Prebiotic RNA unstuck

Non-enzymatic copying of an RNA template is appealing as a transition from pre-life to an RNA world, but it has been difficult to demonstrate in the laboratory. Now, two separate studies focusing on RNA's backbone connectivity offer partial solutions to some of the problems raised with this hypothesis for the origin of life.

The 'RNA world' hypothesis describes an early stage of life on Earth, prior to the evolution of coded protein synthesis, in which RNA served as both information carrier and functional catalyst^1,2. Not only is there a significant body of evidence to support this hypothesis³, but the 'ribo-organisms' from this RNA world are likely to have achieved a significant degree of metabolic sophistication⁴. From the perspective of the origins of life, the path from pre-life chemistry to the RNA world probably included cycles of template-directed RNA replication, with the RNA templates assembled from prebiotically generated ribonucleotides (Fig. 1)⁵. RNA seems well suited for the task of replication because its components pair in a complementary fashion. One strand of nucleic acid could thereby serve as a template to direct the polymerization of its complementary strand.

Figure 1: The path from prebiotic chemistry to the RNA world is likely to have involved template-directed RNA replication.

Activated RNA monomers would have been synthesized de novo from prebiotic organic and inorganic chemicals on primordial Earth (1). Random oligomerization of activated monomers would lead to the formation of functional and non-functional (random) oligonucleotides (2), which could serve as templates (3) that direct the copying of the complementary strand (4). Strand separation must occur for the newly copied RNA transcripts (5) to progress to a new round of replication (6). Governed by the forces of molecular selection, RNAs with increasing structural and functional sophistication would appear over time (7), leading to the emergence of the first protocells and the dawn of the RNA world (8).

Nevertheless, even given abundant ribonucleotides and prebiotically generated RNA templates, significant problems are believed to stand in the way of an experimental demonstration of multiple cycles of RNA replication. For example, non-enzymatic RNA template-copying reactions generate complementary strands that contain 2′,5′-phosphodiester linkages randomly distributed amongst the 3′,5′-linkages (Fig. 2a)⁶ rather than the solely 3′,5′-linkages found in contemporary biology. This heterogeneity has been generally presumed to preclude the evolution of heritable RNA with functional properties. A second problem with the RNA replication cycle concerns the high 'melting temperatures' required to separate the strands of the RNA duplexes formed by template copying⁷. For example, 14mer RNA duplexes with a high proportion of guanine–cytosine pairs can have melting temperatures well above 90 °C. Such stability would prohibit strand separation and thereby halt progression to the next generation of replication. Yet another difficulty results from the hydrolysis and cyclization reactions of chemically activated mononucleotide and oligonucleotide substrates⁷, which would deactivate them for template copying. Together, these and other issues have precluded the demonstration of chemically driven RNA replication in the laboratory. Now, two new studies reported in Nature Chemistry — one from the Szostak laboratory and the other from the Sutherland group — offer potential solutions to these problems^8,9.

Figure 2: RNA backbone linkages and substrate activation chemistry.

a, Szostak and co-workers have shown that functional RNAs having a mixture of 2′,5′- and 3′,5′-phosphodiester linkages could serve as ribozymes and aptamers⁸. b, Sutherland and co-workers have demonstrated that reacting nucleoside-2′- and 3′-phosphates with plausible prebiotic acetylating reagents, in the presence of free nucleotides (Nuc), biases the formation of 3′,5′-phosphodiester bonds⁹. c, Ligation of an RNA strand to a primer-template duplex under plausible prebiotic conditions.

Szostak and co-workers have challenged the assumption that backbone homogeneity was a requirement for the primordial RNA replication process, and considered whether RNAs that contain significant levels of 2′,5′-linkages can be tolerated within known ribozymes and aptamers⁸. They synthesized two well-known functional RNAs — the flavin mononucleotide (FMN) aptamer and hammerhead ribozyme — containing varying amounts (10–50%) of randomly distributed 2′,5′-linkages. Overall, FMN aptamers and hammerhead ribozymes possessing high levels of 2′,5′-linkages (50%) performed about two orders of magnitude worse than their homogeneous 3′,5′-linked counterparts. In contrast, more modest, yet significant levels of 2′,5′-linkages (∼10%) reduced aptamer binding affinity to FMN by only 10-fold and decreased hammerhead ribozyme activity by only 20%. Taken together, these findings demonstrate that RNAs with considerable backbone heterogeneity can retain functional properties. Moreover, RNAs evolved to function in the context of such heterogeneity could possibly tolerate even higher levels of 2′,5′-linkages.

Szostak and colleagues go on to argue that not only can RNAs tolerate backbone heterogeneity, but that such heterogeneity would confer a functional advantage in the primordial RNA replication process by lowering the melting temperature of RNA duplexes. A lower melting temperature would enable continued replication cycles through thermal strand separation. Supporting this viewpoint, the study showed that 13% of 2′,5′-phosphodiester linkages reduced the melting temperature of a 30mer RNA duplex by more than 16 °C compared with its homogenous 3′,5′-linked counterpart.

In the work described by Sutherland and co-workers, the issue of backbone heterogeneity is tackled head-on⁹. They searched for plausible prebiotic conditions that would have allowed for the formation of activated RNA monomers leading to predominantly 3′,5′-linked oligoribonucleotides. The reactivity of adenosine-3′-phosphate was explored in the presence of various potential electrophiles, co-factors and reaction conditions. It was found that when thioacetate and cyanoacetylene were present, adenosine-3′-phosphate produced a 3′,5′-linked 2′-O-acetylated dimer, indicating both protection of the 2′-hydroxyl group and activation of the 3′-phosphate for formation of a dinucleotide. Presumably this reaction chemistry proceeds by acetylation of the 3′-phosphate monoester to form a mixed carboxy–phosphate anhydride (Fig. 2b). The adjacent 2′-hydroxyl group then attacks the carbonyl carbon atom to mediate acyl transfer. A second acylation of the 3′-phosphate group activates the protected nucleotide for intermolecular attack by the 5′-hydroxyl group of another monomer to form a dimer. In contrast, nucleoside-2′-phosphates undergo 3′-hydroxyl acetylation much less efficiently and become deactivated for oligomerization by undergoing 2′,3′-cyclization. After formation of the mixed anhydride, it seems that the 3′-hydroxyl group preferentially attacks the phosphorus rather than the carbon centre, displacing the carboxylic acid and forming the 2′,3′-cyclic phosphate.

The remarkable chemistry and selectivity observed in the context of mononucleotides extends to oligoribonucleotide-3′-phosphates and 2′-phosphates. Chemoselective acetylation and efficient template-directed ligation occur for the former, but not the latter, thereby strongly favouring ligation products with 3′,5′-linkages (Fig. 2c). Moreover, subsequent deprotection conditions for removal of the 2′-O-acetyl group also result in a further enrichment of 3′,5′-linked RNA by causing partial hydrolysis of 2′,5′-linkages while leaving 3′,5′-linkages intact. These new results also provide a chemical pathway that could continuously reactivate hydrolysed substrates, which ordinarily limit the extent of template copying that can be achieved.

Together, the two studies make it a little easier to visualize the emergence of the RNA world from prebiotic chemistry. Instead of RNAs getting 'stuck' by succumbing to irreversible deactivation chemistry or by engaging in replication cycles that produce non-heritable traits and dead-end complexes, backbone heterogeneity coupled with acetylation could have sustained template-driven RNA replication cycles. Subject to the forces of molecular selection, the replication cycles would gradually produce a more sophisticated replication machinery, including ribo-catalysts for strand separation that could enable the eventual transition to homogenous 3′,5′-linked genetic material.