A versatile communication module for controlling RNA folding and catalysis

Abstract

To exert control over RNA folding and catalysis, both molecular engineering strategies and in vitro selection techniques have been applied toward the development of allosteric ribozymes whose activities are regulated by the binding of specific effector molecules or ligands. We now describe the isolation and characterization of a new and considerably versatile RNA element that functions as a communication module to render disparate RNA folding domains interdependent. In contrast to some existing communication modules, the novel 9-nt RNA element is demonstrated to function similarly between a variety of catalysts that include the hepatitis delta virus, hammerhead, X motif and Tetrahymena group I ribozymes, and various ligand-binding domains. The data support a mechanistic model of RNA folding in which the element is comprised of both canonical and non-canonical base pairs and an unpaired nucleotide in the active, effector-bound conformation. Aside from enabling effector-controlled RNA function through rational design, the element can be utilized to identify sites in large RNAs that are susceptible to effector regulation.

Figure 1

Isolation and characterization of theophylline-dependent HDV ribozymes. (A) Design of initial populations for in vitro selection. The genomic HDV ribozyme was modified to include random-sequence domains of either 9 or 10 total nucleotides (N) and a theophylline-binding domain in place of P4 (P4′). Other modifications to the ribozyme include the replacement of G-C base pairs with A-U base pairs (shaded) and the inclusion of both 5′- and 3′-terminal extensions to enable selection and facilitate manipulation by RT–PCR. Additionally, 3 nt at positions 41–43 of the genomic HDV ribozyme were deleted (ΔCAA; shaded) as their absence or presence has little effect on ribozyme activity. C75 (boxed) is the nucleotide that functions at the active site of the catalyst as a general acid/base. (B) Progress of selection for theophylline-dependent HDV catalysts. The self-cleavage activity of the initial population (G0) and each subsequent population derived by in vitro selection for theophylline-dependent function is shown for reactions in the absence (open bars) or presence (filled bars) of theophylline under selection conditions incubated for 2 min. (C) Self-cleavage activity of an allosteric HDV ribozyme identified from the final population (G7). The inset depicts the precursor (filled arrowhead) and 3′-cleavage product (open arrowhead) of the theophylline-dependent HDV catalyst containing cm⁺theo6 as separated by denaturing PAGE following reaction under selection conditions in the absence or presence of theophylline for various lengths of time. The observed rate constants in the absence (open circles) or presence (filled circles) of theophylline were derived by plotting the natural logarithm of the fraction of uncleaved RNA versus time, and establishing the negative slope of the resulting lines (solid lines). The biphasic kinetic profile (dashed line) exhibited by the catalyst in the presence of theophylline indicates that ∼50% of the molecules might be irreversibly misfolded or only slowly convert to the active conformation.

Figure 2

Mutational analysis of theophylline-dependent HDV ribozymes. (A) Theophylline-dependent HDV ribozymes and mutant constructs. Shown is the communication module sequence and flanking G40·G74 and A5·G29 base pairs (shaded) of the catalyst and theophylline-binding domain, respectively, for each ribozyme containing cm⁺theo6 and cm⁺theo7. The secondary structures depicted are those proposed for the active, ligand-bound conformations of the catalysts. Arrows denote the relationship of each construct to other constructs (arbitrarily numbered 1–8) containing single nucleotide mutations (boxed relative to cm⁺theo6 sequence). (B) Theophylline-dependent activities of allosteric HDV ribozymes and mutant constructs. The activity of each construct is represented as the logarithm of the initial k_obs determined by reaction in the absence (open circle) or presence (filled circle) of theophylline under selection conditions. Numbers in parentheses indicate the fold-activation elicited by effector. Asterisks denote those catalysts and conditions under which the catalysts exhibit biphasic kinetic profiles similar to that shown in Figure 1C.

Figure 3

Versatile control of HDV and hammerhead ribozymes. (A) Allosteric HDV ribozymes. Catalysts are comprised of the modified genomic HDV ribozyme, cm⁺theo6 (shaded), and theophylline-, FMN- or ATP-binding domains (HDV-T, HDV-F and HDV-A, respectively). Depicted is P4′ for each construct. (B) Effector specificities and activities of allosteric HDV ribozymes. Shown is the uncleaved ribozyme (filled arrowheads) and 3′-cleavage product (open arrowheads) separated by denaturing PAGE following reaction in the absence (no ligand) or presence of theophylline (+theo), FMN (+FMN) or ATP (+ATP) at 23°C for 45 min. Unreacted ribozyme is indicated (no rxn). (C) Effector specificities and activities of allosteric hammerhead ribozymes. The hammerhead ribozyme was similarly modified by replacing all but the G10.1·C11.1 base pair of stem II with cm⁺theo6 and theophylline-, FMN- or ATP-binding domains (HH-T, HH-F and HH-A, respectively). Each allosteric hammerhead ribozyme was analyzed identically as described for allosteric HDV ribozymes with the exception that open arrowheads denote the 5′-cleavage product.

Figure 4

Identification of effector-dependent sites in X motif and Tetrahymena group I ribozymes. (A) An allosteric X motif ribozyme. An X motif ribozyme was modified by replacing all but two base pairs of stem IV with cm⁺theo6 and a theophylline-binding domain (X-T). The inset depicts the activity of the allosteric ribozyme. Shown is the uncleaved ribozyme (filled arrowhead) and 5′-cleavage product (open arrowhead) separated by denaturing PAGE following reaction in the absence (no ligand) or presence of theophylline (+theo) at 37°C for 45 min. Unreacted ribozyme is indicated (no rxn). (B) Design of allosteric Tetrahymena group I ribozymes. Group I ribozymes correspond to the L-21 NheI RNA which lacks the P9.1 and P9.2 extensions, and contain CGAAA in place of nucleotides 322–326 in P9. The ribozyme was modified by replacing a portion of P6 or P8 with cm⁺theo6 and a theophylline-binding domain (P6-T or P8-T, respectively). Alternatively, the ribozyme was modified at both positions (P6-T/P8-T). (C) Activities of unmodified and allosteric Tetrahymena group I ribozymes. Shown is ribozyme (shaded arrowhead), substrate (filled arrowhead) and 5′-cleavage product (open arrowhead) resulting from ribozyme catalysis of the reverse of the second step of splicing (3′-exon ligation). Products were separated by denaturing PAGE following reaction in the absence (no ligand) or presence of theophylline (+theo) at 23°C for 30 min. Unmodified ribozyme is labeled P6/P8. Reaction containing substrate alone is indicated (no ribozyme).