Nucleobase-mediated general acid-base catalysis in the Varkud satellite

Timothy J. Wilsona, Nan-Sheng Lib, Jun Lub, John K. Frederiksenb,2, Joseph A. Piccirillib,1, and David M. J. Lilleya,1

aCancer Research United Kingdom Research Group, MSI/WTB Complex, The University of Dundee, Dow Street, Dundee DD1 5EH, United Kingdom; and bThe University of Chicago Center for Integrative Science, Departments of Biochemistry and Molecular Biology and Chemistry, 929 East 57th Street, Room W406, Chicago, IL 60637

Edited by Olke C. Uhlenbeck, Northwestern University, Evanston, IL, and approved May 19, 2010 (received for review March 30, 2010)

2þ Existing evidence suggests that the Varkud satellite (VS) ribozyme At high concentrations of Mg ions, the pH dependence of accelerates the cleavage of a specific phosphodiester bond using the cleavage reaction of the trans-acting VS ribozyme is bell general acid-base catalysis. The key functionalities are the nucleo- shaped, fitting a model involving proton transfers in the transition K ¼ 5 2 bases of adenine 756 in helix VI of the ribozyme, and guanine 638 state with participating groups of p a . and 8.4 (9). Substitu- in the substrate stem loop. This results in a bell-shaped dependence tion of G638 by diaminopurine shifts the pH profile, correspond- of reaction rate on pH, corresponding to groups with pK ¼ 5.2 and K a ing to new p a values of 4.6 and 5.6. A plausible mechanism 8.4. However, it is not possible from those data to determine which involves general acid-base catalysis by A756 and G638, but it nucleobase is the acid, and which the base. We have therefore is not possible to determine from the pH dependence which ′ made substrates in which the 5 oxygen of the scissile phosphate nucleobase acts as the acid, and which the base. The alternatives is replaced by sulfur. This labilizes the leaving group, removing the predict identical pH profiles. Smith et al. have suggested a requirement for general acid catalysis. This substitution restores resolution of this ambiguity, based on the relationship between full activity to the highly impaired A756G ribozyme, consistent K ionic environment and nucleobase p a (15), but there is no with general acid catalysis by A756 in the unmodified ribozyme. direct evidence enabling the assignment of function to specific The pH dependence of the cleavage of the phosphorothiolate- nucleobases. modified substrates is consistent with general base catalysis by A similar ambiguity existed for the HDV ribozyme, where the nucleobase at position 638. We conclude that cleavage of the critical functionalities are a cytosine nucleobase (16–18) and a substrate by the VS ribozyme is catalyzed by deprotonation of metal ion-bound water (2). Assignment of the general acid and the 2′-O nucleophile by G638 and protonation of the 5′-O leaving group by A756. base was both difficult and controversial. The distinction was made by the introduction of a 5′-phosphorothiolate (5′-PS) substitution ′ 5′-phosphorothiolate ∣ RNA catalysis ∣ nucleolytic ∣ at the scissile phosphate (19). The 5 sulfur atom is a much better catalytic mechanism leaving group than oxygen and therefore no longer requires protonation by a general acid. Thus alterations to the ribozyme ibozyme-mediated catalysis is important for both RNA that impair the function of the general acid (so inhibiting cleavage ′ Rsplicing and translation (1), yet its chemical origins are incom- of the oxy substrate) should have little effect on cleavage of a 5 -PS- pletely understood. The nucleolytic ribozymes bring about the containing substrate. Moreover, the pH dependence of the site-specific cleavage or ligation of RNA, with an acceleration cleavage rate of the 5′-PS substrate should reflect the deprotona- of a millionfold or greater. The intensively studied protein tion of the base alone. We have therefore synthesized VS ribozyme RNase A catalyses an identical cleavage reaction, and much substrates containing a 5′-PS linkage at the cleavage site. The evidence supports the hypothesis that each of these phosphoryl substitution restores full activity to the highly impaired A756G transfer reactions is subject to general acid-base catalysis. This ribozyme, and we conclude that the nucleobase of A756 is the mechanism requires a general base to deprotonate the attacking probable general acid in the cleavage reaction. nucleophile, and a general acid to protonate the oxyanion leaving group (Fig. 1). Results The most common chemical entities implicated in RNA Kinetic Analysis of Cleavage Using Modified VS Ribozyme and catalysis by the nucleolytic ribozymes are the nucleobases (2). Substrate. The analysis of VS ribozyme cleavage reactions was Guanine appears to play a catalytic role in the hairpin, hammer- performed using a ribozyme comprising helices II to VI acting head, GlmS, and Varkud satellite (VS) ribozymes, adenine in the in trans upon a substrate stem loop (Fig. 1A). The ribozyme hairpin, and VS and cytosine in the hepatitis delta virus (HDV) was prepared by transcription from a DNA template, and the BIOCHEMISTRY ribozyme. Crystal structures of the hairpin ribozyme (3) reveal substrate by a combination of chemical synthesis and sequential the presence of guanine (G8) and adenine (A38) bases juxta- ligations (Figs. S1 and S2). A 5′-PS-substituted scissile phosphate posed with the 2′-O and 5′-O, respectively, of the scissile phos- was introduced by synthesis of a GpA dinucleotide containing a phate, where they seem poised to act in general acid-base 5′ sulfur on the adenosine with o-nitrobenzyl protection on the catalysis. This is consistent with the pH dependence of the reac- tion (4) and its variation with functional group modifications (5–8). Author contributions: T.J.W., N.-S.L., J.L., J.K.F., J.A.P., and D.M.J.L. designed research; CHEMISTRY T.J.W., N.-S.L., J.L., and J.K.F. performed research; T.J.W. and D.M.J.L. analyzed data; In its simplest active form, the VS ribozyme comprises five and T.J.W., J.A.P., and D.M.J.L. wrote the paper. helices (II through VI) organized by two three-way junctions, The authors declare no conflict of interest. which acts in trans upon a substrate stem loop (helix I) with an This article is a PNAS Direct Submission. internal loop that contains the scissile phosphate (Fig. 1). The loop – 1To whom correspondence may be addressed. E-mail: [email protected] or jpicciri@ also contains the critical G638 (9). A756 (10 13) is contained chicago.edu. within an internal loop in helix VI. While no crystal structure 2Present address: Department of Pathology and Laboratory Medicine, The University of of the VS ribozyme has yet been solved, a small-angle X-ray scat- Rochester Medical Center, 601 Elmwood Avenue, Box 626, Rochester, NY 14642. tering-derived model places G638 and A756 in proximity to the This article contains supporting information online at www.pnas.org/lookup/suppl/ scissile phosphate (14). doi:10.1073/pnas.1004255107/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1004255107 PNAS ∣ June 29, 2010 ∣ vol. 107 ∣ no. 26 ∣ 11751–11756 Downloaded by guest on September 24, 2021 A C G IV V U U C U U G C U G A A A U U G C G U A G C A G U G A C G U A A C U U U A A C G U A U U G U C A C I G C G C U G C U U C G U G G C C G U A A C 638 A U 621 A or DAP 5‘ III A G C G G G A U A C 3‘ C C G G G A 730 G U A A A A A A G U G U A A G C G G G G C U G UG C G G U A U U G G C U 5‘ GUGUCGCAAUC A C G U U C G C C C C G A CC A G U U A U G A C U G 3‘ U A A G A G G A A C A G A C A or II 756 VI G

B C D O Gua N NH2 O NH N HO N N N NH2 O O H :B O HO N N NH P 2 O O O O oNO2Bn NH2 A H O O P HO OH Ade O N O O N S 2, 6-diaminopurine N N O OH O

HO OH

Fig. 1. The sequence of the VS ribozyme, the proposed mechanism, and the chemical structure of the 5′-phosphorothiolate substitution. (A). The sequence of the trans-acting ribozyme and substrate used here. The cleaved bond is arrowed. (B). Mechanism of general acid-base catalysis for the cleavage reaction of the VS ribozyme. (C). The chemical structure of a protected GpA dinucleotide with a 5′-phosphorothiolate linkage. oNO2Bn ¼ o-nitrobenzyl. (D). The chemical structure of 2,6-diaminopurine nucleoside.

2′-hydroxyl of the guanosine to prevent premature activation of the labile P-S bond during deprotection. Addition of the natural- the nucleophile (Fig. 1C). sequence (A756) ribozyme resulted in significant cleavage of both 0 32 The cleavage of radioactively ½5 - P-labeled substrate was 5′-PO and 5′-PS substrates. The 5′-PO substrate is poorly cleaved studied under single-turnover conditions immediately following by a ribozyme carrying the A756G substitution, consistent with deprotection of the 2′-hydroxyl nucleophile using ultraviolet previous observations (10). By contrast, the 5′-PS substrate was irradiation. Products of ribozyme cleavage were separated by substantially cleaved by the A756G ribozyme. The extent of gel electrophoresis and quantified by phosphorimaging. product formation for this reaction is shown in Fig. 2B, and the reaction progress is plotted in Fig. 2C for each ribozyme reaction. 5′-PS Substitution Restores Cleavage by a Ribozyme with a A756G The A756 ribozyme cleaves the 5′-PO and 5′-PS substrates at Substitution. We have adopted standard conditions of 50 mM 2.4 min−1 and 0.37 min−1, respectively. The A756G ribozyme MES (pH 6.0), 25 mM KCl, and 200 mM MgCl2 at 37 °C. All cleaves the 5′-PO substrate at 0.00023 min−1, yet cleaves the the measured reaction rates are tabulated in Table 1. −1 Under standard conditions, in the absence of ribozyme the 5′-PS substrate 3,700-fold faster at 0.84 min . This rate of normal substrate carrying an 5′-oxy scissile phosphate group at 5′-PS substrate cleavage is twice that observed for the A756 A621 (5′-PO) exhibited no detectable cleavage after 15-min ribozyme and represents complete restoration of A756G ribozyme incubation, whereas a small fraction of product was observed activity by the 5′-PS linkage. This restoration is consistent with for the 5′-PS substrate (Fig. 2A). This cleavage occurs selectively the A756 nucleobase acting as the general acid in the unmodified at the position cleaved by the VS ribozyme and is due to cleavage of cleavage reaction.

k trans Table 1. Observed rates of substrate cleavage ( obs)in measured under standard conditions ′ k ∕ −1 ′ k ∕ −1 k 50 ∕k 50 Rz Substrate 5 -PO, obs min 5 -PS, obs min obs -PS obs -PO — wt <10−5 0.0012 ± 0.0001 >120 wt wt 2.4 ± 0.3 0.37 ± 0.03 0.16 A756G wt 0.00023 ± 0.00006 0.84 ± 0.05 3,700 C755G wt 0.14 ± 0.01 0.35 ± 0.03 2.5 G757A wt 0.39 ± 0.07 0.22 ± 0.04 0.56 A730U wt 0.0042 ± 0.0004 0.038 ± 0.002 9.0 A756C wt 0.029 ± 0.006 0.080 ± 0.009 2.8 wt DAP 0.037 ± 0.004 0.30 ± 0.02 8.1 A756G DAP 0.0003 ± 0.0001 1.05 ± 0.02 3,500 Single-turnover rates were determined for the indicated combinations of ribozyme and substrate and either 5′-PO or 5′-PS substitution. Unmodified ribozyme or substrate sequence is indicated by wt. Each rate is the average of ≥3 independent measurements

11752 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1004255107 Wilson et al. Downloaded by guest on September 24, 2021 A sub + no rz A756 rz A756G rz If A756 is the general acid, G638 is the likely general base and O S O S O S inhibitory substitutions at this position are not expected to be res- cued by the 5′-PS substitution. Replacement of G638 by diami- nopurine (DAP, Fig. 1D) leads to significantly lower rates of substrate substrate cleavage (9). Under standard conditions a 5′-PO substrate carrying the G638DAP substitution was cleaved by the A756 ribozyme at a rate of 0.037 min−1. Introduction of product the 5′-PS substitution to the G638DAP substrate resulted in 8-fold faster cleavage by the A756 ribozyme at a rate of 0.30 min−1 (Table 1). While this increase is consistent with a 1 2 3 4 5 6 mechanism in which G638 acts as a general base (see Discussion), it is necessary to examine the reactivity of 5′-PS substrates over a B A756G rz + 5’ PS substrate broad pH range to clarify the respective roles of G638 and A756, 0.24 0.67 1.3 2.25 3.5 5 min because the pH dependence of the cleavage rate should reflect the pK of the general base alone. substrate a The Rate of Cleavage Reaction of a 5′-PS-Substituted Substrate Is Not product Reduced at pH Values at Which Adenine Would Deprotonate. The pH dependence of the cleavage rate for the natural VS ribozyme reaction is bell shaped, corresponding to the ionization of bases K with apparent p a values of 5.2 and 8.4, assigned to A756 and G638, respectively (9). In both cases the local environment shifts C 0.8 A756G rz + 5’ PS K the p a from the solution values of 3.8 for free adenosine and 9.4 A756 rz + 5’ PO for free guanosine. The pH profile changes when a G638DAP 0.6 substrate is cleaved by the natural-sequence ribozyme, with K apparent p a values of 4.6 and 5.6 assigned to A756 and 0.3 DAP638, respectively, because free DAP has the higher unper- A756G rz 0.4 K 0.2 + 5’ PO turbed p a of 5.1 (9). We may calculate the fraction of the acid A756 rz in the protonated form (f ) and base in its deprotonated form + 5’ PS A 0.1 f K

fraction cleaved fraction ( ) from the observed p values. The rate of a reaction requir- 0.2 B a ing both acid and base should follow the shape of the product 0 f f A 0 1000 2000 3000 A · B, showing the origin of the bell-shaped profile (Fig. 3 ). 0 The rate is reduced at low pH due to increased protonation of 012345 the base, and at high pH due to deprotonation of the acid. How- time / min ever, if labilization of the leaving group leads to a reaction that is independent of general acid catalysis, the reaction rate should Fig. 2. Cleavage of natural-sequence substrate by A756 or A756G VS f ribozyme as a function of the presence or absence of a 5′-phosphorothiolate depend solely on the fraction of unprotonated base, i.e., B (also linkage at the scissile phosphate. (A). Gel electrophoresis of reaction plotted in Fig. 3A). In this case the reaction rate increases to products. The 5′-PO (tracks 1, 3, 5) and 5′-PS (tracks 2, 4, 6) substrates were pH ∼ 6.5 and then forms a plateau. The rate does not diminish incubated with no ribozyme (tracks 1, 2), A756 ribozyme (tracks 3, 4), or at higher pH, because the protonated form of the acid is not A756G ribozyme (tracks 5, 6) for 15 min. (B). Cleavage of the 5′-PS substrate required for the catalytic activity of the ribozyme. catalyzed by the A756G ribozyme as a function to time (indicated above gel). We have measured the rate of cleavage of the G638DAP (C). Plots of reaction progress with time, fitted to single exponential 0 0 substrate with a 5′-PS substitution by the A756 ribozyme as a functions. Filled circles, 5 -PO þ A756 ribozyme; open circles, 5 -PS þ A756 B ribozyme; open squares, 50-PS þ A756G ribozyme; filled squares (Inset), function of pH (Fig. 3 ) and compared this with the profile 0 ′ 5 -PO þ A756G ribozyme. Corresponding axes use the same units. for the 5 -PO substrate measured previously. The latter follows a bell-shaped rate profile (9), but the rate of cleavage of the ′ ∼ 6 We examined the effect of the 5′-PS modification on other 5 -PS substrate increases up to pH after which the reaction rate becomes independent of pH. The data for the 5′-PS substrate variations within the A730 loop of the ribozyme. The C755G −1 fit well to a model involving a single ionization, with an apparent ribozyme cleaves the 5′-PO substrate at 0.14 min (17-fold pK of 5.3. They are fully consistent with a reaction in which the slower than the natural ribozyme); this rate increased slightly to a −1 cleavage of the labilized substrate is subject to general base 0.35 min for the 5′-PS substrate. The G757A ribozyme cleaves BIOCHEMISTRY catalysis by the diaminopurine at position 638. However, these the 5′-PO substrate at 0.39 min−1 under these conditions. This was −1 data alone do not exclude the possibility that A756 acts as the barely affected by the 5′-PS modification (a rate of 0.22 min ). K general base, as the apparent p a is only 0.7 units higher than Thus there is no evidence for a direct role for either C755 or G757 that assigned to A756 in the presence of G638DAP. in leaving group stabilization. The A730U ribozyme is much less If the nucleobase at position 638 acts as the general base, then active, cleaving the 5′-PO substrate at a rate of 0.0042 min−1.It −1 the pH dependence of the natural-sequence substrate (i.e., with cleaves the 5′-PS substrate only 9-fold faster at 0.038 min .We ′ K G638) bearing the 5 -PS modification should correspond to a p a examined a second sequence variant at the critical 756 position, close to 8.4 (9). The expected pH dependence is simulated in CHEMISTRY A756C. This ribozyme cleaved the 5′-PO substrate at a rate of A −1 Fig. 3 (broken lines). The reaction rate should follow the extent 0.0029 min , 80-fold slower than the A756 ribozyme. Interest- of deprotonation of the guanine, and this increases log-linearly up ingly, it cleaved the 5′-PS substrate only 3-fold faster at a rate to approximately pH 8. Conversely, if A756 acts as the general −1 of 0.080 min . These results suggest that cytosine can replace base, the pH dependence of the natural-sequence 5′-PS substrate adenine to some degree under these conditions, unsurprisingly should follow the deprotonation of the adenine and would be because the bases share an amidine functionality with similar similar to that of the G638DAP substrate presented in Fig. 3B. K p as, but the substitution results in a structure with a reduced in- Over the observable pH range the experimental rates for the trinsic activity; this may reflect a change in the active site structure 5′-PS substrate increase log-linearly up to pH 6.5 (Fig. 3C, open K due to the substitution of a pyrimidine for a purine nucleobase. circles). Thus the general base must have a p a greater than 7,

Wilson et al. PNAS ∣ June 29, 2010 ∣ vol. 107 ∣ no. 26 ∣ 11753 Downloaded by guest on September 24, 2021 f (pK = 5.6) even in the presence of the G638DAP substitution, activity is A 1 B a B fully restored in the presence of the highly deleterious A756G 0.4 G638DAP, S fA modification. fB (pKa = 8.4) or -2 10 rate / Discussion f B min-1 0.2 Significant data implicate A756 and G638 in the catalytic -4 mechanism of the VS ribozyme (9–13). The effect of substitution 10 fA (pKa = 4.6) 0 at these positions and the pH dependence of rates are consistent 1 G638DAP, S with general acid-base catalysis by these two nucleobases (9). The fB (pKa = 5.6) 1 change in rate in D2O also suggests that proton transfer occurs in

f .f the rate-limiting step of cleavage (20). A756G substitution lowers A B rate / G638DAP, O fB min-1 10 -2 cleavage activity by 4 orders of magnitude at pH 6.0, but substitu- 10 -2 or tion of the 5′-O leaving group by sulfur enhances the rate 3,800-

fA.f B fold by the A756G ribozyme. Thus, labilization of the leaving group so that protonation is no longer required to facilitate bond 10-4 fB (pKa = 8.4) 10 -4 breakage suppresses the inhibitory effect of the A756G substitu- 5678 tion. The A756G ribozyme cleaves the natural-sequence 5′-PS 45678910 pH pH substrate twice as fast as the A756 ribozyme, suggesting that the structure of the ribozyme is not altered by the substitution C 10 so as to affect its activity. Other substitutions in the A730 loop A756G of helix VI have smaller effects on cleavage rates than those A756 at 756 and show similar rates for 5′-PO and 5′-PS substrates. This 1 confirms the importance of A756 to the reaction mechanism. Identifying A756 as the general acid in the cleavage reaction rate / G638DAP min-1 leaves G638 as the probable general base. The base would be ex- pected to facilitate proton transfer from the 2′-O nucleophile and -1 10 should still be required irrespective of the lability of the leaving group. The log-linear pH dependence below pH 7 of the cleavage of the natural-sequence 5′-PS substrate is fully consistent with 10 -2 general base catalysis by guanine and excludes adenine from this 45678 pH role. The pH dependence of the ribozyme-mediated cleavage changes markedly for the 5′-PS G638DAP substrate. The rate Fig. 3. The pH dependence of cleavage rates. (A) Simulation of reaction rate rises with pH and then remains constant beyond pH 6 giving a Upper f K as a function of pH. ( ) Fraction of protonated nucleobase ( A) with calculated p a of 5.3. This change in pH-rate profile induced K ¼ 4 6 f K ¼ 5 6 p a . , and unprotonated nucleobase ( B) with p a . (unbroken line) by the G638DAP substitution strongly suggests that G638 is K (corresponding to the apparent p a values for the combination of A756 with the general base. f K ¼ 8 4 Low- DAP638), and unprotonated G638 ( B) with p a . (broken line) (9). ( er f f The proposed roles for A756 and G638 also account for other ) Comparison of A · B corresponding to a reaction catalyzed by general ′ K ¼ 4 6 f K observations. At pH 6.0 the A756 ribozyme cleaves the 5 -PS acid-base catalysis (p a . and 5.6), with single B values of p a 5.6 and 8.4 corresponding to reactions catalyzed by general base catalysis alone G638DAP substrate 8-fold faster than the corresponding using DAP and G, respectively. The shaded regions reflect pH values outside 5′-PO substrate. The faster rate probably arises from the increase the experimentally accessible range. (B) Experimental profiles of cleavage in the fraction of ribozyme in the active state. At pH 6.0, for clea- rate as a function of pH for the G638DAP substrate with the natural-sequence vage of the 5′-PO substrate the proportion of acid and base in the ribozyme. (Upper) Rate of cleavage of the G638DAP, 5′-PS substrate plotted ∼0 03 f f A active state is . ( A · B in Fig. 3 ). The active fraction is on a linear scale. (Lower) Logarithmic scale of cleavage rate for the G638DAP, predicted to increase 20-fold for cleavage of the 5′-PS substrate 5′-PO (filled circles, taken from ref. 9) and G638DAP, 5′-PS (open circles) ′ where the fraction of protonated acid is not important substrates. The G638DAP, 5 -PO substrate has been fitted to a double-ioniza- [f ðpK ¼ 5.6Þ in Fig. 3A], potentially accounting for the tion model. By contrast, the pH dependence of the G638DAP, 5′-PS substrate B a observed increase in rate. In contrast, for the A756G ribozyme indicates a single ionization, consistent with a reaction not requiring general ′ ′ acid catalysis. (C) Experimental profiles of cleavage rate as a function of pH the active fraction will be the same for 5 -PO and 5 -PS substrates f ∼1 for the G638 5′-PS substrate with the A756 (open circles) and A756G (closed because guanine is fully protonated at pH 6 and thus A is . The circles) ribozymes. Both give log-linear pH dependence over the range 5–6.5 low activity of the A756G ribozyme for 5′-PO cleavage is likely with near unit gradient. The single-ionization fit for the cleavage of the due to guanine being a much weaker acid than adenine. These G638DAP, 5′-PS substrate cleaved by natural-sequence ribozyme has been results are therefore consistent with A756 acting as the acid superimposed on the plot (broken line, taken from B).

Gua excluding A756 from this role. The simplest, fully consistent mod- N O O el describing the data is that A756 acts as the general acid and G638 as the general base. N O O H N rib Further support for this conclusion comes from the pH depen- NH2 N P N dence of cleavage of the natural-sequence 5′-PS substrate by the O C O H2N G638 A756G ribozyme (Fig. 3 , closed circles). This too increases N N H O Ade log-linearly over the observable pH range, excluding a general rib K N O base with a p a below neutrality. We have also studied the cleavage of a G638DAP substrate by an A756G ribozyme. The A756 O OH A756G ribozyme cleaves the 5′-PO substrate at 0.0003 min−1 at pH 6, similar to the cleavage rate for the natural sequence sub- −1 strate. The 5′-PS substrate is cleaved at 1.05 min , slightly faster Fig. 4. The probable chemical mechanism of cleavage reaction of the VS than the cleavage rate of the natural-sequence substrate. Thus, ribozyme.

11754 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1004255107 Wilson et al. Downloaded by guest on September 24, 2021 k ∼3;700 −1 and G638 as the base for the cleavage reaction (Fig. 4). By cat is s . The fastest observed rate of cleavage for the VS microscopic reversibility, in the ligation reaction A756 and ribozyme acting in trans is ∼12 min−1. However, the observed ′ k G638 should act as base (deprotonating the 5 -O nucleophile) rate ( obs) will be constrained by the small fraction of ribozyme ′ f and acid (protonating the 2 -O leaving group), respectively. molecules that have a protonated A756 (i.e., A) and a deproto- f By itself the reduction in the rate of cleavage of the unmodified nated G638 ( B), i.e., substrate by A756G ribozyme could be consistent with a number k ¼ k f f ; [1] of roles for A756. The nucleobase could stabilize the reaction obs cat · A · B transition state by specific hydrogen bonding, or a protonated adenine might stabilize the dianionic phosphorane electrostati- k f f where cat is the intrinsic rate of catalysis. We calculate A · B cally. Both effects have been proposed for the hairpin ribozyme K from the measured p a values for A756 and G638 (9) to be (3, 21). Were this to be the case in the VS ribozyme, then sub- 3 8 × 10−4 k ∼520 −1 . , and hence cat is s . This will be even faster stitution of the nucleobase at position 756 could lead to lower for the rapidly cleaving cis ribozyme (31). Similarly, the k cleavage rates, as observed. But these impairments should not cat ′ for cleavage by the HDV ribozyme has been estimated to be be suppressed by activation of the leaving group by 5 -PS substi- in the range of 102 to 104 s−1 (2). Thus, the rate of cleavage tution. Our new results strongly suggest that general acid catalysis by ribozyme molecules that are in the correct ionization state by A756 is a major contributor to the catalytic rate enhancement appears comparable to that of RNaseA. All the nucleolytic ribo- of the VS ribozyme. zymes seem to employ general acid-base catalysis. In addition to While general acid catalysis by the A756 nucleobase is not the nucleobases, this class of ribozyme has evolved the use of required for the cleavage of the 5′-PS substrate, the ribozyme still ′ ′ bound hydrated metal ions (2) and glucosamine-6-phosphate contributes to catalysis of 5 -PS cleavage. Although the 5 -PS K substrate undergoes cleavage at a significant rate in the absence (32, 33). All these alternatives have a p a far from neutrality. of ribozyme (0.0012 min−1), addition of the A756 ribozyme In general these ribozymes are required to turn over only a single increases the cleavage rate 300-fold to 0.37 min−1. It is therefore time in their natural function, but the limitation of a low active K likely that the interaction of the substrate with the ribozyme leads fraction imposed by the unfavorable p a values may have been to a structural remodeling that is required for full activity. In an a significant factor holding back the evolution of RNA-based NMR structure of the substrate RNA alone (22), N1 of G638 is catalysts. Ultimately the selection of histidine with its imidazole K 6 Å from the 2′-OH group of G620. Moreover, the 2′-O nucleo- side chain whose p a is close to neutrality enabled protein phile is far from an in-line trajectory. It is likely that an association to achieve significantly greater rates where turnover of the substrate loop with the A730 loop of the ribozyme would is important. generate the structure required for catalytic activity. A similar situation exists in the hairpin ribozyme, where the active geometry Materials and Methods results from a close interaction between the A and B loops (23). Preparation of RNA. VS ribozyme was transcribed from a DNA template and purified by standard methods. To prepare substrates, GpA dinucleotides were The local structure (3, 24) is significantly altered from those of the ′ loops in isolation (25, 26). We have previously noted that the synthesized containing either a 5 oxygen or sulfur on the adenosine and o-nitrobenzyl protection on the 2′-hydroxyl of the guanosine. Dinucleotides topological placement of the key components within the two inter- were first ligated to the 5′ portion of the substrate with RNA ligase I, then to acting loops of the hairpin and VS ribozymes is very similar (9). the 3′ sequence using RNA ligase II and a DNA splint. The methods are fully In summary, the VS ribozyme catalyzes the cleavage of its sub- described in SI Text. strate by making an intimate loop-loop interaction with the A730 loop in helix VI. This most probably brings about a structural Analysis of Ribozyme Kinetics. The cleavage of radioactively ½50-32P-labeled rearrangement that aligns the 2′-O for nucleophilic attack and substrate was studied under single-turnover conditions (9). Immediately prior moves the nucleobase of G638 into position to act as a general to use substrates were irradiated for 15 min at 365 nm to activate the 2′-hy- base to deprotonate the hydroxyl of G620 thereby enhancing its droxyl. Following equilibration to 37 °C in reaction buffer, reactions were nucleophilicity. The association of the loops is expected to posi- initiated by mixing ribozyme and substrate. Products of ribozyme cleavage tion A756 to protonate the 5′-oxyanion leaving group. Our results were separated by gel electrophoresis and quantified by phosphorimaging. indicate that general acid catalysis contributes substantially to Reaction progress curves were fitted by nonlinear regression analysis to the observed rate enhancement by the VS ribozyme. It is very single or double exponential functions. The pH dependence of observed clea- ′ probable that the hairpin ribozyme uses an essentially identical vage rates of 5 -PO substrates were fitted to a double-ionization model catalytic mechanism, in which G8 and A38 are the general base appropriate for analysis of general acid-base catalysis, assuming a require- ment for one protonated and one deprotonated form. The pH dependence and acid, respectively. of observed cleavage rates of 5′-PS substrates were fitted to a single-ioniza- We may ask how the rate of chemical catalysis by the VS tion model in which the deprotonated form is assumed to be active. Full ribozyme compares with protein enzymes ? Pancreatic ribonu- descriptions of kinetic methods and analysis are given in SI Text. clease carries out the same cleavage of a phosphodiester linkage BIOCHEMISTRY in RNA using two histidine side chains as general acid and base ≤1;400 −1 K ACKNOWLEDGMENTS. We thank Selene Koo for technical assistance and (27, 28), with an observed rate of s (29). Given the p a Cancer Research UK, Howard Hughes Medical Institute, and National for acid and base of 6.2 and 5.8, respectively (30), the maximum Institutes of Health (1R56AI081987-01) for financial support.

1. Lilley DMJ, Eckstein F (2008) Ribozymes and RNA Catalysis (Royal Soc Chemistry, 9. Wilson TJ, McLeod AC, Lilley DMJ (2007) A guanine nucleobase important for catalysis Cambridge, UK), pp 1–318. by the VS ribozyme. EMBO J 26:2489–2500. 2. Nakano S, Chadalavada DM, Bevilacqua PC (2000) General acid-base catalysis in the 10. Lafontaine DA, Wilson TJ, Norman DG, Lilley DMJ (2001) The A730 loop is an CHEMISTRY mechanism of a hepatitis delta virus ribozyme. Science 287:1493–1497. important component of the active site of the VS ribozyme. J Mol Biol 312:663–674. 3. Rupert PB, Massey AP, Sigurdsson ST, Ferré-D’Amaré AR (2002) Transition state 11. Lafontaine DA, Wilson TJ, Zhao Z-Y, Lilley DMJ (2002) Functional group requirements stabilization by a catalytic RNA. Science 298:1421–1424. in the probable active site of the VS ribozyme. J Mol Biol 323:23–34. 4. Bevilacqua PC (2003) Mechanistic considerations for general acid-base catalysis by 12. Sood VD, Collins RA (2002) Identification of the catalytic subdomain of the VS RNA: Revisiting the mechanism of the hairpin ribozyme. Biochemistry 42:2259–2265. ribozyme and evidence for remarkable sequence tolerance in the active site loop. 5. Pinard R, et al. (2001) Functional involvement of G8 in the hairpin ribozyme cleavage J Mol Biol 320:443–454. mechanism. EMBO J 20:6434–6442. 13. Zhao Z, et al. (2005) Nucleobase participation in ribozyme catalysis. J Am Chem Soc 6. Kuzmin YI, Da Costa CP, Fedor MJ (2004) Role of an active site guanine in hairpin 127:5026–5027. ribozyme catalysis probed by exogenous nucleobase rescue. J Mol Biol 340:233–251. 14. Lipfert J, Ouellet J, Norman DG, Doniach S, Lilley DMJ (2008) The complete VS 7. Kuzmin YI, Da Costa CP, Cottrell JW, Fedor MJ (2005) Role of an active site adenine in ribozyme in solution studied by small-angle X-ray scattering. Structure 16:1357–1367. hairpin ribozyme catalysis. J Mol Biol 349:989–1010. 15. Smith MD, Mehdizadeh R, Olive JE, Collins RA (2008) The ionic environment deter- 8. Wilson TJ, et al. (2006) Nucleobase catalysis in the hairpin ribozyme. RNA 12:980–987. mines ribozyme cleavage rate by modulation of nucleobase pKa. RNA 14:1942–1849.

Wilson et al. PNAS ∣ June 29, 2010 ∣ vol. 107 ∣ no. 26 ∣ 11755 Downloaded by guest on September 24, 2021 16. Ferré-d’Amaré AR, Zhou K, Doudna JA (1998) Crystal structure of a hepatitis delta virus 25. Cai ZP, Tinoco I (1996) Solution structure of loop A from the hairpin ribozyme from ribozyme. Nature 395:567–574. tobacco ringspot virus satellite. Biochemistry 35:6026–6036. 17. Perrotta AT, Shih I, Been MD (1999) Imidazole rescue of a cytosine mutation in a 26. Butcher SE, Allain FH, Feigon J (1999) Solution structure of the loop B domain from the Science – self-cleaving ribozyme. 286:123 126. hairpin ribozyme. Nat Struct Biol 6:212–216. 18. Ke A, Zhou K, Ding F, Cate JH, Doudna JA (2004) A conformational switch controls 27. Wyckoff HW, et al. (1970) The three-dimensional structure of ribonuclease-S. hepatitis delta virus ribozyme catalysis. Nature 429:201–205. Interpretation of an electron density map at a nominal resolution of 2 Å. J Biol Chem 19. Das SR, Piccirilli JA (2005) General acid catalysis by the hepatitis delta virus ribozyme. – Nat Chem Biol 1:45–52. 245:305 328. 20. Smith MD, Collins RA (2007) Evidence for proton transfer in the rate-limiting step of a 28. Wodak SY, Liu MY, Wyckoff HW (1977) The structure of cytidilyl (2′,5′) adenosine when fast-cleaving Varkud satellite ribozyme. Proc Natl Acad Sci USA 104:5818–5823. bound to pancreatic ribonuclease S. J Mol Biol 116:855–875. 21. Lebruska LL, Kuzmine I, Fedor MJ (2002) Rescue of an abasic hairpin ribozyme by 29. delCardayré SB, Raines RT (1994) Structural determinants of enzymatic processivity. cationic nucleobases. Evidence for a novel mechanism of RNA catalysis. Chem Biol Biochemistry 33:6031–6037. – 9:465 473. 30. Raines RT (1998) Ribonuclease A. Chem Rev 98:1045–1066. 22. Hoffmann B, et al. (2003) NMR structure of the active conformation of the Varkud 31. Zamel R, et al. (2004) Exceptionally fast self-cleavage by a Neurospora Varkud satellite satellite ribozyme cleavage site. Proc Natl Acad Sci USA 100:7003–7008. ribozyme. Proc Natl Acad Sci USA 101:1467–1472. 23. Murchie AIH, Thomson JB, Walter F, Lilley DMJ (1998) Folding of the hairpin ribozyme Chem in its natural conformation achieves close physical proximity of the loops Molec. 32. McCarthy TJ, et al. (2005) Ligand requirements for glmS ribozyme self-cleavage. Biol – Cell 1:873–881. 12:1221 1226. 24. Rupert PB, Ferré-D’Amaré AR (2001) Crystal structure of a hairpin ribozyme-inhibitor 33. Klein DJ, Ferré-D’Amaré AR (2006) Structural basis of glmS ribozyme activation by complex with implications for catalysis. Nature 410:780–786. glucosamine-6-phosphate. Science 313:1752–1756.

11756 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1004255107 Wilson et al. Downloaded by guest on September 24, 2021