Imidazole Rescue of a Cytosine Mutation in a Self-Cleaving Ribozyme

See allHide authors and affiliations

Science  01 Oct 1999:
Vol. 286, Issue 5437, pp. 123-126
DOI: 10.1126/science.286.5437.123


Ribozymes use a number of the same catalytic strategies as protein enzymes. However, general base catalysis by a ribozyme has not been demonstrated. In the hepatitis delta virus antigenomic ribozyme, imidazole buffer rescued activity of a mutant with a cytosine-76 (C76) to uracil substitution. In addition, a C76 to adenine substitution reduced the apparent pK a (whereK a is the acid constant) of the self-cleavage reaction by an amount consistent with differences in the pK a values of these two side chains. These results suggest that, in the wild-type ribozyme, C76 acts as a general base. This finding has implications for potential catalytic functions of conserved cytosines and adenines in other ribozymes and in ribonuclear proteins with enzymatic activity.

Transphosphoesterification reactions catalyzed by self-cleaving and self-splicing RNAs (ribozymes) require loss of a proton from the participating 2′- or 3′-hydroxyl group to promote its nucleophilic attack on the cleavage-site or splice-site phosphate (1, 2). Metal ions can assist in this reaction, and metal-ion catalysis is one of several strategies that ribozymes share with protein enzymes (1, 2). Enhanced nucleophilicity of the hydroxyl group could also result from base-catalyzed deprotonation (1–3). The pK a values of the nucleoside side chains (pK a ∼ 3.5 to 4.5), however, appear to be too low to provide efficient general acid-base catalysis at physiologic pH (4). Although pK a values can be shifted closer to neutrality in particular RNA structures (2, 5), it has not been demonstrated that an RNA side chain can act as a general base in catalysis (1,2).

The two hepatitis delta virus (HDV) ribozymes are structurally related self-cleaving RNAs (6, 7) that require a 2′-hydroxyl group on the ribose located immediately 5′ of the cleavage site phosphate (8) and that generate products containing a 2′,3′-cyclic phosphate and a 5′-hydroxyl group (9). Thus, implied is a cleavage mechanism that involves nucleophilic attack of the 2′-hydroxyl or 2′-alkoxide on the cleavage-site phosphorus (Fig. 1). In the HDV ribozymes, a specific cytosine (C75 in the genomic ribozyme, designated γC75, and its counterpart C76 in the antigenomic ribozyme) has been hypothesized to accept the proton from the attacking 2′-hydroxyl group (10, 11).

Figure 1

Proposed mechanism for general base catalysis by imidazole and RNA side chains. In the C76u mutant, the ring nitrogen of imidazole is the potential proton acceptor. In the wild-type (top inset) and C76a mutant (bottom inset), cytosine and adenine, respectively, could act as general bases. Other likely players in catalysis for which there is no specific evidence in the HDV ribozymes, including metal ions and the proton donor for the 5′ leaving group, are not shown.

To establish that the cytosine base at position 76 was essential for cleavage in the antigenomic ribozyme, we tested whether exogenous cytosine could rescue activity of C76 mutants. We introduced mutations at C76 into the PEX1 antigenomic ribozyme sequence (12); consistent with previous findings (13), self-cleavage activity of C76u and C76g was undetectable under standard conditions (Fig. 2A). At 37°C, the rate constants were down by a factor of 106. Cleavage activity of the C76u ribozyme was partially restored when cytosine was added to the reaction mixture (14). Rescue of activity by exogenous bases and base analogs has previously been demonstrated in hammerhead ribozymes containing abasic residues (15). In those studies, rescue occurred through compensation of structural changes introduced by the abasic residue. To test whether cytosine rescue of the C76u mutant might reflect a catalytic role for the base, we substituted imidazole for cytosine. For C76u and C76g, addition of 200 mM imidazole (pH 7.4) to the reaction mixtures enhanced cleavage activity at least 250- and 25-fold, respectively (Fig. 2A and Table 1). The 3′ product band in these reactions was the same size as the normal 3′ product, which suggests that imidazole-dependent cleavage occurred at the wild-type cleavage site in the 101-nucleotide (nt) precursor. We tested several other buffers, but only imidazole and 4(5)-methylimidazole enhanced cleavage activity (14). A divalent cation (Mg2+, Ca2+, or Mn2+) was required for cleavage of all constructs (Fig. 2A) (16).

Figure 2

(A) Dependence of ribozyme cleavage activity on imidazole and Mg2+. Radiolabeled precursor RNA was heated for 1 min to 95°C in 0.1 mM EDTA (pH 8.0) and then preincubated at 37°C for 10 min in 0.5 mM spermidine containing 1 mM EDTA buffered with either 40 mM tris-HCl (pH 7.5) or 200 mM imidazole (pH 7.4). All pH values given in this paper were measured under final reaction conditions. Reactions (37°C) were initiated by adding MgCl2 (11 mM final concentration). Samples were quenched with formamide containing 50 mM EDTA and fractionated on a 6% polyacrylamide gel containing 7 M urea. Reaction time for PEX1 was 5 s; reaction time for all mutants was 4 hours. (B) Cleavage of wild-type and C76u in imidazole generates the same 5′ product. Cleavage products of wild-type (wt) and C76u precursors (5′-end-labeled with 32P) were fractionated on a sequencing gel under conditions in which small fragments containing 2′,3′-cyclic and 2′(3′)-terminal phosphates were well resolved (19% acrylamide, 1% bisacrylamide). The shortened 5′ sequence (3 nt, 5′-GAC-) facilitated resolution of the 5′ product on the basis of charge. Precursor (Prec) lanes are untreated (lanes ø), alkaline hydrolysis ladder (lanes Hy), T1 G ladder (lanes T1), U2 A ladder (lanes U2), or allowed to self-cleave (lanes Clv). The wild type (wt) cleaved for 10 s in MgCl2, and the C76u cleaved for 4 hours in MgCl2 with imidazole. Product lanes (Prdt) are the isolated 5′ product receiving no further treatment (lanes −) or the same fragment treated with 0.1 M HCl for 4 hours at room temperature (lanes H+). Positions of the bands corresponding to mono-, di-, tri-, and tetranucleotides containing either a terminal (p) or cyclic (>p) phosphate are indicated on the left. Arrow is the position corresponding to the ribozyme cleavage site in the ladder of fragments.

Table 1

Rate constants for cleavage of PEX1 (wild type) and C76 mutants in the absence and presence of 200 mM imidazole. In reactions without imidazole, the buffer was 40 mM tris-HCl (pH 7.5). All reactions in imidazole were done at pH 7.4. The reaction temperature for PEX1 was 25°C to allow for manual measurements; the other reactions were done at 37°C. Otherwise, the conditions were the same as those described in Figs. 2 and 3. Maximal rate constants for cleavage of C76g and C76u in the absence of imidazole were estimated from the average extent of cleavage after 20 hours; there was a twofold variation in these numbers between experiments. Rate constants for the other reactions were obtained from a fit to a mechanism for a single exponential reaction. The values reported are the averages of three determinations, and, other than for PEX1, variation between rate constants in individual experiments was within 10%. Enhancement (fold) is defined as [k obs(imidazole)]/[k obs (no imidazole)].

View this table:

Imidazole would most likely be acting as either a general base (pK a ∼ 7.0) or a nucleophile in the cleavage reaction. If imidazole acted as a nucleophile in a single-displacement reaction, it should show up in one of the products. However, a 2′,3′-cyclic phosphate would be generated if the adjacent 2′-hydroxyl was the nucleophile. Therefore, we characterized the 5′ cleavage product. For this analysis, the sequence 5′ to the cleavage site in both PEX1 and C76u was shortened from 8 to 3 nt. Wild-type and mutant precursor RNAs were 5′-end-labeled and allowed to cleave in the absence and presence of imidazole, respectively. The 5′ cleavage products for both ribozymes comigrated on polyacrylamide gels under denaturing conditions in which short fragments containing 3′(2′)-terminal phosphates were resolved from fragments containing 2′,3′-cyclic phosphates (Fig. 2B). To confirm generation of a cyclic phosphate, we isolated 5′ products from both reactions and hydrolyzed them with acid. We observed a shift in mobility after hydrolysis when the products were fractionated by gel electrophoresis (Fig. 2B) or thin-layer chromatography (16). This shift is consistent with the presence of an additional negative charge carried by a terminal phosphate. These results demonstrate that the mutant ribozyme cleaved at precisely the wild-type position and suggest that cleavage most likely occurred by attack of the adjacent 2′-hydroxyl group on the phosphorus.

The magnitude of imidazole rescue was greatest for the C76u mutant (Fig. 3A and Table 1); therefore, we focused on the mechanism of rescue of this ribozyme. The cleavage rate constant for the C76u ribozyme was linearly dependent on imidazole concentration (25 to 200 mM), and the slope of the line increased with increasing pH (Fig. 3B). Extrapolation to zero imidazole gave an imidazole-independent rate constant of 2.1 × 10−3hour−1 at pH 7.4. At higher imidazole concentrations, the increase in the rate constant showed only slight curvature (k obs = 0.71 hour−1 at 500 mM imidazole, pH 8.3); however, degradation of RNA became more pronounced at these higher concentrations. If there is a saturable imidazole binding site, the equilibrium dissociation constant (K d) is estimated to be >1.2 M. The pH dependence of the apparent second-order rate constant indicated an apparent pK a of 7.0 ± 0.1 (Fig. 3C). The concentration and pH dependence data for the C76u mutant suggest that imidazole base is the active species, and, although other mechanisms dependent on the concentration of imidazole base are plausible (see below), these data are consistent with imidazole-enhanced cleavage by a general base mechanism.

Figure 3

(A) Kinetics of cleavage of C76u (circles) and C76g (triangles) with (closed symbols) and without (open symbols) imidazole. RNA was preincubated, as described in Fig. 2, in either 200 mM imidazole (pH 7.4) or 40 mM tris-HCl (pH 7.5). MgCl2 was added to start the reactions, and aliquots were quenched and fractionated on polyacrylamide gels. Results were quantified by PhosphorImager analysis, and the fraction cleaved was calculated from the amount of radioactivity in the precursor and 3′ product after correcting for the label in the 5′ product. Rate constants are reported in Table 1. (B) Concentration and pH dependence of the cleavage rate constant for C76u in imidazole. Pseudo-first-order rate constants in increasing imidazole concentrations (25 to 200 mM) at the pH indicated were obtained as described above. In the pH range 6.0 to 8.3, the fraction of unprotonated imidazole ranges from 10% to 95% (assuming a pK a of 6.95). k obs was linearly dependent (R values ≥ 0.95) on imidazole concentration in this range with apparent second-order rate constants (M−1 hour−1) of 7.8 × 10−2(pH 6.0), 4.7 × 10−1 (pH 6.6), 9.9 × 10−1 (pH 7.0), 1.4 (pH 7.4), and 1.8 (pH 8.3). (C) The pH dependence of C76u cleavage rate constants in imidazole. The apparent second-order rate constant for cleavage in imidazole is plotted as a function of pH. Fitting the data to the expression k app =k max/(1 + 10p K a −pH) gives an apparent pK a of 7.0 ± 0.1 (arrow). (D) Comparison of pH dependence of wild-type (PEX1) and the C76a ribozymes. Conditions for cleavage reactions were the same as for Fig. 2 except for the buffer system, which was 25 mM acetic acid, 25 mM MES, 50 mM tris (pH 4.0 to pH 8.0) or 50 mM MES, 25 mM tris, 25 mM 2-amino-2-methyl-1-propanol (pH 7.0 to 9.0) (24) instead of 40 mM tris-HCl. The triple buffer systems were titrated with HCl or NaOH, and the ionic strength was maintained within 10% of 0.5. The observed rate constants (min−1) are plotted as a function of pH. Apparent pK a values (arrows) for the two ribozymes were obtained by fitting the data to k obs =k max/(1 + 10p K a −pH). Data analysis and graph preparation were done with KaleidaGraph (Synergy Software, Reading, PA).

There are, however, at least two additional mechanisms that could explain imidazole-dependent cleavage in the C76u mutant. First, imidazole may structurally compensate for the mutation but not participate in the chemistry. Pyrazole (1,2-diazole; pK a ∼ 2.5), a close structural analog of imidazole but a much weaker base, did not rescue activity (16). Competition experiments indicated that pyrazole inhibited the imidazole-dependent reaction [inhibition constant (K i) ∼ 300 mM], so it appeared to bind but not to facilitate cleavage. The second possibility is that imidazole could coordinate a catalytic metal ion, presumably replacing a ligand lost with the mutation of C76. For example, the N-3 of C60 in yeast tRNAPhe is a Mg2+ligand. That interaction is through a water molecule (17), and mutating the cytosine affected the concentration and specificity of Mg2+-catalyzed hydrolysis of the tRNA (18). Thus, catalytic metal ion coordination appears to be a potentially viable mechanism. However, if imidazole substituted for C76 as a metal ligand in the ribozyme reaction, pyrazole should have fulfilled a similar function. Additionally, increasing the MgCl2 concentration from 10 to 250 mM or adding Mn2+ (2 and 10 mM, with and without Mg2+) did not restore site-specific self-cleavage activity in the C76u mutant in the absence of imidazole. In the presence of imidazole, addition of 2 mM Mn2+ did not further stimulate the reaction (16). Finally, metal coordination to γC75 in the crystal structure of the HDV genomic ribozyme cleavage product was not observed (10), although, if metal binding requires coordination to both γC75 and the scissile phosphate, the metal ion may not remain strongly bound in the cleavage product (10).

The simplest explanation for imidazole-dependent cleavage of the C76u ribozyme is that imidazole acts as a general base to accept a proton from the 2′-hydroxyl group of the nucleotide 5′ to the cleavage site (Fig. 1). This interpretation could mean that, in the wild-type antigenomic sequence, the N-3 of C76 accepts the proton, as proposed for γC75 in the genomic ribozyme (10). If C76 is acting as a general base, replacing it with a base of a different pK a would be expected to shift the apparent pK a of the cleavage reaction in a predictable direction and amount. Because the C76a mutant was marginally active, we examined it in this regard. If it is assumed that there are structural similarities for the two HDV ribozymes, the adenine N-6 amino group could potentially form the same hydrogen bond network as the cytosine N-4 amino group, thus positioning the N-1 of adenine near the N-3 of the cytosine it replaced. However, the N-1 of adenosine (pK a ∼ 3.5) is a poorer proton acceptor than the N-3 of cytidine (pK a ∼ 4.2) (4). The pH dependence of the cleavage rate constants for the two ribozymes revealed a decrease in the apparent pK a from 6.1 for wild type to 5.6 for the C76a mutant (Fig. 3D). Finding that the apparent pK ahad decreased by an amount roughly equivalent to the pK a difference of the two nucleosides is consistent with both cytosine and adenine functioning as general bases at position 76.

The general base mechanism for imidazole in these studies is supported by a strong precedent for the role of imidazole, or an imidazole side chain, as a general base in RNA cleavage reactions (3,19). There is also precedent for exogenous imidazole acting as a general base—for example, to rescue the activity of histidine mutations of bacterial luciferase (H44A) (20) and horseradish peroxidase (H42A) (21). Imidazole rescue of a side-chain mutant of a naturally occurring RNA enzyme expands the application of this method and suggests that general acid-base catalysis by nucleotide side chains is a mechanism that can be used in RNA-catalyzed reactions.

  • * To whom correspondence should be addressed. E-mail: been{at}


View Abstract

Stay Connected to Science

Navigate This Article