soup, then computer viruses can be considered to be the digital equivalent of primordial self-replicators.
One of the simplest computer viruses, Tinba, is only 20 kilobytes long: very short compared to most computer programs. Yet Tinba successfully attacked the computers of large banks in 2012, burrowing into their browsers and stealing login data; so it was clearly a formidable self-replicator. While 20 kilobytes may be very short for a computer program, it nonetheless comprises a relatively long string of digital information as, with 8 bits in a byte, it corresponds to 160,000 bits of information. Since each bit can be in one of two states (0 or 1) we can easily calculate the probability of randomly generating particular strings of binary digits. For example, the chances of generating a particular three-bit string, say, 111, is ½ × ½ × ½, or 1 chance in 23. Following the same mathematical logic, it follows that arriving by accident at a specific string 160,000 bits long, the length of Tinba, is 1 chance in 2160,000. This is a mind-bogglingly small number, and tells us that Tinba could not have arisen by chance alone.
Perhaps there are, just as we conjectured for RNA molecules, very many self-replicating codes out there that are far simpler than Tinba and that might have arisen by chance. But if that were the case, then surely a computer virus would, by now, have arisen spontaneously from all the zillions of gigabytes of computer code that are flowing through the Internet every second. Most of these codes are after all just sequences of ones and zeros (think of all the images and movies that are being downloaded every second). These codes are all potentially functional in terms of instructing our CPUs to perform basic operations, such as to copy or to delete; yet all of the computer viruses that have ever infected anyone’s computer show the unmistakable signature of human design. As far as we know, the vast stream of digital information that flows around the world every day has never spontaneously generated a computer virus. Even within the replication-friendly environment of a computer, self-replication is hard, and, so far as we know, it has never happened spontaneously.
This excursion into the digital world exposes the essential problem in the quest for life’s origin, which boils down to the nature of the search engine used to bring its necessary ingredients together in the correct configuration to form a self-replicator. Whatever chemicals were available in the primordial soup, they would have had to explore a huge space of possibilities to hit upon an exceedingly rare self-replicator. Could our problem be that we are confining the search routine to the rules of the classical world? You may remember from chapter 4 that the quantum theorists at MIT were initially highly skeptical of the New York Times report that plants and microbes were implementing a quantum search routine. But they eventually came around to the idea that photosynthetic systems were indeed implementing a quantum search strategy, called a quantum walk. Several researchers, ourselves included,3 have explored the idea that the origin of life could similarly have involved some kind of quantum search scenario.
Imagine a tiny primordial pool enclosed within a pore of those serpentine rocks extruded from a mud volcano under the ancient Isua sea three and a half billion years ago, when Greenland’s gneiss strata were being formed. Here is Darwin’s “warm little pond, with all sorts of ammonia and phosphoric salts, light, heat, electricity etc. present,” in which “a protein compound … ready to undergo still more complex changes” might have formed. Now, further imagine that one “protein compound” (it could just as easily be an RNA molecule), made by the kind of chemical processes that Stanley Miller discovered, is a kind of proto-enzyme (or ribozyme) that has some enzymatic activity but is not yet a self-replicating molecule. Further imagine that some of the particles in this enzyme could move to different positions but are prevented from doing so by classical energy barriers. However, as we discussed in chapter 3, both electrons and protons are able to quantum tunnel through energy barriers that forbid their classical transfer, a feature that is crucial in enzyme action. In effect, the electron or proton exists on both sides of the barrier simultaneously. If we imagine this happening within our proto-enzymes, then we would expect the different configurations—finding the particle on either side of the energy barrier—to be associated with different enzyme activities, that is, abilities to accelerate different types of chemical reactions, perhaps including a self-replication reaction.
Just to make the numbers easy to work with, let us imagine there are a total of 64 protons and electrons within our imaginary proto-enzyme that are each capable of quantum tunneling into any one of two different positions. The total amount of structural variation available to our imaginary proto-enzyme is still enormous: 264—an awful lot of possible configurations. Now imagine that just one of these configurations has what it takes to become a self-replicating enzyme. The question is: How easy is it to find the particular configuration that could lead to the emergence of life? Will the self-replicator ever be realized in our tiny warm pond?
Consider first the proto-enzyme as an entirely classical molecule unable to do any quantum tricks, such as superposition or tunneling. The molecule must, at any given moment, be in just one of the possible 264 different configurations, and the probability that this proto-enzyme will be a self-replicator is 1 divided by 264—an exceedingly small chance indeed. With overwhelming odds, the classical proto-enzyme will be stuck in one of the boring configurations that can’t self-replicate.
Of course, molecules do change, as a result of general thermodynamic wear and tear, but in the classical world such change is relatively slow. For one molecule to change, the original arrangement of atoms must be dismantled and its constituent particles rearranged to form a new molecular configuration. As we discovered in chapter 3 with the long-lived dinosaur collagen, chemical changes can sometimes take place over geological timescales. Considered classically, our proto-enzyme would take a very long time to explore even a tiny fraction of those 264 chemical configurations.
However, the situation is radically different if we consider the 64 key particles in the proto-enzyme to be electrons and protons that can tunnel between their alternative positions. Being a quantum system, the proto-enzyme can exist in all its possible configurations simultaneously as a quantum superposition. The reason for our choice of the number 64 above now becomes clearer; it is the same number we explored when we were using the Chinese emperor’s chessboard blunder to illustrate the power of quantum computing in chapter 8, with the tunneling particles taking the role of the squares on the board or qubits. Our proto-self-replicator could, if it survived long enough, act as a 64-qubit quantum computer; and we have already discovered how powerful such a device would be. Perhaps it can use its huge quantum computational resources to compute the answer to the question: What is the correct molecular configuration for a self-replicator? In this guise, the problem and its potential solution become clearer. Consider the proto-enzyme to be in such a quantum superposition, and the search problem of finding the one in 264 possible structures that is the self-replicator becomes solvable.
There is a hitch, though. You will remember that qubits have to remain coherent and entangled in order to perform quantum computing. Once decoherence kicks in, the superposition of 264 different states collapses and just one remains. Does this help? On the face of it, no, because the chance of the quantum superposition collapsing into the single self-replicating state is the same as before: a minuscule 1 divided by 264, the same as the chances of throwing heads 64 times in a row. But what happens next is where the quantum description diverges from its classical counterpart.
If a molecule is not behaving quantum mechanically and finds itself, as it almost certainly will, with the wrong arrangement of atoms that is unable to self-replicate, trying out a different configuration would have to involve the geologically slow process of dismantling and rearranging molecular bonds. But, after decoherence of the equivalent quantum molecule, each of the 64 electrons and protons of our proto-enzyme will, almost instantaneously, be ready to tunnel again into a superposition of both of their possible positions to reestablish the original quantum superposition of 264 different configurations. In its 64-qubit state, the quantum proto-replicator molecule could repeat its search for self-replication in the quantum world continuously.
Decoherence will rapidly collapse the superposition once again; but this time the molecule will find itself in another of its 264 different classical configurations. Once again, decoherence will collapse the superposition, and once again the system will find itself in another configuration; and this process will continue indefinitely. Essentially, in this relatively protected environment, the making and breaking of the quantum superposition state is a reversible process: the quantum coin is being continually tossed by the processes of superposition and decoherence, processes that are far more rapid than the classical making and breaking of chemical bonds.
But there is one event that will terminate the quantum coin-tossing. If the quantum proto-replicator molecule eventually collapses into a self-replicator state, it will start to replicate and, just as in the starving E. coli cells we discussed in chapter 7, replication will force the system to make an irreversible transition into the classical world. The quantum coin will have been irreversibly thrown, and the first self-replicator will have been born into the classical world. Of course, this replication will have to involve some sort of biochemical process within the molecule, or between it and its surroundings, that is distinctly different from those that took place before the proto-replicator arrangement was found. In other words, there needs to be a mechanism that anchors this special configuration in the classical world before it is lost and the molecule moves on to the next quantum arrangement.
The proposition we have outlined above is, of course, speculative. But if the search for the first self-replicator was performed in the quantum rather than the classical world, it does at least potentially solve the self-replicator search problem.
In order for this scenario to work, the primordial biomolecule—the proto-self-replicator—must have been capable of exploring lots of different structures by the quantum tunneling of its particles into different positions. Do we know what kind of molecules would be capable of such a trick? Well, to a certain extent we do. As we have already discovered, the electrons and protons in enzymes are held relatively loosely, which enables them to tunnel into different positions with ease. The protons in DNA and RNA are also capable of tunneling, at least across the hydrogen bond. So we might imagine our primordial self-replicator to be something like a protein or RNA molecule that was loosely held together by hydrogen bonds and weak electronic bonds that allowed its particles—both protons and electrons—to travel freely through its structure to form a superposition of its trillions of different configurations.
Is there any evidence for such a scenario? Apoorva D. Patel, a physicist at the Centre for High Energy Physics at the Indian Institute of Science in Bangalore, is one of the world’s experts on quantum algorithms—the software of quantum computers. Apoorva suggests that aspects of the genetic code (the sequences of DNA bases that code for one amino acid or another) betray its origin as a quantum code.4 This is not the place to go into any technical detail (for this would take us too deeply into the mathematics of quantum information theory), but his idea should not come as such a surprise. In chapter 4 we saw how, in photosynthesis, the photon’s energy is transferred to the reaction center by following multiple pathways at once—a quantum random walk. Then, in chapter 8, we discussed the idea of quantum computation and whether life might make use of quantum algorithms to enhance the efficiency of certain biological processes. Similarly, origin-of-life scenarios that involve quantum mechanics, while speculative, are nothing more than an extension of these ideas: the possibility that quantum coherence in biology played the kind of role in the origin of life as it currently does in living cells.
Of course, any scenario involving quantum mechanics in the origin of life three billion years ago remains highly speculative. But, as we have discussed, even classical explanations of life’s origin are beset with problems: it isn’t easy to make life from scratch! By providing more efficient search strategies, quantum mechanics may have made the task of building a self-replicator a little easier. It almost certainly was not the whole story; but quantum mechanics could have made the emergence of life in those ancient Greenland rocks a lot more likely.
