Living by RNA Alone

There may be ribozymes and regulatory RNAs in modern organisms, but why are there no longer any free-living descendants of the RNA world?25 Why do even the simplest organisms rely on DNA and protein for sequence storage and catalytic function? The answer seems to be that systems built exclusively of RNA sequences and ribozymes are subject to limitations which compromise their ability to cross von Neumann’s threshold of complication.

In particular, the RNA world suffered from three handicaps. First, RNA is not a stable molecule. Second, RNA replication is subject to a high error rate. And third, as an enzyme, the ribozyme is not terribly potent. These disadvantages do not mean the RNA world never existed—on the contrary, the evidence is strong that it did—but they did place limits on its scalability and evolutionary potential.26

The first handicap is instability. In The Wizard of Oz, the Wicked Witch of the West is unstable in water. So is RNA. Unlike DNA, which is stored as two complementary strands that reinforce each other, RNA is stored as a single strand. The single-stranded RNA molecule is susceptible to hydrolysis; in water the backbone of the sequence can cleave spontaneously.27 Given that water is ubiquitous and essential to life, this is not a good thing. It is especially problematic in replication, which must be fast enough to outrun decay. “In order for a seed copy of replicating RNA to germinate,” says Joyce, “it must produce additional copies of itself faster than the existing copies become degraded.”28

This leads to the second handicap, a high mutation rate in replication. In the living world there is an ever-present tradeoff between speed and reliability; the faster the replication, the lower the fidelity. Compared with DNA, the copying of RNA sequences is intrinsically error-prone.29 The evolutionary impact of this is to limit genome size and complexity. “If copying is very error ridden, and selection is weak,” say Kim Sterelny and colleagues, “then noise can swamp selection, and cumulative selection will be unable to build complexly adapted interactors.”30

In the RNA world, sequences were under selection pressure to develop error-correcting mechanisms to improve replication fidelity.31 We don’t know how well they succeeded. “When self-replication is first established, fidelity is likely to be poor and there is strong selection pressure favoring improvement of the fidelity,” write Joyce and Orgel.

As fidelity improves, a larger genome can be maintained. This allows exploration of a larger number of possible sequences, some of which may lead to further improvement in fidelity, which in turn allows a still larger genome size, and so on.32

Here we begin to get a glimpse of the runaway explosive quality von Neumann is after.

RNA may be unstable and may suffer from poor replication fidelity, but it gets worse. The third handicap is that, compared to protein enzymes, ribozymes are just not very good catalysts. Modern proteins are “intrinsically a millionfold fitter as catalysts than RNA,” write Steven Benner and colleagues.33 This is partly because ribozymes are so much smaller than proteins but also because the chemical bonds that give a folded ribozyme its shape are not that strong. Remember that ribozymes need to fold for function but also must unfold for replication. There is a “tradeoff between favorable folding stability and ease of unfolding,” write computational biologist Abe Pressman and colleagues. “More effective ribozymes, with likely higher folding stability, are expected to be less prone to unfold for use as a replication template, and would therefore have lower fitness.”34

These built-in handicaps kept a lid on the evolutionary potential of the RNA world. “The replicating RNA enzyme is the only known molecule that can undergo self-sustained Darwinian evolution,” write Joyce and biochemist Michael Robertson, “but it has limited genetic complexity, and therefore limited capacity for the invention of novel function.”35 The diversity and complexity we see in the living world today only arose when these problems were solved. DNA was recruited to improve storage stability and replication fidelity, and proteins were recruited to improve catalytic function. Division of labor allowed the RNA world to cross the threshold of complication.36,37

“In the beginning, the same entities had to perform both functions necessary for selection,” writes David Hull. “However, because replication and interaction are fundamentally different processes, the properties which facilitate them tend also to be different. None too surprisingly, these distinct functions eventually were apportioned to different entities.”38

As a storage medium for one-dimensional patterns, DNA has advantages over RNA. First, the DNA backbone is more stable and less prone to hydrolysis.39 Second, DNA is stored in a double-stranded form in which complementary base pairing adds another layer of stability and error correction.40 DNA also replicates with higher fidelity than RNA, which allows for a substantial increase in genome size and complexity. “The primary advantage of DNA over RNA as a genetic material is the greater chemical stability of DNA,” says Joyce, “allowing much larger genomes based on DNA. Protein synthesis may require more genetic information than can be maintained by RNA.’41 And crucially for regulatory function, a stable sequence like DNA enables random access; if you are searching for a specific one-dimensional pattern, it is there to be found. DNA sequences were, in Woese’s words, “something unique in the (RNA) world, namely, nucleic acids whose primary value lay in their coding capacity.”42

Replacing RNA with DNA for storage and replication of linear patterns was one key step in the division of labor. The other was the replacement of weakly catalytic ribozymes with strongly catalytic protein enzymes, which improve upon ribozymes in three ways. First, their size and binding specificity enable them to differentiate among a much larger class of potential substrates. Second, their size and complexity enable a larger range of allosteric conformational changes. Finally, their stronger bonds and consequent ability to remain folded allows their catalytic power to increase by orders of magnitude.

Proteins are so much more versatile and powerful that once nature figured out how to use protein enzymes, there was no way any legacy ribozyme-based system could remain dominant. “A life form that did not exploit proteins as catalysts could not have competed with life that did,” write Benner and colleagues.43 This is why there are no free-living RNA-only descendants of the RNA world. Concludes Joyce: “The invention of protein synthesis, instructed and catalysed by RNA, was the crowning achievement of the RNA world, but also began its demise.’44

What are we to make of the RNA world? While remarkable, it was limited in its ability to achieve scale and explosive evolutionary creativity. It “should be viewed as a milestone,” say Robertson and Joyce, “a plateau in the early history of life on earth.”43 We do not know how long it took for life to upgrade from the ancient RNA-only system to the more robust and versatile DNA-RNA- protein system. Nor do we know the steps involved, whether DNA or protein came first or if they were incorporated simultaneously. But we do know this transition allowed the system to cross von Neumann’s threshold of complication, making possible the diversity of living forms on Earth today.