Report

Kinetic Intermediates Trapped by Native Interactions in RNA Folding

See allHide authors and affiliations

Science  20 Mar 1998:
Vol. 279, Issue 5358, pp. 1943-1946
DOI: 10.1126/science.279.5358.1943

Abstract

In the magnesium ion–dependent folding of theTetrahymena ribozyme, a kinetic intermediate accumulates in which the P4-P6 domain is formed, but the P3-P7 domain is not. The kinetic barriers to P3-P7 formation were investigated with the use of in vitro selection to identify mutant RNA molecules in which the folding rate of the P3-P7 domain was increased. The critical mutations disrupt native tertiary interactions within the P4-P6 domain and increase the rate of P3-P7 formation by destabilizing a kinetically trapped intermediate. Hence, kinetic traps stabilized by native interactions, and not simply by mispaired nonnative structures, can present a substantial barrier to RNA folding.

RNA forms complex structures that are able to perform a variety of functions ranging from ligand binding to catalysis. However, the mechanism by which an RNA molecule folds into a unique three-dimensional structure remains poorly understood. To study the Mg2+-dependent kinetic folding pathways of large, highly structured RNA molecules such as theTetrahymena ribozyme and ribonuclease (RNase) P, we have previously developed a kinetic oligonucleotide hybridization assay (1, 2). This assay exploits the selective accessibility of unfolded RNAs to sequence-specific oligodeoxynucleotide probes, the binding of which confers sensitivity to cleavage by RNase H. Folding is initiated by the addition of Mg2+, and the fraction of unfolded RNA at various times is scored in a cleavage reaction containing DNA probes and RNase H. On addition of Mg2+ to the Tetrahymena ribozyme, the two structural domains that constitute the catalytic core—P4-P6 [base-paired (P) regions 4 to 6, positions 104 to 261] and P3-P7 (P3, P7, and P8)—form sequentially as kinetic folding units (1, 3). Formation of P4-P6 is rapid (60 min−1) (4), whereas P3-P7 forms slowly, on the minute time scale (1). This order of kinetic folding events is supported by chemical modification (5), ultraviolet cross-linking (6), and x-ray footprint (4) analysis. In the proposed folding pathway (1,3), an intermediate (I2) accumulates in which only P4-P6 is folded, and the rate-limiting step for P3-P7 formation is the unimolecular rearrangement of I2 to intermediate I3. Slow unimolecular folding steps have also been identified for the group I intron b15 (7) and RNase P (2), and they may be a general feature in the folding of large RNAs.

Mutations that increase the rate of folding of proteins have provided insight into the mechanism of slow folding steps (8). We developed an in vitro selection scheme to identify mutantTetrahymena ribozymes in which the slow P3-P7 folding step (I2 → I3) is accelerated (9). Ribozymes that fold rapidly after Mg2+ addition were selected from a pool of RNAs containing an average of four mutations per molecule. Slow-folding RNAs were selectively depleted from the pool by kinetic oligonucleotide hybridization with probes targeting P3 and P7. A step was included in each cycle of selection to ensure that fast folding mutants formed an intact catalytic core (9). After nine rounds, the folding rate of the pool (G9) had increased by a factor of 4 relative to that of the initial pool (G0) and by a factor of 2 relative to that of the wild type (Fig. 1). Twenty-four individual molecules were cloned from the G9 pool, and the folding rate of the P3-P7 domain for five of these clones was at least three to five times that of the wild type at 37°C (Fig. 1 and Table1) (10, 11).

Figure 1

Isolation of fast folding RNAs after nine rounds of in vitro selection. The kinetics of P3-P7 formation for ribozyme generations G0 to G9 and cloned individual molecules from G9 were probed by kinetic oligonucleotide hybridization. Initiation of folding and the quench reaction were as described (9). The fraction cleaved at each folding time was determined by denaturing PAGE and Phosphorimager analysis (Molecular Dynamics). The apparent folding rate constant (k fold) was calculated by fitting curves to a single exponential (10). Data were normalized to allow direct comparison. RNAs and k fold values: □, G0 pool (0.63 min−1), ▪, wild type (1.2 min−1), ○, G9 pool (2.33 min−1), and •, clone G9-10 (5.0 min−1).

Table 1

Kinetic and thermodynamic constants for the folding and activity of fast folding mutants. P3-P7 folding kinetics for the clones and corresponding point mutants (second and fourth columns, respectively) were measured as in Fig. 1. Mutations that did not affect folding are not listed. The precision of thek fold values was generally a factor of ∼1.5. The stability of P3-P7 for each point mutant ([Mg2+]1/2) was measured by Mg2+titration as described (1). RNA molecules were equilibrated for 25 min at 37°C in 0.2 to 10.0 mM MgCl2. At each Mg2+ concentration, the fraction of molecules folded was determined by oligonucleotide hybridization targeting P3. [Mg2+]1/2, the concentration of Mg2+ required for half-maximal folding, was calculated by fitting curves to the Hill equation. For each point mutant, the integrity of the catalytic core was assessed by measuring the rate constant for the chemical step (k c) in a single-turnover cleavage reaction as described (3). Ribozyme (50 to 100 nM) was annealed in TE buffer and allowed to fold for 10 min at 37°C in k c buffer [50 mM MES (pH 5.5), 10 mM MgCl2, 10 mM NaCl, and 1 mM dithiothreitol]. The reaction was initiated by adding 5′32P-labeled substrate (CCCUCUAAAAA) and guanosine triphosphate (final concentrations, 0.5 nM and 0.5 to 2.0 mM, respectively) in k c buffer. The fraction of molecules cleaved at various times was determined by denaturing PAGE, and k c was calculated by fitting curves to a single exponential. k c values were constant (±0.02 min−1) in the range of ribozyme and guanosine triphosphate concentrations tested, confirming that the conditions were saturating. The fraction of active molecules was similar for wild-type and mutant ribozymes (28).

View this table:

Because each fast folding clone contained at least three mutations, individual point mutants were constructed. For the four clones analyzed, a single mutation was sufficient to reproduce the fast folding phenotype (Table 1). The A183U (A at position 183 → U), A171G, U167C, and +G174 (G insertion at position 174) mutations are all localized in the P5abc region of the P4-P6 domain (Fig.2A), suggesting a common mechanism of action. The mutations did not greatly affect either catalysis or the stability of P3-P7 (Table 1). Thus, although P5abc makes no direct contacts with P3-P7 in the three-dimensional structure model (12) and the P4-P6 domain acquires its native structure before P3-P7 formation (4), positions in P5abc affect the folding rate of P3-P7.

Figure 2

Localization of fast folding mutations in the P5abc region of the P4-P6 domain. Positions of fast folding mutations in the ribozyme secondary structure (A) and the P4-P6 domain crystal structure (B) (13) are indicated.

The P5abc mutations are clustered throughout the “magnesium core” of P4-P6 in the domain crystal structure (Fig. 2B) (13). Single-atom changes that disrupt the Mg2+binding sites destabilize the entire P4-P6 domain (13). Given that all of the fast folding mutations are likely to disrupt the highly cooperative network of interactions within the core, destabilization of P4-P6 in I2 may increase the rate of P3-P7 formation. The potentially destabilizing effects of the A171G and A183U mutations are especially apparent. A171 is within an adenosine platform motif in the loop of P5c (L5c) (13) that may stabilize P5c or mediate pairing between L5c and the P2 loop in the proposed P14 helix (12, 14). Furthermore, the Rpphosphate oxygen of A171 directly coordinates Mg2+(13). The A183U mutation disrupts two hydrogen bonds that directly bridge the helical stacks in P4-P6 (13).

If the fast folding mutations destabilize P4-P6, then other destabilizing mutations such as A186U (15, 16) should also accelerate folding. A186U disrupts five hydrogen bonds with three different nucleotides in the three-way junction (13) (Fig.2). This mutation accelerates P3-P7 formation by the same factor as that observed for the selected mutations (Table 1), suggesting that destabilization of P4-P6 in I2 is the common mechanism of action (17).

Given that destabilization of native P4-P6 interactions increases the rate of P3-P7 formation, we propose that I2 is a kinetic trap. In protein and RNA folding, the hallmark of a kinetic trap is that the folding rate is increased in the presence of a denaturant. Although denaturants typically reduce the rate of protein folding, they can also increase the rate of protein (18) and RNA (19) folding by destabilizing kinetically trapped intermediates. Urea markedly increased the rate of P3-P7 formation in the wild-type ribozyme (Fig. 3) but had a much smaller effect on the folding rate of the A183U mutant. Hence, a kinetic trap that is present in the folding of the wild-type ribozyme is diminished by the A183U mutation (20).

Figure 3

Destabilization of a kinetic trap by fast folding mutations. The kinetics of P3 formation for wild-type (•) and A183U (○) ribozymes were measured in the presence of various concentrations of urea. RNAs were folded as in Fig. 1, with the exception that 2× folding buffer also contained urea. Folding was quenched with RNase H and oligonucleotides targeting P3 (positions 270 to 279).

Kinetic traps in both RNA and protein folding are often misfolded structures that slow folding because stable nonnative interactions must be disrupted to achieve the transition state. Incorrect heme coordination in cytochrome c (21) and mispairing in tRNA (22) are examples of nonnative interactions that must be disrupted for folding to proceed. However, in the folding of bovine pancreatic trypsin inhibitor, a trapped intermediate is stabilized solely by native interactions (18). Furthermore, in the major folding pathway of lysozyme, an intermediate with a folded α domain slows folding of the β domain (23). It has been suggested that the lysozyme intermediate is a kinetic trap in which the stability of a native α domain renders the polypeptide rigid and “aggravates” the conformational search of the β domain (24). We propose a similar mechanism to explain the slow folding of P3-P7.

The kinetic trap (I2) in ribozyme folding exhibits both native (P4-P6) and nonnative (P3-P7) structures and may also exhibit nonnative interactions at the domain interface (6). Each of these structural features might restrict the conformational flexibility of I2 and slow P3-P7 formation. However, because mutations that destabilize P4-P6 diminish the kinetic trap, the contribution of the nonnative structures to the stability of the trap is either minimal or strictly dependent on the presence of a stable P4-P6 domain. I2 is thus a native kinetic trap because its stability is derived primarily from native interactions.

The proposed native kinetic trap differs fundamentally from the canonical mispairing traps observed in tRNA (22) and ribozymes (25) and may define a new class of barriers in the folding of multidomain RNAs. Furthermore, our data show that stable intermediates are prevalent but nonessential features of RNA folding and that destabilization of an intermediate accelerates folding by facilitating escape from a native kinetic trap. These conclusions support theoretical folding models and emphasize the parallels between protein and RNA folding (3, 26).

  • * Present address: Center for Cellular and Genetic Therapies, Department of Surgery, Duke University Medical Center, Durham, NC 27710, USA.

REFERENCES AND NOTES

View Abstract

Navigate This Article