| Nucleic Acids Research | Pages |
Hairpin ribozyme cleavage catalyzed by aminoglycoside antibiotics and the polyamine spermine in the absence of metal ions
Introduction
Materials And Methods
Determination of ribozyme kinetic parameters
Inhibition of ribozyme cleavage by antibiotics
pH profiles
Competition cleavage assays
Filter binding
Results
The effect of aminoglycoside antibiotics on hairpin ribozyme cleavage
Cleavage of the hairpin ribozyme by polyamines
Discussion
Cleavage and inhibition of the hairpin ribozyme by aminoglycoside antibiotics
The mechanism of hairpin ribozyme cleavage
Implications for therapeutic use of hairpin ribozymes
Acknowledgements
References
Hairpin ribozyme cleavage catalyzed by aminoglycoside antibiotics and the polyamine spermine in the absence of metal ions
ABSTRACT
INTRODUCTION
A number of small self-cleaving RNAs undergo trans-esterification reactions which are chemically identical in that the two cleavage products contain a 5[prime]-hydroxyl group and 2[prime],3[prime]-cyclic phosphate, respectively. Yet the structures and mechanistic pathways used to achieve such cleavages appear quite dissimilar (for recent reviews see 1,2). The hairpin ribozyme occurs in the negative strand of the satellite RNA of tobacco ringspot virus (TRSV), as well as in two other closely related nepovirus satellite RNAs, and is the second smallest, naturally occurring, self-cleaving RNA (reviewed in 3-6).
The minimum catalytic motif consists of two helical domains (A and B) each containing an internal loop flanked by two regions of Watson-Crick double helix. The hairpin is commonly studied in trans-cleavage form, with a substrate strand and a ribozyme component consisting of one or two strands (Fig.
Much recent effort has gone into understanding the steps involved in the cleavage activity. The association of substrate strand with the substrate binding strand to form domain Aoccurs extremely rapidly, if the substrate sequence is modifiedto avoid self-complementarity, with values of kon in the region of 2 × 108/M/min, one of the fastest association rates of any known ribozyme (7-9). The complex is quite stable with a koff of <0.01/min. The kinetics of hairpin cleavage are biphasic showing a fast phase with a kcat of ~0.2/min, some 5-fold lower than efficient hammerhead ribozyme reactions (7-9). The subsequent slow phase has been shown to be due to an inactive conformation where the two domains A and B are stacked in an extended configuration (8,10). The conformers are non-exchangeable and the inactive complex must dissociate in order to rebind to form an active ribozyme. The active conformation is achieved through an initial docking of the two domains A and B in an antiparallel fashion. The docking process has been monitored by attachment of fluorescent dyes to each domain and by measurement of fluorescence resonance energy transfer (FRET) upon folding induced by metal ions such as calcium, magnesium and strontium, all of which support cleavage. Docking was found to be required for both cleavage and ligation but in the case of cleavage, docking is a fast process (~0.5/min) that is not rate-limiting (11).
In the context of the TRSV satellite RNA, the hairpin motif is part of a four-way helical junction, the folding of which in the presence of magnesium or other group IIA metal ions in the micromolar range also favours the interaction of domains A and B (12). Further metal ions are needed for cleavage activity which is achieved in the millimolar range. However, some metal ions which allow correct folding, such as manganese, are less able to support cleavage (13). This dual role of metal ion in the natural hairpin context is paralleled by earlier studies of the ionic requirements for cleavage of the hinged two-domain hairpin (14).
In the case of the well-documented hammerhead ribozyme, in addition to playing an important structural role, metal ions have been shown to participate directly in the chemical cleavage step. The precise role is controversial, but proposals for metal ion involvement have included interaction with the attacking 2[prime]-hydroxyl group, direct chelation to the pro-R oxygen atom of the cleaved phosphate in the transition state, and/or interaction with the leaving 5[prime]-oxygen atom (discussed in 15,16). In contrast, recent studies of the metal ion requirements for hairpin ribozyme cleavage have failed to provide any evidence for direct participation of metal ions in the cleavage mechanism, either by 2[prime]-hydroxyl activation or by chelation to a non-bridging oxygen atom in the transition state of the cleaved phosphate. Furthermore, cobalt (III) hexammine, which cannot form inner sphere co-ordination complexes, was found to be particularly efficient at supporting hairpin ribozyme cleavage (17-19). Although participation of metal ion as an outer sphere complex cannot be ruled out, a more likely possibility is that metal ions play a structural role in a step just prior to chemical cleavage, perhaps a metal-dependent conformational change.
The polyamine spermidine was the first non-metallic ion shown to support hairpin cleavage, but at 2 mM the cleavage rate is extremely slow, some 103 lower than for 10 mM magnesium ions (14). More interestingly, low concentrations of spermidine were found to be able to stimulate the magnesium-dependent cleavage rate somewhat (14). To gain further insight into the structural and unusual ionic requirements of hairpin ribozyme cleavage, we decided to investigate a wider range of non-metallic polycations, such as aminoglycoside antibiotics and other polyamines. We now report that aminoglycoside antibiotics cause inhibition of hairpin cleavage when added to a magnesium ion-catalyzed reaction. Neomycin B gives moderate inhibition, but only one, 5-epi-sisomicin, inhibits strongly. More significantly, and in contrast to other self-cleaving domains, aminoglycoside antibiotics stimulate hairpin ribozyme cleavage in the absence of metal ions, with the best (neomycin B, apramycin and kanamycin B) being only 13-20-fold lower in rate at 10 mM than for the same concentration of magnesium ions. We have further found that the tetra-amine spermine supports a very efficient hairpin cleavage reaction, almost as fast as magnesium ions at the same concentration. In addition, certain concentration combinations of spermine and magnesium ions lead to substantial stimulation of hairpin ribozyme cleavage and more efficiently than the triamine, spermidine. The results have important implications for potential therapeutic use of the hairpin ribozyme in mammalian cells.
MATERIALS AND METHODS
Oligoribonucleotides were synthesized by 1 µmol solid-phase synthesis on controlled pore glass and purified by anion exchange chromatography on a NucleoPac PA-100 column (Dionex, Surrey, UK) as described previously (20,21). 2[prime]-O-methyladenosine phosphoramidite was obtained from Glen Research (via Cambio, Cambridge, UK). Desalting was achieved via extensive dialysis against water. Oligoribonucleotides were 5[prime]-end-labeled with [[gamma]-32P]ATP and T4 polynucleotide kinase (22). Purities were checked by electrophoresis on a 20% denaturing polyacrylamide gel (PAGE) and by MALDI-TOF mass-spectrometry (21). Aminoglycoside antibiotics were purchased from either Sigma or Fluka except for neamine which was obtained from Affiniti (Exeter, UK) and 5-epi-sisomicin which was kindly provided by Julian Davies. Neomycin (Sigma) was a mixture of neomycin B (>90%) and neomycin C, with the difference being the stereochemistry of the carbon at position 1 of the B ring. Kanamycin A (Fluka) contained minor components B and C (<5%). Gentamicin (Fluka) was a mixture of two components as illustrated (Fig.
Determination of ribozyme kinetic parameters
Kinetic parameters of cleavage of metal-, antibiotic- and polyamine-dependent reactions were determined under single turnover conditions (20). Separate solutions of ribozyme (equimolar mixture of RzA and RzB strands, 20-200 nM, 90 µl) in Tris-HCl (pH 7.5) and [gamma]-32P-labeled substrate RNAs (10 nM, 10 µl) in water were each incubated at 70°C for 1 min and then cooled to room temperature over 15 min. The cofactor solution was added to its required concentration in the ribozyme solution and then both were incubated at 37°C for 15 min. Reactions were initiated by mixing the ribozyme and substrate solutions to give a final volume of 100 µl of 40 mM Tris-HCl in the case of metal ion reactions or 100 mM in the case of other co-factors. Non-metal catalyzed reactions contained 1 mM EDTA and 1 mM EGTA which were added to the ribozyme strand prior to the 70°C incubation. No differences in rates were observed in the concentration range of 0.1-2 mM for these chelators but 10-fold rate reductions were observed at 50 mM EDTA, most likely due to the inhibition by sodium ions (data not shown) (14).
Aliquots (10 µl) were removed at six suitable time intervals and the reactions quenched by addition to 10 µl of urea stop mix(7 M urea, 50 mM EDTA, 0.04% w/v xylene cyanol, 0.04% w/v bromophenol blue). Samples were loaded onto a 20% denaturing polyacrylamide gel and subjected to electrophoresis at 12 W for 80 min. The resultant gels were dried and scanned using a PhosphorImager (Molecular Dynamics, Buckinghamshire, UK), and the data processed using the programme Image Quant (Molecular Dynamics) and quantitated by use of the Geltrak programme as described previously (20). The initial velocities of the reactions (kobs) at different ribozyme concentrations were determined from a plot of product formation against time, usually up to 30% cleavage. Data from at least three independent experiments were obtained and usually fell within 20% of the mean. kobs values were plotted against ribozyme concentration/kobs using Eadie-Hofstee plots as described previously (20). Multiple turnover kinetic parameters were carried out using substrate concentrations of 50-1000 nM and ribozyme concentrations of 1-80 nM as described previously (23,24). In the cases of aminoglycosides, it was necessary to keep the ribozyme concentration constant or within a narrow range and to vary substrate concentration.
Inhibition of ribozyme cleavage by antibiotics
The inhibition constants (Ki) of hairpin ribozyme cleavage for the antibiotics were obtained by varying the concentration of antibiotic from 0.1 µM to 100 mM and by measuring the kobs at 10 mM MgCl2 under conditions described for the single turnover conditions. Ribozyme (RzA and RzB) concentration was 50 nM and 32P-labeled substrate (10 nM) was used to initiate the reaction. Control cleavage reactions with antibiotic alone and with magnesium ions alone were also carried out in parallel. For neomycin B inhibition, the fraction of inhibition at each concentration of antibiotic (1 - kneo/kcon) was calculated from the kobs for the antibiotic reaction in the presence of 10 mM magnesium ions (kneo) and the kobs for the control reaction of 10 mM magnesium ions alone (kcon) (25). The fraction of inhibition was plotted against neomycin B concentration (Fig.
In the delayed inhibition reactions (Fig.
pH profiles
The pH dependences of the kobs of hairpin cleavage reactions (Fig.
Competition cleavage assays
Ribozyme cleavage rates (kobs) for competitive cleavage reactions (Figs
Filter binding
Filter binding assays utilizing a 96-well 0.45 µM `MultiScreen'-HA Millipore mixed cellulose ester filtration plate were carried out essentially as described previously for protein-RNA filter binding (21,26) and similarly to the method reported for measurement of neomycin B binding to RRE RNA (27). To prevent substrate cleavage during the binding assay, an uncleavable 2[prime]-O-methyl analogue at the A-1 position of the substrate strand was used. Hairpin ribozyme (25 nM) with one strand 5[prime]-32P radio-labelled (200 000 c.p.m.) was annealed in 40 mM Tris-HCl (pH 7.5) by incubation of equimolar amounts of the three strands at 70°C for 1 min and cooled slowly to 4°C. Antibiotic was added to give concentrations of 50 nM to 1200 µM and a volume of 250 µl.
Parallel experiments were also carried out in the presence of 10 mM MgCl2. Binding reactions were incubated on ice for 60 min in individual wells of the 96-well plate which was presoaked and washed in ice-cold 40 mM buffer. After 60 min the plate was vacuum-filtered, and the wells washed once with ice-cold 40 mM buffer (200 µl), dried and counted by liquid scintillation as described previously (21). Apparent Kd values were determined by plotting the percentage of labeled ribozyme bound to the filter against the antibiotic concentration and calculation of the concentration at which binding was half maximal.
In control experiments we found that ribozyme B and substrate strands alone did not bind aminoglycosides significantly, but ribozyme A strand alone was retained on the filter in the presence of antibiotics. Therefore, to check that binding of the complete hairpin was measured, three parallel experiments with neomycin B were carried out where one strand in turn was 32P-labelled. Identical filter retention results were found in each case. In the tRNA competition reactions, 250 nM and 2.5 µM concentrations of yeast tRNA (Sigma) were also present.
RESULTS
The effect of aminoglycoside antibiotics on hairpin ribozyme cleavage
Several years ago, aminoglycoside antibiotics were shown to inhibit the self-splicing activity of the group I intron (28,29). More recently, it was shown that a range of aminoglycosides inhibit the cleavage of the hepatatis delta virus ribozyme (30) and the hammerhead ribozyme (25,31), with neomycin B being particularly effective. We therefore investigated the effect of increasing concentrations of neomycin B on the magnesium ion-dependent hairpin ribozyme cleavage reaction. For these experiments we utilized our well-characterized three-stranded hairpin ribozyme (Fig.
Figure 1. Secondary structure of a three-stranded model hairpin ribozyme showing the two domains that are required to dock in order to effect cleavage. The cleavage site in domain A is shown by an arrow. The residue A-1 adjacent to the cleavage site is highlighted in green. 2[prime]-O-methyl A subsitution of this residue allows efficient docking but the ribozyme cannot be cleaved (11). The proposed ribose zipper which involves the 2[prime]-hydroxyl groups of residues A10 and G11 in domain A, and A24 and C25 in domain B that may facilitate docking are shown in magenta (20). See Corrigenda Figure 2. Concentration dependence of neomycin B inhibition of the magnesium-induced cleavage of the hairpin ribozyme. The data was fitted to a hyperbolic, bimolecular binding equation to give Ki = 190 ± 29 µM. kneo = kobs of the 10 mM magnesium-induced reaction at the particular neomycin concentration. kcon = kobs of the 10 mM magnesium-alone catalyzed reaction. We screened a range of other antibiotics in the inhibition of magnesium-dependent hairpin ribozyme cleavage. These included other members of the neomycin family (Fig. Figure 3. (a) Structure of 4,5-disubstituted deoxystreptamine class of aminoglycoside antibiotics. Neomycin B, R = NH2; paromomycin, R = OH; ribostamycin contains all rings except IV; neamine lacks rings III and IV; lividomycin A has a mannose sugar attached to O4[prime][prime][prime] of ring IV; butirosin has l-amino-[alpha]-hydroxybutyric acid at position 1 of ring II and has no ring IV. The pKa values in free solution are shown for neomycin B (53). (b) Table showing the concentrations (µM) of antibiotic required to inhibit the 10 mM Mg2+-catalyzed cleavage (Ki), rate parameters for single turnover cleavage for the co-factor alone (Mg2+ or aminoglycoside), and the apparent dissociation constant (Kd) of the co-factor on the 2[prime]-O-methyl A-1 substituted hairpin ribozyme as determined by filter binding, together with the percentage filter retention. a, apparent Kd in the presence of 10 mM Mg2+;. nd, not determined. Figure 4. (a) Structure of 4,6-disubstituted deoxystreptamine aminoglycosides. Tobramycin, R1 = NH2, R2 = OH, R3 = H, R4 = OH, R5 = NH2; kanamycin A, R1 = NH2, R2/R3 = OH, R4 = OH, R5 = NH2; kanamycin B, R1 = NH2, R2/R3 = OH, R4 = NH2, R5 = NH2; dibekacin, R1 = NH2, R2 = H, R3 = H, R4 = OH, R5 = NH2; amikacin, R1 = NH2, R2/R3 = OH, R4 = OH, R5 = NHCOCH(OH)CH2CH2NH2. The pKa values are shown for tobramycin (54). (b) Table as described in Figure 3b. Figure 5. (a) Structures of gentamicin, R2 = NH2, R1= CH3; or R2 = NHCH3, R1 = H, R3 = H, R4 = H, and gentecin G418, R1 = OH, R2 = CH3, R3 = OH, R4 = OH. (b) Structure of apramycin. The two amino groups marked * have indistinguishable pKa values (55). (c) Structures of sisomicin and 5-epi-sisomicin (epimerization of 5-hydroxyl group on ring II). (d) Table as described in Figure 3b. Figure 6. pH dependence of aminoglycoside-induced cleavage. Graph showing the variation in the kobs values over a range of pH values for aminoglycosides at 5 mM concentration in 100 mM buffer, or for buffer alone (see Materials and Methods). Inset, the pH dependence of hairpin ribozyme cleavage for 5 mM Mg2+. The unusual ionic dependence of the hairpin ribozyme prompted us to test whether neomycin B could carry out the hairpin ribozyme cleavage reaction in the absence of metal ions. We found that at 37°C, pH 7.5 and in the presence of chelators EDTA and EGTA, 10 mM neomycin B catalyzed specific cleavage of the hairpin with products which migrated identically to those seen in the magnesium-dependent reaction. No other cleavage products were seen (data not shown). The k[prime]cat was only 18-fold slower than for the same concentration of magnesium ions alone (Fig. All other members of the neomycin family showed significantly poorer k[prime]cat values, although butirosin and neamine showed K[prime]M values which were 10-fold lower (Fig. In order to obtain an estimate of the relative affinities ofthe various antibiotics for the hairpin ribozyme, we measured their apparent dissociation constants by determination of the concentrations required for 50% retention of the ribozyme-drug complex on an ice-cold mixed cellulose ester filter (Figs Figure 7. Delayed inhibition of magnesium-dependent cleavage by antibiotics. Graph showing fraction of cleavage of substrate as a function of time of reaction. Either 10 mM neomycin B or 10 mM tobramycin was added at 90 s to a 10 mM magnesium-initiated hairpin ribozyme cleavage reaction. For comparison, plots of the 10 mM magnesium alone, neomycin B alone and tobramycin alone hairpin ribozyme cleavage reactions are also shown. We next compared the pH profiles for antibiotic-induced and magnesium-dependent cleavages (Fig. A number of recent studies have suggested that, in the cases of the hammerhead and hepatitis delta virus ribozymes and the self-splicing group I intron, aminoglycoside antibiotics achieve inhibition of ribozyme cleavage by replacement of structurally important magnesium ions by protonated amino groups (25,30,32,33). To address whether this is also the case for hairpin ribozyme cleavage, we first carried out a delayed addition reaction, where 10 mM magnesium ions was used to initiate ribozyme cleavage and then 10 mM of the antibiotic was added after 90 s. For both neomycin B and tobramycin addition, a reduction in cleavage rate occurred very rapidly as compared with the reaction with magnesium ions alone (Fig. Secondly, the concentration dependence of the neomycin B alone cleavage reaction was studied and this showed a biphasic character with a hyperbolic curve up to 25 mM neomycin B followed by an approximately linear dependence thereafter and no evidence of saturation (Fig. Figure 8. Rate-dependence of cleavage on neomycin B concentration. Graph showing kobs of cleavage as a function of neomycin B concentration for the antibiotic alone reaction and also antibiotic in the presence of 1 mM Mg2+. The kobs of the 1 mM Mg2+ alone cleavage reaction is marked (0.0143/min). Inset, the effect of low concentrations of neomycin B (<1 mM) on cleavage for the antibiotic in the presence of 1 mM Mg2+. Finally, we determined the effect of 1 mM neomycin B on the magnesium dependence of hairpin ribozyme cleavage at three pH values (Fig. Figure 9. pH dependence of neomycin B inhibition of the magnesium-dependent cleavage reaction. Graph showing kobs of cleavage as a function of Mg2+ concentration for three pH conditions (pH 6, 7.5 and 8.5) in the presence or absence of 1 mM neomycin B. It has been reported previously that a very slow, but specific cleavage of the hairpin ribozyme could be effected by the triamine spermidine in the presence of EDTA and EGTA (kobs 0.008/min at 50 mM spermidine) (14). We found that under single turnover conditions 10 mM spermidine specifically cleaved the hairpin ribozyme with k[prime]cat of 0.003/min, ~45-fold slower than the k[prime]cat for magnesium ions (Fig. The shorter polyamine diethylene triamine catalyzed only an extremely slow cleavage reaction, some 3000-fold poorer than spermine (data not shown). Similarly, diaminoalkanes of C2-C9 chain length at 50 mM concentrations showed very low rates of hairpin cleavage, the best, 1,6-diaminohexane, still having a 1000-fold lower rate than spermine under single turnover conditions. 1,12-Diaminododecane, which is the same length as spermine but has only two amino groups, was poorly water soluble but in the presence of 2.5% DMSO, which does not affect the magnesium reaction significantly, it showed a rate still some 500-fold poorer than spermine (data not shown). Figure 10. (a) Structures of the tetramine, spermine, and the triamine, spermidine, and their pKa values (44). (b) Table showing single and multiple turnover kinetic parameters for the polyamine co-factors and in combination with Mg2+. Also for spermine and spermidine the apparent dissociation constants (Kd ) and percentage filter retention are shown. The pH dependence of the spermine-induced cleavage showed very little change in rate between pH 5 and 8 (data not shown). There was a slight maximum at pH 7 just 2-fold above the rates at pH 5 and 8, which were almost identical. The rate dropped to a very low value at pH 9. More significantly, there was a substantial stimulation of the initial rate of the magnesium-dependent reaction by addition of spermine. Under single turnover conditions, maximum stimulation (3.5-fold) was obtained for spermine at 500 µM when the magnesium concentration was ~5 mM to give a kobs of 0.35/min (Fig. Figure 11. Graph showing the rate (kobs) of hairpin cleavage as a function of magnesium concentration in the presence of 500 µM spermine or spermidine. The magnesium concentration dependence for the Mg2+ alone reaction is also shown. A concentration of 500 µM spermine or spermidine was found to be optimal for the enhancement of the magnesium-induced cleavage, since both 250 and 750 µM concentrations gave rise to less stimulation (data not shown). Figure 12. Graph showing the rate (kobs) of hairpin cleavage as a function of either spermine or spermidine concentration in the presence of 500 µM Mg2+. The spermine and spermidine alone concentration dependences of cleavage are also shown. A concentration of 500 µM magnesium ions was found to be optimal for the enhancement of the polyamine-induced cleavage, since 250 and 750 µM were found to be less effective (data not shown). We have shown that specific cleavage of the hairpin ribozyme can be effected by a range of aminoglycoside antibiotics in the absence of metal ions. The rates of cleavage at pH 7.5 under single turnover conditions are at best 13-18-fold less than that for magnesium ions, but at pH 5.5 the rates of cleavage for neomycin B andkanamycin B are more comparable with that for magnesium ions. The antibiotic-induced cleavage reaction is identical in specificity to the magnesium-dependent reaction and occurs much faster than internucleotide hydrolysis. For example, the pseudo-first order rate constant for hydrolysis of ApA by 300 mM neomycin B is1.6 × 10-5/min at 50°C and pH 8 (34). Furthermore, ribozyme cleavage is not a general property of aminoglycoside antibiotics, since we found no cleavage of the hammerhead ribozyme up to 100 mM neomycin B (data not shown). A detailed aminoglycoside structure-activity relationship is beyond the scope of this discussion, but we can draw a number of general conclusions. First, the most important feature of aminoglycoside antibiotics for hairpin ribozyme cleavage seems to be the presence of at least four protonated amino groups. For example, at pH 7.5 in the neomycin family, those antibiotics lacking rings III and IV are less active, but at lower pH, ribostamycin, which lacks ring IV, becomes almost as active as neomycin B. All of the antibiotics tested in the neomycin and tobramycin families are more active at lower pH, suggesting that protonation of the amino group at position 3 of ring II, which has a nominally low pKa value, is important for activity. A second determinant of activity is the presence and orientation of certain hydroxyl groups; for example, as seen in the tobramycin family, where the removal of 3[prime]- and 4[prime]-hydroxyl groups results in loss of activity and in the substantial difference in activity obtained with sisomicin and its synthetic derivative, 5-epi-sisomicin. Structurally important hydroxyl contacts have been observed, for example, in interaction of aminoglycoside antibiotics with the A site of 16S ribosomal RNA (35). Apramycin is unusual in having a pH optimum at pH 8. This was unexpected because at this pH several of the amino groups would be expected to be unprotonated or less protonated. The different ring structure, arrangement of amino groups and the narrower spread of pKa values (6.6-8.2) suggest that apramycin does not follow the same mode of action throughout the pH range tested. For example, at lower pH there may perhaps be a different conformation or binding mode that stabilizes an inactive form of the hairpin. Under conditions when magnesium ions are present, aminoglycoside antibiotics inhibit the hairpin ribozyme cleavage. The degree of inhibition is correlated with the apparent binding constant of the antibiotic; in general the tighter the binding, the greater the degree of inhibition. Thus, 5-epi-sisomicin is by far the most potent inhibitor, but neomycin B also has significant inhibition properties. All others tested were much poorer inhibitors. Neomycin B is the only antibiotic tested which is both a good inhibitor of magnesium-dependent cleavage and is also reasonably active in the metal-free cleavage reaction. Competition experiments were therefore particularly revealing. For example, we have provided several lines of evidence that the inhibition by neomycin B occurs by replacement of metal ions by the amino groups of the antibiotic, similarly to the hammerhead, HDV and group I intron ribozymes (25,30,32,33). From curve fitting of the neomycin B inhibition and from preliminary isothermal calorimetry experiments (D.J.Earnshaw and M.J.Gait, unpublished results), the stoichiometry of neomycin B binding to the hairpin appears to be 1:1. In the competition experiment (Fig. Early studies of the ionic requirements for hairpin ribozyme cleavage showed that at least two types of ion were needed for efficient hairpin ribozyme cleavage (14). One of these types seems to play a role in the folding and docking of the two domains and can be satisfied by a range of metal ions (magnesium, calcium, strontium, manganese) (11,13). In contrast, it has been shown by FRET studies of the hinged hairpin that 10 mM spermidine in the absence of metal ion does not promote efficient docking, but instead is able to boost the docking rate of 2 mM magnesium or manganese ions (11). It was suggested that spermidine stabilizes the metal-dependent docking. In the case of the natural hairpin junction, spermidine does promote docking, albeit at much higher concentrations than metal ions (13). Our kinetic data show that the longer polyamine spermine must be able to fulfill the docking requirement very efficiently in the hinged hairpin. It may be noted that spermine has been shown to promote folding of the hammerhead ribozyme and to allow ions such as cadmium to catalyze efficient ribozyme cleavage, which are unable to do so in the absence of spermine (36). We have no information at this stage concerning the effect of aminoglycoside antibiotics on hairpin docking, and therefore as to whether the lower rates of the antibiotic-induced cleavage reaction compared with magnesium ions are due to impaired docking or to a suboptimal ability to replace catalytically important magnesium ions in the transition state, or to both. Once docking has taken place, a second type of ion is required for attainment of a catalytically competent hairpin configuration. Magnesium, calcium and strontium ions are all effective in this role (14). Use of manganese ions has been reported to result in 10-fold (14) or 2.6-fold (18) lower cleavage rates for the hinged hairpin and to be catalytically inactive in the case of the natural junction hairpin (13). Cobalt hexammine is quantitatively the most effective ion in both docking (11,13) and in hairpin cleavage (17-19). We have also found that under single turnover conditions at pH 7.5 at 37°C, 0.25 mM Co3+(NH3)6 or Ru3+(NH3)6 promote hairpin ribozyme cleavage with k[prime]cat values 2.9- and 2.3-fold higher, respectively, than are achieved with 10 mM concentrations of Mg2+, but Pt2+(NH3)4 was completely inactive (D.J.Earnshaw and M.J.Gait, unpublished results). In the current study we have shown that the polyamine spermine can carry out this second ionic role very efficiently, but the effectiveness of spermidine in this role is less clear. Clearly, both polyamines can synergize with 5 mM magnesium ions to boost cleavage rates. But at low magnesium concentration, where docking is effective but cleavage is slow, we found that spermidine is much less effective than spermine in enhancing the cleavage rate. Under such conditions, the polyamine may be required to play predominantly the role of attainment of the catalytically active structure, and for this spermidine appears sub-optimal. It is interesting to note that addition of a low concentration of either spermine or spermidine dramatically enhanced the rate of the magnesium-dependent cis cleavage of the Neurospora VS ribozyme and higher concentrations facilitated the trans cleavage reaction (37). Furthermore, spermidine was shown to be essential to allow self-cleavage of the coconut cadang cadang viroid but magnesium ions could not promote such cleavage (38). Thus, it seems likely that there is an important general effect of polyamines in facilitating interstrand interactions in ribozymes, which may be due partly to the reduction in charge repulsion between nucleic acid strands expected upon polyamine binding, as observed for tRNA (39). However, the particular success of spermine in promoting hairpin ribozyme cleavage suggests that there is in addition an important structural component both to hairpin interdomain interaction and subsequent configuration for cleavage that is satisfied by the particular length and arrangement of the four positive charges on the spermine backbone. If fully extended, spermine would contain regularly spaced amino groups spanning a distance of 15 Å, which is very similar to the spacing and distance spanning the amino groups in rings I and II of aminoglycoside antibiotics (32). Curve fitting of preliminary isothermal calorimetry experiments at concentrations of <120 µM suggest that one molecule of spermine may bind to the hinged hairpin (D.J.Earnshaw and M.J.Gait, unpublished results). This is in line with FRET data showing a 1:1 stoichiometry of spermidine binding to the natural hairpin junction (13). Our data support the hypothesis that the positively charged amino groups on spermine replace structurally important metal ions, one or more of which may be required to be located between the two domains to ligand functional groups or to form ionic interactions, as part of a network of interdomain contacts that may include, for example, the recently proposed `ribose zipper' between pairs of neighbouring hydroxyl groups on the two domains (Fig. Spermine, and to a lesser extent aminoglycoside antibiotics, must also be able to mimic the second role of metal ion in attaining a catalytically competent configuration. This role also appears to be structural, since no direct catalytic role for metal ion has yet been shown and indeed as we have shown, metal ion can be dispensed with completely. Furthermore, whilst this manuscript was in preparation we have learnt that 4 M ammonium ions can also induce efficient hairpin ribozyme cleavage (40). Our pH titration data confirm previous studies showing the lack of a log linear pH dependence of the magnesium-dependent reaction, such that the cleavage rate cannot be dependent on the concentration of MgOH- (18). The pH dependence is approximately linear in the range 5-8 (Fig. In contrast, the pH profiles of the aminoglycoside- and polyamine-dependent cleavages would be expected to be dominated by the pKa values of the various amino groups on the ligand. For spermine, the pKa values of all four amino groups are high (8.6-11.2) (44). The flat pH profile between pH 5 and 8 and the very low activity at pH 9 indicates that there is a requirement for all four positive charges, but it seems very unlikely that any of the amino groups can directly participate either as a general base or a general acid. The pH profiles for the aminoglycoside antibiotic-induced cleavages (Fig. Overall we favour a structural role for the ligand in hairpin ribozyme cleavage where the metal ions, or positively charged amino groups, are involved in a rate-determining structural rearrangement that follows docking and precedes chemical cleavage. Our previous functional group studies of the requirements for hairpin cleavage in loop B (23,24) indicated that there may be a secondary structure in the transition state different from that proposed in the ground state by chemical probing, molecular modelling and other structural data (20,45,46). Another precedent comes from the finding that neomycin B, spermine and cobalt hexammine, but not platinum tetrammine, share common structural motifs which are important in another structural rearrangement, the conversion of B- into A-DNA (47). There have been a number of studies where the hairpin ribozyme has been engineered for vector delivery into mammalian cells (reviewed in 4,5,48). Further, intracellular cleavage has been demonstrated by a hairpin ribozyme construct (49). It has been estimated that the concentration of free magnesium within cells is in the 1 mM range (49 and references therein). Whereas at 5 mM magnesium ions, both spermidine and spermine boost cleavage rates, at lower magnesium ion concentration, we have found that spermine is much more effective. Polyamines are found in all cells and are essential for eukaryotic proliferation. Of the well known polyamines, spermine is the most abundant polyamine in eukaryotes, for example in hepatocytes (50) and in human colon tumour cells (51), and occurs in amounts up to twice as much as spermidine. Although the quantities of free intracellular polyamines are unknown, it seems likely that hairpin ribozyme cleavage inside cells could be enhanced significantly by spermine. This may help to explain why such promising preliminary results have already been obtained in cleavage by the hairpin ribozyme inside mammalian cells and may lead to improved prospects for the use of the hairpin ribozyme as a therapeutic agent. Furthermore, the ability of aminoglycosides and polyamines to promote hairpin ribozyme cleavage adds strength to suggestions that low molecular weight compounds may have played roles as modulators of RNA-catalyzed reactions during evolution (52). We thank Mohinder Singh for valuable technical assistance, Richard Grenfell for synthesis of oligoribonucleotides, Julian Davies for supply of various aminoglycoside antibiotics and Chris Johnson for instruction in isothermal calorimetry. We are also grateful for helpful discussions with Mark Farrow, Tom Barlow and Eric Westhof.
Cleavage of the hairpin ribozyme by polyamines
DISCUSSION
Cleavage and inhibition of the hairpin ribozyme by aminoglycoside antibiotics
The mechanism of hairpin ribozyme cleavage
Implications for therapeutic use of hairpin ribozymes
ACKNOWLEDGEMENTS
REFERENCES
This article has been cited by other articles:
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 3 Dec 1998
Copyright©Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
M. Roychowdhury-Saha and D. H. Burke
Distinct reaction pathway promoted by non-divalent-metal cations in a tertiary stabilized hammerhead ribozyme
RNA,
June 1, 2007;
13(6):
841 - 848.
[Abstract]
[Full Text]
[PDF]
R. Ting, J. M. Thomas, L. Lermer, and D. M. Perrin
Substrate specificity and kinetic framework of a DNAzyme with an expanded chemical repertoire: a putative RNaseA mimic that catalyzes RNA hydrolysis independent of a divalent metal cation
Nucleic Acids Res.,
December 29, 2004;
32(22):
6660 - 6672.
[Abstract]
[Full Text]
[PDF]
A. Inoue, Y. Takagi, and K. Taira
Importance in catalysis of a magnesium ion with very low affinity for a hammerhead ribozyme
Nucleic Acids Res.,
August 9, 2004;
32(14):
4217 - 4223.
[Abstract]
[Full Text]
[PDF]
S. H. Najafi-Shoushtari, G. Mayer, and M. Famulok
Sensing complex regulatory networks by conformationally controlled hairpin ribozymes
Nucleic Acids Res.,
June 15, 2004;
32(10):
3212 - 3219.
[Abstract]
[Full Text]
[PDF]
A. V. Vlassov, B. H. Johnston, L. F. Landweber, and S. A. Kazakov
Ligation activity of fragmented ribozymes in frozen solution: implications for the RNA world
Nucleic Acids Res.,
May 25, 2004;
32(9):
2966 - 2974.
[Abstract]
[Full Text]
[PDF]
K. F. Blount and Y. Tor
Using pyrene-labeled HIV-1 TAR to measure RNA-small molecule binding
Nucleic Acids Res.,
October 1, 2003;
31(19):
5490 - 5500.
[Abstract]
[Full Text]
[PDF]
S. Corvaisier, V. Bordeau, and B. Felden
Inhibition of Transfer Messenger RNA Aminoacylation and trans-Translation by Aminoglycoside Antibiotics
J. Biol. Chem.,
April 18, 2003;
278(17):
14788 - 14797.
[Abstract]
[Full Text]
[PDF]
P. Kumar, G. Dhawan, R. Chandra, and K. C. Gupta
Polyamine-assisted rapid and clean cleavage of oligonucleotides from cis-diol bearing universal support
Nucleic Acids Res.,
December 1, 2002;
30(23):
e130 - e130.
[Abstract]
[Full Text]
[PDF]
J.-M. Zhou, D.-M. Zhou, Y. Takagi, Y. Kasai, A. Inoue, T. Baba, and K. Taira
Existence of efficient divalent metal ion-catalyzed and inefficient divalent metal ion-independent channels in reactions catalyzed by a hammerhead ribozyme
Nucleic Acids Res.,
June 1, 2002;
30(11):
2374 - 2382.
[Abstract]
[Full Text]
[PDF]
J. Wrzesinski, M. Legiewicz, B. Smolska, and J. Ciesiolka
Catalytic cleavage of cis- and trans-acting antigenomic delta ribozymes in the presence of various divalent metal ions
Nucleic Acids Res.,
November 1, 2001;
29(21):
4482 - 4492.
[Abstract]
[Full Text]
[PDF]
Y. Takagi, M. Warashina, W. J. Stec, K. Yoshinari, and K. Taira
SURVEY AND SUMMARY: Recent advances in the elucidation of the mechanisms of action of ribozymes
Nucleic Acids Res.,
May 1, 2001;
29(9):
1815 - 1834.
[Abstract]
[Full Text]
[PDF]
