Goz's Quote of the Week 18/09/2019

The RNA World

Because RNA has the capacity for heredity and metabolism, evolutionary biologists have been fascinated by its role in early life. Gerald Joyce describes this fascination as an archaeological mystery:



In 1986, Walter Gilbert coined the phrase “RNA world” to capture the idea that early evolution—from about 4 billion to 3.5 billion years ago—may have been RNA based (Gilbert 1986). There is some evidence from laboratory work mimicking conditions on early Earth that ribose, phosphate, purines, and pyrimidines—all the essential parts of RNA—likely existed in the prebiotic environment (Robertson and Miller 1995a). Evolutionary biologists hypothesize that these molecules may have emerged from conditions similar to those simulated in the prebiotic-soup experiment, or they may have arrived on meteorites, which often contain high amounts of carbon. In either case, if these compounds had bonded together, RNA-based life-forms might have resulted (Benner et al. 1989; Joyce 2002).

Experimental Evidence on the Origins of Natural Selection

A critical step in understanding how basic chemical reactions could lead to entities capable of variation, multiplication, and heritability came in a fascinating experiment by Sol Spiegelman and his colleagues (Mills et al. 1967; Spiegelman
1970). In an early experiment on the origins of natural selection, Spiegelman and his colleagues placed a “primer” strand of RNA, made up of about 4000 RNA nucleotides—adenine (A), guanine (G), cytosine (C), and uracil (U)—into a small
test tube. To this mixture, they added more A, G, C, and U nucleotides, and a replicase enzyme, which functions to make copies of an RNA molecule. The researchers heated and incubated the mixture, and then they transferred a small drop to a new test tube.

The new test tube contained the replicase enzyme and nucleotides A, G, C, and U, but it did not contain primer RNA. This new test tube was heated and incubated, followed by yet another transfer to a new test tube, and so on, for 75 serial transfers. Spiegelman and his colleagues found that RNA made copies of itself in these test tubes. The interesting thing was not that the RNA was copied—Spiegelman had added replicase enzyme to ensure that this would happen—but rather that natural selection took place on these copies, leading to a change in their characteristics. Let us examine how.

The process of RNA replication carried out by the replicase enzyme involved errors; these errors produced new mutant forms of RNA that differed both in their length and their nucleotide sequence. This generated variation in RNA types on which natural selection might act. Because the replicase enzyme copies whatever strands are present, these changes should be heritable as well.

We already have mentioned that two of the three ingredients for natural selection were present in the experiment: variation and heritability (which was built into the design of Spiegelman’s experiment). The third ingredient for natural selection is differential survival or replication, and we would expect this to be present in Spiegelman’s experiment as well. We know that shorter RNA sequences will replicate more quickly, and so, in general, selection should favor sequences shorter than the original 4000-nucleotide primer strand. But, if RNA is very short—fewer than 50–100 nucleotides or so—the error rate in replication is so very high that these strands can no longer be copied reliably by the replicase enzyme. As a result, we would expect strands shorter than 50–100 base pairs to be selected against.

At the end of the serial transfer experiment, if natural selection was operating in Spiegelman’s test tubes, we would expect to see a strand that is less than 4000 nucleotides long and greater than 50–100 nucleotides in length. And that is what Spiegelman found—a strand that was a little over 200 nucleotides long. Apparently, selection for a moderately short strand was very strong, as the end product RNA was much closer to the 50–100 nucleotide minimal length we discussed than the 4000-nucleotide-long original primer. Shorter RNA sequences, which took a shorter time to replicate, were favored by selection.

Using Spiegelman’s protocol, Manfred Sumper and his colleagues further examined natural selection and the early stages of RNA-based life (Kuppers and Sumper 1975; Sumper and Luce 1975; Eigen et al. 1981). They ran an experiment similar to the one described earlier; in addition, they added a chemical called acridine orange. Acridine orange is a dye used in fluorescence microscopy that binds to the RNA and typically inhibits replication by the replicase enzyme. Researchers found that, while replication was initially inhibited, within a few hours there arose RNA variants that could replicate effectively, despite the presence of acridine orange—and in fact these variants replicated faster with acridine orange present than without. Moreover, these “acridine orange adapted” variants did not evolve in other experiments where acridine orange was not present. Taken together, these observations suggest strong natural selection for the ability to replicate in the particular chemical environment that the RNA molecules experienced.

In essence, the results from the studies by Spiegelman and by Sumper show how the genetic building blocks of life may have evolved. If small variations in RNA sequences exist, if the sequences can replicate themselves, and if there are fitness differences between these sequences, natural selection will lead to evolutionary change. But as we mentioned earlier, these experiments all required that the replicase enzyme be added by the experimenters at the start of each incubation period. RNA plus replicase is sufficient to get natural selection going—but from where would the replicase have come? How could RNA ever have evolved in the first place in the absence of enzymes to facilitate its replication?

One possible solution to this problem was suggested in the early 1980s when Thomas Cech and Sidney Altman independently discovered that enzymes need not be proteins. Rather, RNA itself can act as an enzyme (Kruger et al. 1982; Guerrier-Takada et al. 1983). Such RNA enzymes are called ribozymes. Although ribozymes are much less stable and much less efficient than protein enzymes, a number of different ribozymes have been documented (Lilley
2003; Orgel 2004).

Scientists have hypothesized that, in the RNA world, RNA molecules would have replicated by using ribozymes. But because ribozymes do not replicate themselves in the modern world, nor have they done so for billions of years, evolutionary biologists have created laboratory conditions under which reproduction could occur in order to examine whether RNA molecules are capable of using ribozymes for independent reproduction (Issac and Chmielewski 2002; Voytek and Joyce 2007).

In an ingenious experiment, Natasha Paul and Gerald Joyce demonstrated that RNA can catalyze reactions involved in its own assembly, using what is called the R3C ribozyme (Paul and Joyce 2002). R3C ligates—that is, joins together—two RNA molecules, let’s call them RNA A and RNA B, by forming a bond between them. Using genetic engineering techniques, Paul and Joyce redesigned R3C so that the product formed when RNA A is joined to RNA B by R3C is identical to the R3C itself. In ligating RNA A and RNA B, R3C operates both as a template—positioning A and B in relation to one another—and as an enzyme, catalyzing the chemical reaction that joins them. From simpler precursors RNA A and RNA B, R3C creates a copy of itself; these copies could in turn create copies of themselves. But Paul and Joyce found that in practice their system could only undergo about two rounds of replication before further copying reactions failed. This is because the RNA A and RNA B were producing AB compounds instead of binding to the R3C ribozyme. But this is in no way a fundamental limitation of these types of self-replicating systems; in a follow-up experiment, Lincoln and Joyce (2009) were able to design a similar system that could replicate indefinitely so long as substrate was present.

Whereas the Paul and Joyce experiment demonstrated that RNA species can self-replicate, that system was not capable of evolution by natural selection because it lacked variation for natural selection to act on. To construct a system capable of evolution by natural selection, researchers needed to design a system that featured variation that affects something about successful replication in the RNA molecules present. Moreover, these self-replicating molecules needed to have the property illustrated in Figure 11.5C: a variant template T′ needs to catalyze synthesis of other variant T′ molecules, rather than copies of the original template T.

In 2009, Lincoln and Joyce found a way to do this. They constructed a system much like the one from the 2002 experiment, albeit now with two different templates and four different substrates. After establishing that the template RNAs in this system could self-replicate, they conducted an experiment in which several variant template molecules were supplied with variant substrate molecules. In this system, new variant forms of the templates arose via various mutations. This variation was heritable, and different templates replicated at different rates. The conditions for natural selection were now met in their system. Lincoln and Joyce found that those self-replicating ribozymes that had more efficient catalytic activity and the ability to grow quickly soon began to dominate their populations of self-replicating ribozymes (Lincoln and Joyce 2009).