Directed Evolutionary Descriptions of Natural RNA Structures

Abstract

Selection experiments have frequently been used to optimize or alter the functions of biopolymers and have also proven useful for discerning biopolymer structure. Just as covariations observed in ribosomal RNA phylogeny may pinpoint Watson-Crick base pairs, sequence information recovered from in vitro and in vivo selection experiments can be used to establish and refine the structure of natural and unnatural nucleic acids. For example, RNA molecules selected from partially or completely randomized pools typically fall into a limited number of sequence families; the relationships within and between these families constitute an artificial phylogeny in which relationships are based on function rather than history. Nevertheless, comparative sequence analysis can still be used to identify what portions of primary sequence are essential for function, which residues potentially pair with one another, whether base pairs act as structural stabilizers or contribute directly to function, and even where tertiary structural interactions may occur. Such results can be combined to generate artificial evolutionary models for RNA structure.

Although crystal or nuclear magnetic resonance (NMR) structures of molecules provide much greater spatial detail, evolutionary studies delve into function as well as structure and can provide insights into how structures may change through time. Evolutionary models can complement more physical determinations: For example, a detailed knowledge of the structure of rRNA or HIV-1 reverse transcriptase can reveal why particular mutations impart resistance to drugs such as streptomycin or AZT, but cannot yet be used to predict all mutations that will lead to drug resistance. In contrast, selection...