The Skaggs Institute
for Chemical Biology

Scientific Report 2006

Intracellular RNA Folding and Catalysis

M.J. Fedor, J.W. Cottrell, C.P. Da Costa, S.B. Daudenaurde, J.W. Harger, E.M. Mahen, M.R. Saha

RNAs play key roles in normal growth and development, providing potential therapeutic targets and reagents. RNAs are transcribed as linear polymers of 4 nucleotides, which must adopt unique 3-dimensional structures in order to carry out their biological functions. We focus on the pathways through which RNAs form specific functional structures even though they can form many alternative nonfunctional structures that have similar thermodynamic stabilities. Understanding how RNA assembly pathways produce unique, functional outcomes will help explain the pathologic changes associated with RNA misfolding and will guide the development of therapeutic approaches for disrupting the RNA structures needed for proliferation of viral and microbial pathogens.

We devised a way to evaluate RNA assembly during transcription in simple solutions and in living cells by using catalytic RNA self-cleavage rates to quantify partitioning between functional and nonfunctional RNA structures. We began by engineering ribozyme variants with upstream and downstream inserts that potentially can form complementary base pairs with either upstream or downstream parts of the ribozyme sequence (Fig. 1). Complementary base pairing between an insert and the ribozyme sequence creates an AltH1 stem loop that is incompatible with assembly of the H1 stem loop that is required for ribozyme function. We found that complementary inserts located upstream of the ribozyme inhibited ribozyme assembly more than did downstream inserts during transcription in vitro, consistent with a sequential folding model in which the outcome of assembly is determined by the structure that forms first. In contrast, both upstream and downstream inserts strongly inhibited assembly of the same ribozyme variants when the variants were expressed as chimeric mRNAs in yeast.

Fig. 1. Models for intracellular RNA folding. A, In a rapid exchange model, secondary structures that form during transcription can interconvert at a rate that might be accelerated through the action of RNA chaperone proteins (lavender spheres). Structures with the greatest thermodynamic stability are expected to dominate the folding outcome. B, In a delayed folding model, secondary structure assembly does not occur during RNA synthesis, perhaps because nascent transcripts associate with RNA binding proteins (turquoise ellipses). Stable structures might form only after an entire folding domain has been released from bound proteins. In a delayed folding model, stable structures that form most rapidly are expected to dominate the folding outcome.

This evidence that an inhibitory stem loop can form downstream, after an entire ribozyme has been transcribed, indicates that some element that is missing from simple in vitro assembly reactions can contribute to RNA-folding reactions in living cells. Two models might account for this striking result (Fig. 1).

In a “rapid exchange” model, alternative structures interconvert rapidly, allowing assembly to reach an equilibrium in which the most thermodynamically stable structure predominates. In a “delayed folding” model, on the other hand, assembly does not occur during transcription, perhaps because proteins that associate with nascent transcripts prevent intramolecular base pairing. Once an entire domain has been released from protein, perhaps concomitant with export of RNA from the nucleus to the cytoplasm, partitioning among alternative structures might be governed by differences in assembly rates.

In the first set of variants we examined, AltH1 stem loops assemble from contiguous sequences while the H1 stem loops assemble from separate ends of the RNA so that nonfunctional AltH1 stem loops might be expected to assemble faster than H1 stem loops do. If so, the ability of AltH1 stem loops to dominate the folding pathway could reflect the importance of fast folding kinetics, consistent with the delayed folding model. To learn whether assembly kinetics affect partitioning between alternative structures, we next examined circularly permuted ribozymes in which H1 stem loops assemble from contiguous strands (Fig. 2). In these circular permutants, H1 is expected to assemble faster than AltH1 stem loops do that form between noncontiguous sequences.

Fig. 2. Circular permutants. Functional structures that include the essential H1 helix of the ribozyme predominate over nonfunctional structures that include AltH1 stem loops when H1 forms from contiguous sequences even when AltH1 has greater thermodynamic stability than H1 does.

Chimeric mRNAs containing circularly permuted ribozymes were less abundant in yeast and had the accelerated decay indicative of efficient intracellular self-cleavage of functional ribozyme structures. The high self-cleavage activity of circular permutants suggests that contiguous H1 stem loops are better able to resist competition from noncontiguous AltH1 stem loops, even when upstream and downstream AltH1 stem loops have high thermodynamic stability. Thus, contiguity, which would accelerate stem loop assembly, evidently contributes to folding outcomes, consistent with the delayed folding model of secondary structure assembly.