J. Mol. Biol. (1995) 249, 59–68 Pseudoknot in Domain II of 23 S rRNA is Essential for Ribosome Function Gunnar Rosendahl, Lykke Haastrup Hansen and Stephen Douthwaite* Department of Molecular The structure of domain II in all 23 S (and 23 S-like) rRNAs is constrained Biology, Odense University by a pseudoknot formed between nucleotides 1005 and 1138, and between DK-5230 Odense M 1006 and 1137 (Escherichia coli numbering). These nucletoides are exclusively Denmark conserved as 1005C·1138G and 1006C·1137G pairs in all Bacteria, Archaea and chloroplasts, whereas 1005G·1138C and 1006U·1137A pairs occur in Eukarya. We have mutagenised nucleotides 1005CG, 1006CU, 1137GA and 1138GC, both individually and in combinations, in a 23 S rRNA gene from the bacterium E. coli . The ability of 23 S rRNA to support cell growth is reduced when either of these base-pairs is disrupted, and it is completely abolished upon disruption of both base-pairs. Each mutant 23 S rRNA is assembled into 50 S subunits, but the mutant subunits do not stably interact with 30 S to engage in protein synthesis. Enzymatic and chemical probing of ribosomal particles reveals increased accessibility in the rRNA structure close to the sites of the mutations. The degree to which the mutations increase rRNA accessibility correlates with the severity of their phenotypic effects. Nucleotide 1131G is extremely reactive to dimethyl sulphate modification in wild-type subunits and ribosomes, but is rendered unreactive when either the pseudoknot is broken or when the r-proteins are removed. The structure of the pseudoknot region is possibly influenced by interaction of an r-protein at or close to the pseudoknot. Re-establishing the pseudoknot Watson-Crick interactions with one ‘‘eukaryal’’ (1005G·1138C or 1006U·1137A) pair and one ‘‘bacterial’’ C·G pair largely restores the structure and function of the rRNA. Bacterial ribosomes containing both these eukaryal pairs also participate in protein synthesis, although at much reduced efficiency, and the structure of their pseudoknot region is partially open and accessible. Keywords: RNA tertiary structure; protein-RNA interactions; rRNA *Corresponding author function; comparative sequence analysis Introduction One of the tenets of molecular biology is that the function of a biologically active macromolecule is defined by its structural conformation. The confor- mation of an RNA molecule is primarily determined by specific interactions between its nucleotides. Elucidation of the nucleotide interactions that determine the folding of rRNAs is therefore progressing hand in hand with an increased understanding of how these molecules engage in protein synthesis. The secondary structures of rRNAs were initially revealed by comparative sequence analysis (Noller, 1984; Woese et al ., 1983), and these structures were subsequently confirmed by various biochemical and molecular genetic methods (reviewed by Dahlberg, 1989; Egebjerg et al ., 1990b; Noller, 1991). Comparative analysis of rRNA sequences has additionally uncovered details of the tertiary folding of these molecules. These include numerous nucleotide interactions within and be- tween the six domains of 23 S rRNAs (Gutell et al ., 1994). Such tertiary interactions receive support from biochemical probing (Egebjerg et al ., 1990b) and cross-linking data (Do ¨ring et al ., 1991). The inter- actions would serve to compact the rRNA structure, bringing functionally important regions of different domains into close proximity. Domain II of 23 S rRNA is associated with the steps of translation that involve hydrolysis of GTP (reviewed by Cundliffe, 1990). The structural con- formation of domain II determines the orientation and function of the GTPase region, as well as orchestrating links with other domains of the rRNA. Comparative sequence analysis has revealed a pseudoknot structure within domain II that would 0022–2836/95/210059–10 $08.00/0  1995 Academic Press Limited