4 billion years before present: the surface of a newly formed planet around a medium-sized star is beginning to cool down. It’s a violent place, bombarded by meteorites and riven by volcanic eruptions, with an atmosphere full of toxic gases. But almost as soon as water begins to form pools and oceans on its surface, something extraordinary happens. A molecule, or perhaps a set of molecules, capable of replicating itself arises.
This was the dawn of evolution. Once the first self-replicating entities appeared, natural selection kicked in, favouring any offspring with variations that made them better at replicating themselves. Soon the first simple cells appeared. The rest is prehistory.
Billions of years later, some of the descendants of those first cells evolved into organisms intelligent enough to wonder what their very earliest ancestor was like. What molecule started it all?
As far back as the 1960s, a few of those intelligent organisms began to suspect that the first self-replicating molecules were made of RNA, a close cousin of DNA. This idea has always had a huge problem, though – there was no known way by which RNA molecules could have formed on the primordial Earth. And if RNA molecules couldn’t form spontaneously, how could self-replicating RNA molecules arise? Did some other replicator come first? If so, what was it? The answer is finally beginning to emerge.
When biologists first started to ponder how life arose, the question seemed baffling. In all organisms alive today, the hard work is done by proteins. Proteins can twist and fold into a wild diversity of shapes, so they can do just about anything, including acting as enzymes, substances that catalyse a huge range of chemical reactions. However, the information needed to make proteins is stored in DNA molecules. You can’t make new proteins without DNA, and you can’t make new DNA without proteins. So which came first, proteins or DNA?
The discovery in the 1960s that RNA could fold like a protein, albeit not into such complex structures, suggested an answer. If RNA could catalyse reactions as well as storing information, some RNA molecules might be capable of making more RNA molecules. And if that was the case, RNA replicators would have had no need for proteins. They could do everything themselves.
It was an appealing idea, but at the time it was complete speculation. No one had shown that RNA could catalyse reactions like protein enzymes. It was not until 1982, after decades of searching, that an RNA enzyme was finally discovered. Thomas Cech of the University of Colorado in Boulder found it in Tetrahymena thermophila, a bizarre single-celled animal with seven sexes.
After that the floodgates opened. People discovered ever more RNA enzymes in living organisms and created new ones in their labs. RNA might be not be as good for storing information as DNA, being less stable, nor as versatile as proteins, but it was turning out to be a molecular jack of all trades. This was a huge boost to the idea that the first life consisted of RNA molecules that catalysed the production of more RNA molecules – “the RNA world”, as Harvard chemist Walter Gilbert dubbed it 25 years ago.
These RNA replicators may even have had sex. The RNA enzyme Cech discovered did not just catalyse any old reaction. It was a short section of RNA that could cut itself out of a longer chain. Reversing the reaction would add RNA to chains, meaning RNA replicators might have been able to swap bits with other RNA molecules. This ability would greatly accelerate evolution, because innovations made by separate lineages of replicators could be brought together in one lineage.Evolving replicators
For many biologists the clincher came in 2000, when the structure of the protein-making factories in cells was worked out. This work confirmed that nestling at the heart of these factories is an RNA enzyme – and if proteins are made by RNA, surely RNA must have come first.
Still, some issues remained. For one thing, it remained unclear whether RNA really was capable of replicating itself. Nowadays, DNA and RNA need the help of many proteins to copy themselves. If there ever was a self-replicator, it has long since disappeared. So biochemists set out to make one, taking random RNAs and evolving them for many generations to see what they came up with.
By 2001, this process had yielded an RNA enzyme called R18 that could stick 14 nucleotides – the building blocks of RNA and DNA – onto an existing RNA, using another RNA as a template (Science, vol 292, p 1319). Any self-replicating RNA, however, needs to build RNAs that are at least as long as itself – and R18 doesn’t come close. It is 189 nucleotides long, but the longest RNA it can make contains just 20.
A big advance came earlier this year, when Philipp Holliger of the MRC Laboratory of Molecular Biology in Cambridge, UK, and colleagues unveiled an RNA enzyme called tC19Z. It reliably copies RNA sequences up to 95 letters long, almost half as long as itself (Science, vol 332, p 209). To do this, tC19Z clamps onto the end of an RNA, attaches the correct nucleotide, then moves forward a step and adds another. “It still blows my mind that you can do something so complex with such a simple molecule,” Holliger says.
“It stills blows my mind that you can do something so complex with such a simple molecule”
So biologists are getting tantalisingly close to creating an RNA molecule, or perhaps a set of molecules, capable of replicating itself. That leaves another sticking point: where did the energy to drive this activity come from? There must have been some kind of metabolic process going on – but RNA does not look up to the job of running a full-blown metabolism.
“There’s been a nagging issue of whether RNA can do all the chemistry,” says Adrian Ferré-D’Amaré of the National Heart, Lung and Blood Institute in Bethesda, Maryland. RNA has only a few chemically active “functional groups”, which limit it to catalysing just a few types of chemical reaction.
Functional groups are like tools – the more kinds you have, the more things you can do. Proteins have many more functional groups than RNAs. However, there is a way to make a single tool much more versatile: attach different bits to it, like those screwdrivers that come with interchangeable heads. The chemical equivalents are small helper molecules known as cofactors.
Proteins use cofactors to extend even further the range of reactions they can control. Without cofactors, life as we know it couldn’t exist, Ferré-D’Amaré says. And it turns out that RNA enzymes can use cofactors too.
In 2003, Hiroaki Suga, now at the University of Tokyo, Japan, created an RNA enzyme that could oxidise alcohol, with help from a cofactor called NAD+ which is used by many protein enzymes. Months later, Ronald Breaker of Yale University found that a natural RNA enzyme, called glmS, also uses a cofactor.
Many bacteria use glmS, says Ferré-D’Amaré, so either it is ancient or RNA enzymes that use cofactors evolve easily. Either way, it looks as if RNA molecules would have been capable of carrying out the range of the reactions needed to produce energy.
So the evidence that there was once an RNA world is growing ever more convincing. Only a few dissenters remain. “The naysayers about the RNA world have lost a lot of ground,” says Donna Blackmond of the Scripps Research Institute in La Jolla, California. But there is still one huge and obvious problem: where did the RNA come from in the first place?
RNA molecules are strings of nucleotides, which in turn are made of a sugar with a base and a phosphate attached. In living cells, numerous enzymes are involved in producing nucleotides and joining them together, but of course the primordial planet had no such enzymes. There was clay, though. In 1996, biochemist Leslie Orgel showed that when “activated” nucleotides – those with an extra bit tacked on to the phosphate – were added to a kind of volcanic clay, RNA molecules up to 55 nucleotides long formed (Nature, vol 381, p 59). With ordinary nucleotides the formation of large RNA molecules would be energetically unfavourable, but the activated ones provide the energy needed to drive the reaction.
This suggests that if there were plenty of activated nucleotides on the early Earth, large RNA molecules would form spontaneously. What’s more, experiments simulating conditions on the early Earth and on asteroids show that sugars, bases and phosphates would arise naturally too. It’s putting the nucleotides together that is the hard bit; there does not seem to be any way to join the components without specialised enzymes. Because of the shapes of the molecules, it is almost impossible for the sugar to join to a base, and even when it does happen, the combined molecule quickly breaks apart.
This apparently insurmountable difficulty led many biologists to suspect to RNA was not the first replicator after all. Many began exploring the possibility that the RNA world was preceded by a TNA world, or a PNA world, or perhaps an ANA world. These are all molecules similar to RNA but whose basic units are thought to have been much more likely to form spontaneously. The big problem with this idea is that if life did begin this way, no evidence of it remains. “You don’t see a smoking gun,” says Gerald Joyce, also of the Scripps Research Institute.
In the meantime John Sutherland, at the MRC Laboratory of Molecular Biology, has been doggedly trying to solve the nucleotide problem. He realised that researchers might have been going about it the wrong way. “In each nucleotide, you see a sugar, a base and a phosphate group,” he says. “So you assume you need to make those building blocks first and then stick them together… and it doesn’t work.”
Instead he wondered whether simpler molecules might assemble into a nucleotide without ever becoming sugars or bases. In 2009, he proved it was possible. He took half a sugar and half a base, and stuck them together – forming the crucial sugar-base link that everyone had struggled with. Then he bolted on the rest of the sugar and base. Sutherland stuck on the phosphate last, though he found that it needed to be present in the mixture for the earlier reactions to work.Goldilocks chemistry
Sutherland was being deliberately messy by including the phosphate from the start, but it gave the best results. That’s encouraging: the primordial Earth was a messy place and it may have been ideal for making nucleotides. Sutherland now suspects there is a “Goldilocks chemistry” – not too simple, not too complex – that would produce many key compounds from the same melting pot.
“Sutherland had a real breakthrough,” Holliger says. “Everyone else was barking up the wrong tree.”
The issue isn’t entirely solved yet. RNA has four different nucleotides, and so far Sutherland has only produced two of them. However, he says he is “closing in” on the other two. If he succeeds, it will show that the spontaneous formation of an RNA replicator is not so improbable after all, and that the first replicator was most likely made of RNA.
Many questions remain, of course. Where did the first replicators arise? What was the first life like? How did the transition to DNA and proteins, and the development of the genetic code, occur? We may never know for sure but many promising avenues are being explored. Most biologists think there must have been something like a cell right from the start, to contain the replicator and keep its component parts together. That way, individuals could compete for resources and evolve in different ways.
Jack Szostak of Harvard University has shown that the same clay that produces RNA chains also encourages the formation of membrane-bound sacs rather like cells that enclose cells. He has grown “proto-cells” that can carry RNA and even divide without modern cellular machinery.
Another idea is that life began in alkaline hydrothermal vents on the sea floor. Not only are these vents laced with pores and bubbles, but they also provide the same kind of electrochemical gradient that drives energy production in cells to this day. Conditions may have been ideal for producing long RNA chains.
Holliger has a rather surprising idea: maybe it all happened in ice. At the time life began, the sun was 30 per cent dimmer than today. The planet would have frozen over if the atmosphere hadn’t been full of greenhouse gases, and there may well have been ice towards the poles.
Cold RNA lasts longer, and ice has many other benefits. When water laced with RNA and other chemicals is cooled, some of it freezes while the rest becomes a concentrated brine running around the ice crystals. “You get little pockets within the ice,” Holliger says. Interestingly, the R18 RNA enzyme works better in ice than at room temperature.
Right now, there’s no way to choose between these options. No fossilised vestiges remain of the first replicators as far as we know. But we can try recreating the RNA world to demonstrate how it might have arisen. One day soon, Sutherland says, someone will fill a container with a mix of primordial chemicals, keep it under the right conditions, and watch life emerge. “That experiment will be done.”