Evolution of RNA Folding Kinetics

The preimages of RNA secondary structures in sequence space are represented by graphs called neutral networks. Common structures form extended neutral networks which span entire sequence space. Pairs of networks approach each other closely at intersection points of the compatibele sets in which they are embedded. Eight years ago the existence of neitral networks and their intersection points was predicted by a mathematical analysis of sequence-structure relations based on random graph theory and computer simulations. This year it was also proven to be true by an elegant experiment designed to evolve an RNA molecule with two different catalytic activities (Schulten & Bartel, Science 289: 448). RNA secondary structures are coarse grained versions of full structures and need to be refined in two complementary aspects: (i) Consideration of minimal free energy (mfe) structures as RNA phenotypes has to be extended to ensembles including suboptimal conformation and kinetically determined structures (In thermodynamic terms such an extension corresponds to a replacement of the "0 K and infinite time" scenario by more elaborate and realistic notions of structure that account for non-zero temperatures and finite folding times), and (ii) incorporation of classifiable tertiary contacts into secondary structures in shows punctuation caused by alternation of short adaptive phases and long epochs of phenotypical stasis which are determined by the topology of neutral networks.

The current proposal suggests to explore the distribution of refined properties, like accessibility of suboptimal stuctures, folding efficiences, and folding times, on neutral networks formed by sequences folding into identical mfe-secondary structures. We expect the appearence of classes, which allow to group molecules according to different folding characteristics, and we shall try to prove that molecules with given refined properties are not confined to certain areas, but are distributed (almost uniformly) over large parts of sequence space. The validity of this conjecture has direct consequences for the evolutionary design of biomolecules since it allows topevelop new search strategies. The mapping of RNA sequences into kinetically controlled structures will be investigated by extensive computer simulations and by a mathematical model in the spirit of the random graph approach to neutral networks of mfe-structures.

Already existing algorithms conceived and implemented by our group (Vienna RNA Package) and new developed techniques will be applied to a study of kinetically controlled sequence-structure relations of RNA molecules. Our algorithm for kinetic folding RNA is not restricted to confentional secondary structures. Versions exist, which account already for H-type pseudoknots and co-axial stacking. Provided enough empirical data are available, the algorithm can be applied to almost all other kinds of classifiable tertiary interactions (B-type pseudoknots, base triplets and others). The ultimate goal is to include a sufficiently large number of tertiary interactions in order to come close to well defined 3D structures.