RNA World (Molecular Biology)

The "RNA World" is a term coined in 1986 by Walter Gilbert to describe the hypothesis that the origin of life proceeded through a stage that did not require DNA (1). The evidence from modern biochemistry that RNA is the more ancient nucleic acid obtains from many different inferences, including: 1) the 2 ‘ -deoxyribose nucleotides are synthesized enzymatically from the corresponding ribose precursors in modern biochemistry, 2) the DNA base thymine is formed by thymidylate synthase from the corresponding uracil nucleotide, and 3) RNA oligonucleotides function as primers in DNA biosynthesis. The molecular biology of DNA points repeatedly to the requirement for preexisting RNA components, suggesting that the evolutionary origin of DNA depended on an ancient world in which RNA was the library of genetic information storage. In this view, DNA arose as a more stable vector by virtue of its superior resistance to self-cleavage, associated with the 2 ‘ -hydroxyl group, and the ability of thymine to circumvent the deleterious thermal and photochemical addition reactions typical of uracil.

Beginning with the self-splicing of introns, the types of reactions promoted by ribozymes has expanded with the help of in vitro selection techniques to include various ligation processes that mimic the action of RNA polymerases (2). An important step toward the emergence of a true "replicase" has come with the creation of a ribozyme that catalyzes elongation of a primer sequence by 11 to 14 nucleotides, with fidelities from 92.0% (for the incorporation of adenosine) to 99.96% (for guanosine) in a mixture of nucleoside triphosphates (3). Not only can ribozymes promote the formation of peptide bonds (4), but the recently unveiled X-ray structure of the ribosome suggests that an adenine base within an RNA chain (instead of an amino acid in a protein) plays the primary catalytic role in the active site: thus, the ribosome is itself a ribozyme (5).

1. Oligomerization of Phosphorimidazolides

The origin of the RNA world has posed a formidable challenge. Self-replication must have preceded the evolution of ribozymes, and, thus, Leslie Orgel and his coworkers (6) have extensively studied the non-enzymatic template-directed oligomerization of activated nucleotides (eg, Ia in Fig. 1). Modern biochemistry employs nucleoside triphosphates, in which pyrophosphate serves as the leaving group, but these compounds react too slowly in laboratory simulations. A seminal discovery in 1980 was the zinc-catalyzed formation of oligoguanylates up to 40 nucleotides in length using a polycytidylate template; moreover, the internucleotide linkage in the presence of this metal ion was predominantly the "natural" 3′ -5′ bond (7). Activated precursors containing 2-methylimidazole (Ib in Fig. 1) gave even better results without the need for added metal ions; subsequent work in Orgel’s group established the fidelity of template copying at better than 99% for the Watson-Crick base when present in a mixture with the other three nucleotide phosphorimidazolides (8). While the precursors were chosen for their chemical efficacy rather than their prebiotic relevance, the results established that template-directed replication of a polynucleotide did not require the presence of a complex protein polymerase.

Figure 1. Activated guanine nucleosides and analogs: (a) guanosine 5′ -phosphoimidazole (R = H) and guanosine 5′ – phospho-2-methylimidazole (R = CH3); (b) 9-[(1-hydroxy-3-phosphoimidazole-2-propoxy)methyl]guanine; (c) 9- [(1,3-diphosphoimidazole-2-propoxy)methyl]guanine; (d) PNA guanine dimer.

Despite the aesthetic appeal of this model system, Orgel and his associates have highlighted serious limits to the templates and activated nucleotides that can be used. 2-Methylphosphorimidazolides prepared from a mixture of D- and L-guanosine do not oligomerize efficiently on a polycytidylate template due to chain termination by the "wrong" stereoisomer (9). Polyuridylates cause problems because of their tendency to form triple helices with adenine derivatives, while polyguanylates assemble into tetraplexes that inhibit replication (10). Mixed templates of guanosine and cytosine work as long as the amount of cytosine is high, but research on the deoxyribonucleotide series has shown that thymidine (and, by implication, uridine) effectively blocks the reaction (11). Because the goal of any self-replicating RNA model requires that the product also serve as a template, the problems of thymine (and uracil) arise whenever adenine is present in the strand being replicated. If these results with DNA analogs apply to their RNA counterparts, then the implications for the RNA world could be: 1) a smaller set of bases was employed early in the origins of replication (with severe limits on the amount of information storage); 2) a different suite of bases was used (with the attendant difficulties for the requisite transition); or 3) a natural catalyst (as yet unidentified) was necessary to achieve the conformation for thymine-containing templates to be replicated. Current research on the second and third alternatives may provide a solution to this dilemma.

Clay minerals on the early Earth could have played a role in the origin of the first oligonucleotide templates. In a remarkable series of papers by James P. Ferris and his colleagues (12), montmorillonite has been shown to promote the oligomerization of nucleoside phosphorimidazolide monomers: oligonucleotides up to 11 U long with a predominance of 3′ -5′ linkages have been obtained from the activated adenosine derivative. The chain length can be extended further (up to the 50-mer) by repeated feeding of the oligoadenylates with additional adenosine phosphorimidazolide (13). The pyrophosphate-linked diadenosine (AppA) appears to be an intermediate in the reaction, and its presence as a starting material augments the regioselectivity of the oligoadenylate product. Montmorillonite also promotes the self-condensation of uridine phosphorimidazoles (14) and cytidine phosphorimidazolides (15), but the linkages are mainly of the non-biological (2 ‘ to 5′ ) type. An intriguing observation is that a mineral-bound template of exclusively 2 ‘ ,5′ -oligocytidylate is a catalyst for the reaction of activated guanosine nucleotides (15). These results demonstrate that naturally occurring minerals cannot only accelerate such reactions but, in some cases, control the orientation of the monomers such that the products contain the biochemically important 3′ -5′ internucleotide bonds.

2. Pre-RNA Alternative Worlds

A major question in the study of the origins of life is what might have preceded the RNA world, and much attention has centered on simpler monomers that possess fewer stereocenters (or even none). The glycerol derivative (II) shown in Figure 1 is attractive because glycerol is a product of prebiotic simulations, and the proposed nucleotide analog has only one stereogenic carbon, but the activated phosphate undergoes a facile intramolecular cyclization that prevents template-directed polymerization (16). A related structural analog, the 1,3-diphosphoimidazolide (III), does react catalytically on a polycytidylate template to give pyrophosphate-linked oligomers up to the 20-mer (17).

Other models for the pre-RNA world have sought to avoid the problems associated with phosphate (low reactivity and dilute natural concentrations) through peptide-linked nucleic acids (PNAs). Although the specific monomer units employed by Nielsen and his colleagues (18) are not asserted to be prebiotic compounds, he and Orgel have demonstrated that the guanosine PNA dimer (IV) undergoes oligomerization up to the decamer on a decadeoxycytidine template. Interestingly, the reaction of the phosphoimidazolide (but not 2-methylimidazolide) of guanosine is promoted by a complementary cytosine-containing analog, thus showing that information transfer between peptide nucleic acid and RNA (in either direction) is feasible. However, studies of autocatalysis in oligopeptide ligation test the assumption that chains of amino acids by themselves cannot replicate (19, 20).

In conclusion, RNA oligonucleotides up to 50 base units long can be synthesized on clay mineral surfaces, and this size lies within the lower regime in which catalytic activity becomes feasible. Self-replication of polynucleotides is possible without the need for enzymes, but only cytidine-rich templates can be efficiently copied. The challenge is to find pathways to this RNA world, either with the discovery of new catalysts or through a genetic takeover from a nucleic acid analog that is truly "prebiotic."