Evolution of RNA ecologies and functional networks

Research project P 13887 Evolution of RNA Ecologies Peter SCHUSTER 11.10.1999 The goal of the proposal is to study RNA-hybridization systematically and to use it as a tool to investigate coevolution. Hybridization complexes will be studied first on the level of secondary structures and later on extended by consideration of tertiary interactions. The systematic study will be based on the concept of understanding the cofolding of two RNA molecules as a mapping from the combined sequence space of both molecules into a shape space of RNA hybrids. As previously done with single RNA molecules the investigations will be performed at three levels: (i) folding all sequences of short chain length and exhaustive enumeration of structures, (ii) statistics of structures derived from intermediate and long chains, and (iii) development of a mathematical model for the mapping based on random graph theory. Coevolution will be studied by means of a flowreactor model simulating an assay of functionally interacting RNA molecules. The computer implementation of the flowreactor model has been used already to study evolution of independently replicating RNA molecules and optimization of RNA folding. The simplest application will be template-induced and catalyzed RNA synthesis. In this case we plan also to perform full qualitative analysis of the underlying kinetic differential equations. The model for functionally coupled organizations of ecologies of RNA molecules will be studied by computer simulation. Catalytic functions are assumed to be properties of model ribozymes. A model ribozyme has to meet both, structural conditions and sequence requirements. The catalytic activities of RNA such as template action, cleavage, ligation, and assistence in replication will be mediated through structural criteria applied to the hybridization complexes of the two cofolded molecules. Computer simulation will be complemented by a model based on a kind of metadynamics which combines stochastic and deterministic aspects of coevolution.

Evolutionary optimization shapes nature and it is of fundamental importance in the design of molecules by means of trial and error techniques. RNA molecules play a special role because they can be replicated and evolved in laboratory systems and, at the same time, can act as catalysts in biological processes. An essential feature of natural evolution based on variation and selection is a dichotomy between genotypes and phenotypes. The genotype, or the sequence of the carrier of genetic information, is varied by mutation and recombination whereas selection operates exclusively through evaluation of the properties of the phenotype, which is a molecular structure, a cell or an entire multi-cellular organism. The relation between genotypes and phenotypes is of fundamental importance for evolution, therefore. In case of RNA evolution this relation boils down to a relation between sequences and molecular structures. Research on the project was targeted to the fundamental question of nearness in evolution: Given a structure of an RNA molecule, how many structures can be reached by mutation? What do these neighboring structures look like, and which are their functions? The answer we are able to give has two issues: (i) The relation of sequences to structures is many-to-one in the sense that many sequences form the same structure. `Neutrality` with respect to sequences as found with RNA molecules is basic to biological evolution in general. Not a single sequence corresponds to a structure but a whole set of sequences which can be connected to a graph called `neutral network`. The success of evolution depends critically on the existence of such networks. (ii) The question of nearness between two structures turns out to be closely related to the problem to define nearness between two sets or two graphs. We could show by the research work done within the frame of the project that nearness of structures, or phenotypes in general, cannot be measured by an ordinary distance but it is of weak topological nature. If structure A is near structure B, the inverse need not be true, B need not be near A. In other word it may be easy to mutate A to yield B, but it is unlikely to yield A as the result of a mutation on B. The predictions from our model have direct consequences for optimal protocols in the design of molecules by evolution. Two of our predictions have been verified already by suitable experiments in evolutionary biotechnology. The knowledge of sequence-structure relations of RNA was applied to studies on `molecular ecologies`. Two examples were conceived and investigated by means of mathematical techniques and computer simulation: (i) A genetic switch involving two genes which has two different states of gene expression, and (ii) a prebiotic model system, which is now experimentally investigated within an EU-project aiming at de novo design and chemical synthesis of a simple cell.