Dissociative Transition State in Hepatitis Delta Virus Ribozyme Catalysis

Abstract

Ribonucleases and small nucleolytic ribozymes are both able to catalyze RNA strand cleavage through 2'-O-transphosphorylation, provoking the question of whether protein and RNA enzymes facilitate mechanisms that pass through the same or distinct transition states. Here, we report the primary and secondary ¹⁸O kinetic isotope effects for hepatitis delta virus ribozyme catalysis that reveal a dissociative, metaphosphate-like transition state in stark contrast to the late, associative transition states observed for reactions catalyzed by specific base, Zn²⁺ ions, or ribonuclease A. This new information provides evidence for a discrete ribozyme active site design that modulates the RNA cleavage pathway to pass through an altered transition state.

Figure 1.

Transition states for RNA cleavage and where to find them. (a) Two-dimensional reaction coordinate diagram with 5′O–P bond breaking and 2′O–P bond formation proceeding along the horizontal and vertical axes, respectively. Regions of transition states for associative versus dissociative (red diagonal) pathways having early versus late (blue diagonal) character are shown inside the box, whereas metastable intermediates are depicted outside the box. Potential concerted reaction pathways are shown as gray dashed lines. Stepwise mechanisms with either phosphorane or metaphosphate intermediates proceed sequentially along individual axes. (b) Two classes of the transphosphorylation mechanism: stepwise with an intermediate (Int.) and two transitions states (TS1 and TS2) and concerted with a single TS1. (c–e) Experimentally guided, computationally derived transition state models of RNA 2′-O-transphosphorylation consistent with ¹⁸O KIEs reported previously (Table 1) for reactions catalyzed by (c) specific base (OH⁻), (d) Zn²⁺ ions, and (e) RNase A. Key distances in Å are indicated. (f) Catalytic mechanism of HDVr showing the proposed roles of protonated C75 acting as a general acid and active site Mg²⁺ ion coordinated to a non-bridging phosphoryl oxygen and participating in catalysis via coordination of the 2′O (red) or as general base via a coordinated hydroxide (orange).

Figure 2.

Normal 18k_NUC for HDVr-catalyzed RNA 2′-O-transphosphorylation determined by two alternative methods (a) sequence and secondary structure of HDVr (black), the wildtype substrate (blue), and the mutant substrate with low binding commitment (red, see the Supporting Information). (b) Positions of ¹⁸O-substitution for measurements of ¹⁸k_NUC (red), ¹⁸k_LG (green), and ¹⁸k_NPO (black). (c) Distribution of individual measurements of ¹⁸k_NUC by the ³²P/³³P remote label. The data set is fit to a Gaussian distribution (dotted line) with a peak centered at 1.019 matching the observed KIE (Table S1). The intrinsic ¹⁸k_NUC corrected for commitments and isotopic enrichment of the ¹⁸O substrate is reported in Table 1. (d) Comparison of the distribution of ¹⁸k_NUC measurements for opposite remote labeling orientation with ³²P labeled 2′-¹⁶O substrate (red) versus with ³²P labeled on the 2′-¹⁸O substrate RNA (blue). (e) ESI-TOF mass spectrum of a mixture of 2′-¹⁸O and 2′-¹⁶O substrate RNA reacting with HDVr at f = 0.9 (magenta) and f = 0.3 (blue) demonstrating enrichment in 2′-¹⁸O in the residual substrate consistent with a normal ¹⁸k_NUC. (f) Determination of ¹⁸k_NUC from fitting the change in the isotope ratio of the residual substrate (R₀ = ¹⁸O/¹⁶O ratio in the 11 mer substrate starting population; R_s = ¹⁸O/¹⁶O ratio in the residual 11 mer substrate at a fraction of reaction f) as a function of reaction progress, f. Simulated data for KIEs of 0.99, 1.0, 1.022, 1.03, and 1.04 are shown as dotted lines. Fitting of the experimental data to ln(R_s/R₀) = (1/¹⁸k – 1)ln(1 – f) + ln(R₀), where ¹⁸k is the isotope effect and f is the fraction of the substrate consumed as determined by HPLC (solid line).

Figure 3.

Comparison of idealized free energy surfaces for RNA 2′-O-transphosphorylation reactions catalyzed by RNase A (a) and HDVr (b). Experimental KIEs and calculated TS distances are depicted on the left. Transition state locations that correspond to the maxima of the one-dimensional substrate-to-product coordinate (blue) and the minima on the orthogonal coordinate (red), with the corresponding ideal non-enzymatic path (i.e., passing through a symmetric dianionic phosphorane TS) used for reference (transparent dotted lines), are illustrated on the right. (a) RNase A provides extraordinary rate enhancement by using acid/base catalysis to stabilize a transition state that is similar to non-enzymatic reactions under alkaline conditions as inferred by KIEs. (b) HDVr provides rate enhancement using a combination of nucleobase and metal ion catalytic interactions. The free energy surface is distorted relative to ideal non-enzymatic reactions and RNase A toward a dissociative path and metaphosphate-like transition state marked by limited 2′O–P bond formation. These surfaces and pathways are meant to be illustrative, with the relative heights of the enzymatic and ideal non-enzymatic TSs normalized such that their locations on the surface are clearer.

Figure 4.

Summary of computational results. Structural description of the transition states obtained by the dissociative (a) and associative (b) strings further geometry optimized using PBE0/6-31G* for QM description.

Figure 5.

HDV active site comparison. (a) HDVr active state and (b) HHr active state. HDVr and HHr active states use similar Mg²⁺ binding modes to enable different catalytic strategies (general base “γ” catalysis in HDVr and general acid “δ” catalysis in HHr).