Mapping of the functional phosphate groups in the catalytic core of deoxyribozyme 10–23 Barbara Nawrot, Kinga Widera, Marzena Wojcik*, Beata Rebowska, Genowefa Nowak and Wojciech J. Stec Department of Bioorganic Chemistry, Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences, Lodz, Poland The RNA-cleaving DNA enzymes, like most ribozymes, require a divalent metal cation for their cleavage activity [1]. Among metal ion-dependent DNA enzymes, deoxyribozyme 10–23, first selected and characterized by Santoro & Joyce [1,2], has been examined most extensively both in vitro and in vivo [3–5]. This enzyme consists of a 15-nucleotide conserved catalytic core and variable substrate recognition arms (Fig. 1A). Cleavage of an RNA substrate is highly sequence-specific, and occurs between the bulged 5¢-purine and paired 3¢-pyr- imidine nucleosides, resulting in the formation of the two products, a 5¢-terminal product with a 2¢,3¢-cyclic phosphate, and a 3¢-terminal product containing an OH group at its 5¢-end. The enzyme preferentially uses Mg 2+ for its activity, although other divalent metal ions are accepted as cofactors [1,2,6]. To date, the structure of the substrate–deoxyribozyme 10–23 active complex remains unknown [7,8], and the mechanistic details of the catalytic reaction are not fully understood. There- fore, much effort has been devoted to determine the role of individual nucleotides in the 10–23 catalytic core, as well as their relative importance [9–13]. Despite numer- ous studies performed on a mutant deoxyribozyme 10– 23 containing chemical modifications inserted into the catalytic core, the role of particular phosphates within this domain has not been investigated in detail. We have studied this issue by systematic modification of each phosphate of the core with phosphorothioate (PS) ana- logs, in which one of the two nonbridging oxygen atoms of the phosphate group was replaced with a sulfur atom. Keywords catalysis; deoxyribozyme; phosphorothioate; rescue effect; thio effect Correspondence B. Nawrot, Department of Bioorganic Chemistry, Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences, Sienkiewicza 112, 90-363 Lodz, Poland Fax: +48 42 6815483 Tel: +48 42 6816970 E-mail: bnawrot@bio.cbmm.lodz.pl *Present address Medical University of Lodz, Department of Structural Biology, Zeligowskiego, Poland (Received 4 October 2006, revised 29 November 2006, accepted 18 December 2006) doi:10.1111/j.1742-4658.2007.05655.x The RNA phosphodiester bond cleavage activity of a series of 16 thio-de- oxyribozymes 10–23, containing a P-stereorandom single phosphorothioate linkage in predetermined positions of the catalytic core from P1 to P16, was evaluated under single-turnover conditions in the presence of either 3mm Mg 2+ or 3 mm Mn 2+ . A metal-specificity switch approach permitted the identification of nonbridging phosphate oxygens (proR P or proS P ) located at seven positions of the core (P2, P4 and P9–13) involved in direct coordination with a divalent metal ion(s). By contrast, phosphorothioates at positions P3, P6, P7 and P14–16 displayed no functional relevance in the deoxyribozyme-mediated catalysis. Interestingly, phosphorothioate modifi- cations at positions P1 or P8 enhanced the catalytic efficiency of the enzyme. Among the tested deoxyribozymes, thio-substitution at position P5 had the largest deleterious effect on the catalytic rate in the presence of Mg 2+ , and this was reversed in the presence of Mn 2+ . Further experiments with thio-deoxyribozymes of stereodefined P-chirality suggested direct involvement of both oxygens of the P5 phosphate and the proR P oxygen at P9 in the metal ion coordination. In addition, it was found that the oxygen atom at C6 of G 6 contributes to metal ion binding and that this interaction is essential for 10–23 deoxyribozyme catalytic activity. Abbreviations AP, 2-aminopurine; DNAzyme, RNA-cleaving deoxyribozyme; PS, phosphorothioate; s 6 G, 6-thioguanosine. 1062 FEBS Journal 274 (2007) 1062–1072 ª 2007 Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences Journal compilation ª 2007 FEBS The PS modification represents the most conservative elemental replacement for the phosphate, although the sulfur atom is slightly larger than the oxygen atom, and the P–S bond is 0.3 A ˚ longer than the P–O bond [14]. Oligonucleotides possessing stereodefined PS internucle- otide linkages have been found useful for clarifying the function of proR P and proS P positions at the scissile site of oligonucleotide substrates. The ribozyme-assisted cleavage reactions were conducted in the presence of divalent metal cations with different affinities for oxygen and sulfur [15–20]. According to the HSAB (Hard and Soft, Acid and Base) rule [21], a reduction in the clea- vage rate of a ‘soft’ thio-substituted substrate should be observed in the presence of ‘hard’ Mg 2+ cation (the thio effect), and restoration of a normal cleavage rate of a sulfur-containing substrate should occur in the presence of thiophilic cations such as Mn 2+ ,Zn 2+ ,orCd 2+ ,in increasing order (the rescue effect). Analysis of these types of interaction led to a better understanding of the mechanistic aspects of the action of naturally occurring catalytic ribozymes: group I and II introns [22–25], the RNA subunit of RNase P [26,27], and the hammerhead ribozymes [18,28–30]. The successful application of P- chiral phosphorothioates in those mechanistic studies prompted us to establish the role of phosphate groups in the catalytic core of deoxyribozyme 10–23. First, we introduced a P-stereorandom single PS linkage in prede- termined positions of the catalytic core in 16 thio-deoxy- ribozymes 10–23 (P1–P16; Table 1, entries 2–17), and conducted metal-specificity switch experiments with Mg 2+ and thiophilic Mn 2+ . These experiments showed that catalytically important phosphate groups were positioned within the catalytic domain of the enzyme. The role of the particular oxygen atoms of the selected phosphate groups is also discussed. Moreover, we ana- lyzed the function of the oxygen moiety at C6 of nucleo- side G 6 positioned within the catalytic loop, by either its removal [substitution with 2-aminopurine (AP) nucleo- side] or its replacement with a sulfur atom by using the 6-thioguanosine (s 6 G) mutant enzyme. Kinetic measure- ments of these deoxyribozyme variants, along with data obtained by Zaborowska et al. [11], proved the import- ance of the oxygen of the carbonyl group at G 6 for the catalytic activity of deoxyribozyme 10–23. Results and Discussion The influence of PS modification on the catalytic activity of deoxyribozyme 10–23 The functional role of the individual phosphate groups in the catalytic core of deoxyribozyme 10–23 was examined by determination of the thio effect and the Mn 2+ -dependent rescue effect of thio-substituted de- oxyribozymes bearing a single PS linkage from P1 to P16, where the P1 phosphate is a 5¢-phosphate of nuc- leotide 1 (G 1 ) (Table 1, entries 2–17). The PS deoxyri- bozymes were synthesized by automated solid-phase synthesis, in which one of the iodine oxidation steps was replaced by sulfurization [31]. Each oligomer was an R P and S P (c. 1 : 1) diastereomeric mixture (Fig. 1). The activity of thio-substituted deoxyribozymes was tested against a short target substrate homosequential with mRNA of aspartyl protease Asp2 (BACE1, acces- sion number AF190725, between nucleotides 1801 and 1817) (Fig. 1). It has already been demonstrated that deoxyribozyme 10–23 accepts not only short RNA substrates but also modified substrates containing a DNA backbone with RNA nucleotides (5¢-purine and 3¢-pyrimidine ribonucleotides) positioned at the scissile bond of the target oligonucleotide [32–34]. We pre- pared a 17-nucleotide chimeric DNAÆRNA substrate with the sequence 5¢-d(ACAGATGA)GUd(CAACC- CT)-3¢, which was easier to synthesize and chemically more stable than an RNA oligonucleotide. All kinetic experiments were performed at a satur- ating concentration of the unmodified deoxyribozyme 1 or thio-deoxyribozymes 2–17 (10 lm) with 32 P-labe- led substrate (0.1 lm) in the presence of 3 mm MgCl 2 . The cleavage product (9-mer) and the substrate were quantified by autoradiography following electrophor- A B O B O P O S O B O O B O P S O O B O S P R P Fig. 1. (A) The structure of deoxyribozyme 10–23. The target sub- strate is a chimeric DNAÆRNA oligonucleotide homosequential to the mRNA of BACE1 (nucleotides 1801–1817). Substrate–enzyme binding occurs via the Watson–Crick mode of base-pairing. The arrow indicates the cleavage site. The positions of the phosphate groups of the catalytic core are numbered from P1 to P16. (B) P- chiral PS internucleotide bonds in PS DNA of S P -sense and R P - sense of chirality, respectively. B. Nawrot et al. Metal-binding site in deoxyribozyme 10–23 FEBS Journal 274 (2007) 1062–1072 ª 2007 Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences Journal compilation ª 2007 FEBS 1063 esis in 20% polyacrylamide gels. The observed rate constants (k obs ) were calculated according to the equa- tion given in Experimental procedures, and compared with the rate constant of the unmodified deoxyribo- zyme (k rel ). The data presented in Table 1 and Fig. 2A indicate that thio-substitutions at phosphates P2, P4 and P9–13 lowered the k rel values by c. 50%. The replacement of 3mm Mg 2+ with 3 mm Mn 2+ resulted in a restoration of the activity to the level of the k Mn obs of the unmodified enzyme (Table 1, Fig. 2B), implying a possible coordi- nation of the metal cation to one of the two (the proR P or the proS P ) oxygen atoms. However, it should be noted that the k obs values were c. 30-fold higher in the presence of Mn 2+ compared with Mg 2+ (Table 1). A similar high rate of the cleavage reaction in the pres- ence of Mn 2+ was reported previously [1,35]. As pro- posed by Breaker et al. [35], it is possible that the higher activity of deoxyribozyme 10–23 in the presence of Mn 2+ may result from the fact that Mn 2+ ,asa stronger Lewis acid, participates more effectively in catalysis steps such as the acceleration of the ribose 2¢-hydroxyl group deprotonation, stabilization of a negative charge that may develop on the nonbridging oxygen in a transition state, and ⁄ or stabilization of the negative charge on the oxygen atom of the 5¢-leaving group. Among the tested modified enzymes, the biggest thio effect (a 16-fold reduction in the cleavage activity; Table 1, Fig. 2A) was found for the PS enzyme modi- fied at position P5. The reduction was much bigger than the two-fold reduction expected if only one of the diastereomers coordinated the metal ion, suggesting that the sulfur atoms in both the proR P and proS P positions hindered direct contact with metal ions. Interestingly, this PS enzyme regained its activity in the presence of Mn 2+ , with the k Mn obs value being 176- fold higher than the k Mg obs value. This value, however, was still c. 3-fold lower than that measured for the unmodified reference at the same conditions (Table 1). It seems that the slightly lower reaction rate of this PS enzyme in the presence of Mn 2+ might be attributed to Table 1. Single-turnover rate constants of the cleavage reactions catalyzed by unsubstituted and thio-substituted deoxyribozyme 10–23. NA, value not available. Entry DNAzyme abbreviation ⁄ PS position 5¢fi3¢ sequence of the catalytic core a k Mg obs (min )1 ) b k Mg rel c Thio effect k Mn obs (min )1 ) d k Mn rel e k Mn obs ⁄ k Mg obs (rescue effect) 1 d(AGGCTAGCTACAACGAT) 0.27 ± 0.028 1 1.0 8.00 ± 0.42 1 30 2 P1 d(A PS GGCTAGCTACAACGAT) 0.85 ± 0.042 3.10 0.3 0.36 ± 0.031 f 3.0 f NA 3 P2 d(AG PS GCTAGCTACAACGAT) 0.15 ± 0.018 0.56 1.8 8.10 ± 0.57 1.01 54 4 P3 d(AGG PS CTAGCTACAACGAT) 0.24 ± 0.0078 0.89 1.1 8.00 ± 1.10 1.00 33 5 P4 d(AGGC PS TAGCTACAACGAT) 0.16 ± 0.0071 0.59 1.7 8.60 ± 1.40 1.08 54 6 P5 d(AGGCT PS AGCTACAACGAT) 0.017 ± 0.0014 0.06 16 3.00 ± 0.071 0.38 176 7 P6 d(AGGCTA PS GCTACAACGAT) 0.21 ± 0.0071 0.78 1.3 8.20 ± 1.10 1.03 39 8 P7 d(AGGCTAG PS CTACAACGAT) 0.24 ± 0.070 0.89 1.1 8.70 ± 0.071 1.10 36 9 P8 d(AGGCTAGC PS TACAACGAT) 0.44 ± 0.014 1.60 0.6 0.16 ± 0.038 f 1.30 f NA 10 P9 d(AGGCTAGCT PS ACAACGAT) 0.14 ± 0.014 0.52 1.9 8.40 ± 0.64 1.05 60 11 P10 d(AGGCTAGCTA PS CAACGAT) 0.14 ± 0.028 0.52 1.9 9.30 ± 0.71 1.20 66 12 P11 d(AGGCTAGCTAC PS AACGAT) 0.15 ± 0.018 0.56 1.8 8.20 ± 1.10 1.03 55 13 P12 d(AGGCTAGCTACA PS ACGAT) 0.15 ± 0.014 0.56 1.8 10.00 ± 1.70 1.30 67 14 P13 d(AGGCTAGCTACAA PS CGAT) 0.14 ± 0.024 0.52 1.9 9.20 ± 0.71 1.20 66 15 P14 d(AGGCTAGCTACAAC PS GAT) 0.28 ± 0.014 0.96 1.0 9.30 ± 0.64 1.20 33 16 P15 d(AGGCTAGCTACAACG PS AT) 0.38 ± 0.035 1.40 0.7 9.40 ± 0.71 1.20 25 17 P16 d(AGGCTAGCTACAACGA PS T) 0.24 ± 0.030 0.89 1.1 7.40 ± 0.78 0.93 31 18 P1 ⁄ P8 d(A PS GGCTAGC PS TACAACGAT) 0.76 ± 0.080 2.80 0.78 ± 0.048 f 6.5 f NA a The sequences of PS deoxyribozymes 10–23 containing a single PS linkage of stereorandom P-configuration (equal amounts of R P and S P diastereomers) in the selected positions of the catalytic core marked from P1 (phosphate bond between A 0 and G 1 ) to P16 (phosphate bond between A 15 and T 16 ). b, d RNA cleavage reactions were performed in 20 mM Tris ⁄ HCl (pH 7.5), containing 100 mM NaCl, b 3mM Mg 2+ or d 3mM Mn 2+ under single-turnover conditions with 0.1 lM 5¢-end 32 P-labeled substrate and 10 lM deoxyribozyme, at 37 °C. c k Mg rel ¼ the ratio of the k obs values of modified deoxyribozyme to unmodified deoxyribozyme, in the presence of Mg 2+ . e k Mn rel ¼ the ratio of the k obs values of modified deoxyribozyme to unmodified deoxyribozyme, in the presence of Mn 2+ . f Reactions were performed in 20 m M Tris ⁄ HCl (pH 7.5), containing 100 mM NaCl and 0.06 mM Mn 2+ under single-turnover conditions with 0.1 lM 5¢-end 32 P-labeled sub- strate and 10 l M deoxyribozyme, at 37 °C. Values of k obs for unsubstituted and thio-substituted deoxyribozyme reactions represent mean values of four independent experiments, and errors indicate deviations between individual experiments. The obtained data were normal- ized to a k obs of 0.12 ± 0.014 min )1 for reaction of the unmodified deoxyribozyme in 0.06 mM Mn 2+ . Metal-binding site in deoxyribozyme 10–23 B. Nawrot et al. 1064 FEBS Journal 274 (2007) 1062–1072 ª 2007 Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences Journal compilation ª 2007 FEBS the bulky sulfur atom, which could influence the geom- etry of metal ion interactions, or the geometry of the catalytic conformation of the core. However, this par- tial inability of thiophilic metal ions to fully rescue catalysis does not eliminate the possibility that this modified phosphate is in direct contact with a catalyti- cally important cation [36]. Other modified deoxyribozymes, containing thio- substitutions at positions P3, P6, P7, P14, P15 and P16, retained catalytic activity comparable with that of the unmodified enzyme in the presence of Mg 2+ and Mn 2+ . These data constitute strong evidence against direct coordination of a metal cation to both the proR P and proS P phosphate oxygen atoms at these positions during catalysis. Also, it is possible that a sulfur atom in these positions does not alter the struc- ture of the catalytically active core of deoxyribozyme. This observation suggests the possibility of using parti- ally modified PS analogs of deoxyribozymes to improve their stability against intracellular endonuc- leases in cellular systems. The catalytic activity of double PS-substituted deoxyribozyme 10–23 PS modification at positions P1 or P8, surprisingly, accelerated the cleavage rates (3-fold and 1.6-fold, respectively) in the presence of Mg 2+ as well as Mn 2+ (Table 1, Fig. 3). The k obs and k rel values for these enzymes were calculated from the reactions performed in 3 mm Mg 2+ or 0.06 mm Mn 2+ . The concentration of Mn 2+ was reduced 50-fold, because reactions per- formed in the presence of 3 mm Mn 2+ reached comple- tion in less than 5 s, making kinetic analysis impossible. Whereas the P8 substitution had only a minor effect both in the presence of Mg 2+ and in the presence of Mn 2+ , causing a 30–60% increase in k rel , the effect of the double substitution P1 ⁄ P8 was strik- ingly different, depending on the metal ion present. There was no increase of the enzyme efficiency in the presence of Mg 2+ , compared to P1 substitution itself, but in the presence of Mn 2+ the k rel for the P1 ⁄ P8 enzyme was over two-fold higher than the k rel for the P1 enzyme and 6.5-fold higher than that for the unmodified reference (Table 1, Fig. 3). For the P1 ⁄ P8 PS congener, the k Mg obs and k Mn obs values were nearly iden- tical, despite a 50-fold difference in the concentration of metal ions present in the catalysis reaction, and the k Mn obs value for this mutant enzyme was three-fold higher than the k Mn obs value for the unmodified reference. The obtained data demonstrate that the P1 ⁄ P8 A unmodified P2 P3 P4 P5 P6 P7 P9 P10 P11 P12 P13 P14 P15 P16 unmodified P2 P3 P4 P5 P6 P7 P9 P10 P11 P12 P13 P14 P15 P16 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 k rel 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 k rel B Fig. 2. Comparison of the relative rates of cleavage (k rel ) of thio- substituted deoxyribozymes 10–23 in the presence of 3 m M MgCl 2 (A) and 3 mM MnCl 2 (B). 0 1 2 3 4 5 6 7 8 unmodified Mg 2+ unmodified Mn 2+ P1 Mg 2+ P1 Mn 2+ P8 Mg 2+ P8 Mn 2+ P1/P8 Mg 2+ P1/P8 Mn 2+ k rel Fig. 3. Comparison of the relative rates of cleavage (k rel ) of thio- substituted deoxyribozymes 10–23 in the presence of 3 m M MgCl 2 (white bars) and 0.06 mM MnCl 2 (gray bars). B. Nawrot et al. Metal-binding site in deoxyribozyme 10–23 FEBS Journal 274 (2007) 1062–1072 ª 2007 Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences Journal compilation ª 2007 FEBS 1065 PS-mutant enzyme is c. 50-fold more active in the presence of Mn 2+ than in the presence of Mg 2+ , and its k obs value in the presence of 3 mm Mn 2+ could reach a value of c. 40 min )1 . This is a four times higher value than the highest one so far reported in the literature for catalytic nucleic acids [35]. Moreover, this double PS congener is c. 150-fold more active in the presence of Mn 2+ than the unmodified reference in the presence of Mg 2+ . A possible explanation for these results is that both pairs of oxygen atoms at the P1 and P8 phosphates do not directly interact with metal ions, and such a double PS modification, together with the presence of Mn 2+ , facilitates a catalytically favora- ble conformation of the 10–23 core. Moreover, one cannot exclude the possibility that the 10–23 enzyme operates with two metal ions interacting with different sets of residues. Our finding that the introduction of a PS bond at the P1 site of deoxyribozyme 10–23 causes about three-fold stimulation of the cleavage rate, irrespective of the metal ion used, demonstrates that chemical modifications of the deoxyribozyme backbone can be used to improve both its stability and its catalytic effi- ciency in cellular experiments. Effect of P-chirality on the catalytic activity of deoxyribozyme 10–23 In order to obtain a deeper insight into the functional role of the oxygen atoms of the P5 phosphate group in the catalytic core of deoxyribozyme 10–23, we pre- pared two PS deoxyribozymes with stereodefined R P - PS or S P -PS linkages at that position and measured the rate of RNA cleavage under analogous conditions in the presence of 3 mm Mg 2+ (Fig. 4). We found that R P -PS and S P -PS substitutions at position P5 reduced k Mg rel by a factor of 34 and 21, respectively (Table 2, Fig. 5A). As k obs values were measured at a saturating concentration of the PS enzymes, their lowered activity could not be attributed to decreased substrate binding, thus implying that sulfur substitution disrupted specific Mg 2+ interactions with nonbridging phosphate oxy- gens. In 3 mm Mn 2+ buffer, the R P -PS and S P -PS deoxyribozyme P5-mediated cleavage activity was significantly enhanced (73-fold and 108-fold increase of k obs values, respectively; Table 2, Fig. 5B). The observed thio effect and rescue effect values for partic- ular P-chiral diastereomers slightly differed from those determined for the diastereomeric mixture of this PS enzyme, and these differences may result from experi- mental errors. The remarkable increase of the catalytic rate for the reactions carried out in the presence of Mg 2+ and each of the P-chiral diastereomeric deoxyri- bozymes suggests that Mn 2+ can stimulate 10–23 enzyme activity in a way that depends on the simulta- neous metal ion interactions with both nonbridging oxygens at position P5. Thus, earlier suggestions are fully confirmed by our findings [13]. Other stereodefined PS deoxyribozymes with a PS bond at position P9 (prepared synthetically by using the same pair of diastereomeric T PS A phosphoramidite monomers), as well as those modified at positions P3 and P7, were evaluated. The latter two pairs of diaster- 0 5 10 15 20 0 20 40 60 80 100 Time [min] Time [min] Degradation of substrate [%] 0 50 100 150 200 250 0 25 50 75 100 Degradation of substrate [%] 0 4 10 20 30 45 60 90 120 150 180 210 240 min 0 0.16 0.5 1.0 2.0 4.0 6.0 8.0 10 20 30 min A B C D Fig. 4. Comparison of the Mg 2+ -dependent activity of the unmodi- fied deoxyribozyme 10–23 with that of thio-substituted deoxyribo- zyme R P -P5 in the presence of 3 mM MgCl 2 . Time course of cleavage reaction of a chimeric DNAÆRNA oligonucleotide by the unmodified (A, B) and R P -P5 (C, D) deoxyribozymes. Metal-binding site in deoxyribozyme 10–23 B. Nawrot et al. 1066 FEBS Journal 274 (2007) 1062–1072 ª 2007 Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences Journal compilation ª 2007 FEBS eomeric deoxyribozymes were prepared from diastereo- meric R P dimers and S P dimers, G PS C [37]. We chose this sequence following Taira and coworkers’ sugges- tion that the proR P phosphate at position P7 (between G 6 and C 7 ) might be important for the catalytic activ- ity of deoxyribozyme 10–23 (unpublished results). The stereodefined deoxyribozymes (R P and S P at positions P3, P7 and P9) were characterized in cleavage reac- tions similar to those described above. The kinetic parameters calculated for these reactions are listed in Table 2 and shown in Fig. 5. Interestingly, in these experiments we found higher thio effects and rescue effects (10 and 115, respectively) for the R P -PS deoxy- ribozyme P9, and a lack of these effects for its S P counterpart, which implies direct involvement in the metal ion coordination of the proR P , but not proS P , oxygen at position P9. The k rel values for nonbridging phosphate oxygens at positions P3 and P7 reach sim- ilar values, indicating a lack of direct coordination of a metal cation to the proR P and proS P oxygen atoms at these positions. These findings further confirm our previous data obtained for the mixtures of diastereo- mers of PS deoxyribozymes. We compared our data with those published for hammerhead ribozymes containing site-specific PS modifications at either the proR P or proS P positions [17]. Single-turnover relative rates of RNA cleavage, determined at 10 mm Mg 2+ , were reduced three-fold for R P -PS isomers at positions A 13 and A 14 , and S P -PS isomers at positions A 6 and U 16.1 , 10-fold for the R P isomer at position A 9 , and 1000-fold for the R P isomer at position U 1.1 , relative to the reactions performed by the hammerhead enzyme. In the analogous reactions performed in the presence of 10 mm Mn 2+ , k rel values for R P -PS isomers at positions A 9 and U 1.1 increased two-fold and 10-fold, respectively [17]. Thus, the thio and rescue effect values observed in our studies for PS deoxyribozymes were much stronger than those observed for hammerhead constructs, except for the k rel value determined in the presence of Mg 2+ for the R P -PS isomer at position U 1.1 of the hammerhead ribozyme. Mutational analysis of nucleoside in position 6 of the catalytic core We were interested in whether there are any other lig- ands in the 10–23 catalytic core that might be directly involved in stabilization of the catalytically active architecture of the deoxyribozyme. As has already been proven, the hammerhead ribozyme metal-binding site utilizes both nonbridging oxygen atoms of the A 9 phosphate as well as nitrogen N7 of the subsequent guanosine unit G 10.1 [38]. We were interested in deter- mining whether the nucleotide residue following the A 5 unit in deoxyribozyme 10–23 plays any role in cata- lysis. Although the exact metal-binding site of deoxyri- bozyme 10–23 is not yet known, it has already been suggested by Kurreck and coworkers that A 5 and G 6 residues within the catalytic core could be directly involved in metal ion binding [11,13]. To characterize the functional role of the oxygen moiety at C6 of G 6 , we replaced this guanosine with its analogs, s 6 G and AP nucleoside (Fig. 6), creating two analogs of the DNA enzyme, s 6 G-zyme and AP-zyme, respectively (Table 3). The k rel values observed for these enzymes Table 2. Single-turnover rate constants for stereodefined thio-deoxyribozyme-mediated reactions in the presence of Mg 2+ and Mn 2+ . Entry DNAzyme abbreviation ⁄ PS position a k Mg obs (min )1 ) b k Mg rel d Thio effect k Mn obs (min )1 ) c k Mn rel e k Mn obs ⁄ k Mg obs f (rescue effect) 1 Unmodified 0.27 ± 0.028 1 1 8.0 ± 0.42 1 30 4a R P -P3 0.30 ± 0.015 1.1 0.90 9.1 ± 0.57 1.1 30 4b S P -P3 0.33 ± 0.028 1.2 0.83 7.5 ± 0.35 0.94 23 6a R P -P5 0.0077 ± 00078 0.029 34 0.56 ± 0.042 0.070 73 6b S P -P5 0.013 ± 0.0014 0.048 21 1.4 ± 0.14 0.18 108 8a R P -P7 0.29 ± 0.003 1.1 0.91 9.4 ± 0.28 1.2 32 8b S P -P7 0.28 ± 0.028 1.04 0.96 8.6 ± 0.50 1.1 31 10a R P -P9 0.026 ± 0.0007 0.096 10 3.0 ± 0.21 0.38 115 10b S P -P9 0.12 ± 0.005 0.44 2.3 0.98 ± 0.021 0.12 8.2 a R P and S P are absolute configurations at the P-chiral center at a given PS linkage. b, c All RNA cleavage reactions were performed in 20 m M Tris ⁄ HCl (pH 7.5), containing 100 mM NaCl and b 3mM Mg 2+ or c 3mM Mn 2+ under single-turnover conditions with 0.1 lM 5¢-end 32 P- labeled substrate and 10 l M deoxyribozyme, at 37 °C. Values of k obs for nonsubstituted and thio-substituted deoxyribozyme reactions repre- sent mean values of four independent experiments, and errors indicate deviations between individual experiments. d k Mg rel ratio of the k obs values for the modified and unmodified deoxyribozymes in the presence of Mg 2+ . e k Mn rel ratio of the k obs values for the modified and unmodi- fied deoxyribozymes in the presence of Mn 2+ . f The values of the rescue effect were calculated from k Mn obs ⁄ k Mg obs . B. Nawrot et al. Metal-binding site in deoxyribozyme 10–23 FEBS Journal 274 (2007) 1062–1072 ª 2007 Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences Journal compilation ª 2007 FEBS 1067 (shown in Table 3) demonstrate that the stimulation of the catalytic activity in the presence of Mn 2+ was sim- ilar for unmodified and s 6 G-substituted enzymes. The observed thio effect is about 20, and the rescue effect is 28, implying that the oxygen atom of the carbonyl moiety serves as a metal ion ligand. In contrast to the inosine substitution [11], exchange of the G 6 base with AP nucleoside resulted in complete loss of catalytic activity, independent of the metal ion (no substrate cleavage over 8 h; Table 3). These findings clearly indi- cate that the oxygen at C6 is essential for the catalytic activity of deoxyribozyme 10–23, whereas the exo amino group of G 6 is not of functional importance. In addition, we extended our mutational analysis to the nucleoside at position 6 by the replacement of G 6 with a 7-deaza-dG unit. This substitution resulted in a 104-fold loss of activity of the DN 7 -zyme in the pres- ence of Mg 2+ , suggesting that the N7 nitrogen partici- pates in the formation of a functionally important intramolecular hydrogen bond within the deoxyribo- zyme 10–23 catalytic core. The k obs for this enzyme increased by almost three orders of magnitude upon addition of Mn 2+ , and was about 30-fold greater than that for the unmodified reference (Table 3). We do not offer any rational explanation for the nature of the extremely high k Mn obs ⁄ k Mg obs value. One can only speculate that this  1000 rescue value for the 7-deazaguanosine- modified enzyme may result from conformational rear- rangement of this modified 10–23 core in the presence of the soft metal ion, involving hydrogen bond pat- terns within the catalytic loop. Implications and Conclusions The present results support the idea that phosphate oxygens of the catalytic core of deoxyribozyme 10–23 participate in stabilization of the catalytically active conformation. Using sulfur-modified deoxyribozymes, we identified phosphate groups important for catalysis. We found that the metal-binding site of deoxyribo- zyme 10–23 involves both nonbridging oxygens of the P5 phosphate of adenosine at position 5, and the oxy- gen atom of the 6-carbonyl group of the subsequent nucleoside (G 6 ). Our model of the metal-binding site in the catalytic core of deoxyribozyme 10–23 includes the interactions of divalent cations with both the pro- R P and proS P oxygens of P5, and an interaction with the oxygen ligand at C6 of the subsequent guanosine nucleotide (Fig. 7). One can argue that in this model the distances between the oxygen ligands of P5 and the oxygen of G 6 are too large to be spanned by a single metal ion. However, it is possible that the architecture of the active conformation of the catalytic core allows for such interactions, or that more than one metal ion is involved in catalysis. Contributions of other ligands cannot be excluded, and the first candidate is the proR P oxygen of phosphate P9, between the T 8 and A 9 nucleosides (Table 2). It is also possible that other functional groups of the catalytic core serve as metal ion ligands, because, as we have already suggested, there are at least seven more nonbridging phosphate oxygens, at positions P2, P4, P9, P10, P11, P12 and P13, which exhibit remarkable thio and rescue effects. Besides the oxygen ligands of the internucleotide bonds, some other functional groups, as indicated in other studies [11], may form intraloop hydrogen bonds or coordinate to metal ion(s) directly or by water bridges. In conclusion, the reported data, along with results obtained by systematic site-directed PS substitutions, enabled the proposal of a model for the metal-binding site in the catalytic core of deoxyribozyme 10–23. In P3 P R P3 P S P5 P R P5 P S P7 P R P 7 P S P9 P R P 9 P S P 3 P R P 3 P S P 5 P R P 5 P S P 7 P R P 7 P S P 9 P R P 9 P S unmodified unmodified 0.00 0.25 0.50 0.75 1.00 1.25 0.00 0.25 0.50 0.75 1.00 1.25 1.50A B k rel k rel Fig. 5. Comparison of the relative rates of cleavage (k rel ) of PS-ster- eodefined thio-deoxyribozymes 10–23 in the presence of 3 m M MgCl 2 (A) and 3 mM MnCl 2 (B). Metal-binding site in deoxyribozyme 10–23 B. Nawrot et al. 1068 FEBS Journal 274 (2007) 1062–1072 ª 2007 Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences Journal compilation ª 2007 FEBS this model, the plausible ligands for metal coordina- tion are the proR P and proS P oxygen atoms of the P5 phosphate, and the proR P oxygen at position P9, as well as the carbonyl oxygen of the guanosine unit at position 6 of the 10–23 catalytic core. In addition, sev- eral other phosphate oxygens and nucleobase func- tional groups can serve as metal-binding ligands and ⁄ or hydrogen bond acceptors within the catalytic core, but no detailed information is yet available. Therefore, further experiments are required to identify possible metal-binding ligands and to study the struc- ture of deoxyribozyme 10–23 at the atomic level, either by molecular modeling or by solution of the crystal structure. In addition, our observations that nonbridged oxy- gens at phosphates at positions P3, P6, P7, P14 and P15 could be replaced by a sulfur without substantial loss of activity, and that the introduction of the PS bond at the P1 and P8 sites stimulated catalytic activ- ity, provided us wi th a starting point for the creation of variants of deoxyribozyme 10–23 with not only improved catalytic effectiveness but also better stability against cellular endonucleases (such studies are pres- ently being carried out in our laboratory). Fig. 6. Chemical structures of the various nucleotide analogs employed in the current study. Table 3. Single-turnover kinetics of cleavage reaction mediated by the deoxyribozyme 10–23 modified at position 6 of the catalytic core. ND, not determined. Deoxyribozyme Substitution k Mg obs (min )1 ) a k Mg rel c k Mn obs (min )1 ) b k Mn rel d k Mn obs ⁄ k Mg obs e Unmodified (WT) None 0.27 ± 0.028 1 8.0 ± 0.42 1 30 G 6 fi adenosine f ND – – – – G 6 fi inosine f As WT – – – – s 6 G-zyme G 6 fi 6-thio-dG 0.013 ± 0.0028 0.048 0.37 ± 0.035 0.046 28 AP-zyme G 6 fi 2-aminopurine nucleoside ND – ND – – DN 7 -zyme G 6 fi 7-deaza-dG 0.0026 ± 0.00028 0.0096 2.6 ± 0.28 0.33 1000 a, b All RNA cleavage reactions were performed in 20 mM Tris ⁄ HCl (pH 7.5), containing 100 mM NaCl and a 3mM Mg 2+ or b 3mM Mn 2+ , under single-turnover conditions with 0.1 l M 5¢-end 32 P-labeled substrate and 10 lM deoxyribozyme, at 37 °C. Values of k obs for unmodified and mutated deoxyribozyme reactions represent mean values of three independent experiments, and errors indicate deviations between individual experiments. c k Mg rel ¼ ratio of the k obs values of modified deoxyribozyme to unmodified deoxyribozyme, in the presence of Mg 2+ . d k Mn rel ¼ ratio of the k obs values of modified deoxyribozyme to unmodified deoxyribozyme, in the presence of Mn 2+ . e The values of the res- cue effect were calculated from k Mn obs ⁄ k Mg obs f . B. Nawrot et al. Metal-binding site in deoxyribozyme 10–23 FEBS Journal 274 (2007) 1062–1072 ª 2007 Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences Journal compilation ª 2007 FEBS 1069 Experimental procedures Deoxyribozymes and substrate The unmodified deoxyribozyme and its substrate oligonu- cleotide (Fig. 1) were synthesized using an ABI 394 DNA synthesizer (Applied Biosystems, Inc., Foster City, CA) and commercially available phosphoramidite monomers (Glen Research, Sterling, VA). Base-modified deoxyribozymes (AP-zyme and DN 7 -zyme) were synthesized routinely using commercially available monomers (2-aminopurine 2¢-deoxy- ribonucleoside and 7-deaza-2¢-deoxyguanosine phosphoram- idite monomers; Glen Research). The s 6 G-zyme was synthesized routinely using protected 6-thio-2¢-deoxyguano- sine phosphoramidite prepared according to the published procedure [39] and commercially available UltraMILD phosphoramidites (Glen Research). The deprotection step was performed as described previously [39]. Oligomers were purified by RP-HPLC (ODS Hypersil column, Alltech Associates, Inc., Deerfield, IL) followed by preparative elec- trophoresis in a 20% polyacrylamide gel containing 7 m urea. PS-stereodefined oligonucleotides were synthesized by incorporation of PS dinucleoside building blocks into the oligonucleotide chain according to our recently described procedure [37]. The structure and purity of the PS oligonu- cleotides were confirmed by MALDI-TOF MS and RP- HPLC, as well as by PAGE. The absolute configuration at the chiral phosphorus center was assigned enzymatically with stereospecific nP1 (Sigma-Aldrich, St Louis, MO) and svPDE (Boehringer Mannheim, Germany) nucleases. Oligonucleotide labeling The substrate oligonucleotide of an RNAÆDNA chimeric sequence (Fig. 1) was 5¢-labeled with [c- 32 P]ATP and T4 polynucleotide kinase (Amersham, Little Chalfont, UK). A mixture containing 10 mm Tris ⁄ HCl (pH 8.5), 10 mm MgCl 2 ,7mm 2-mercaptoethanol, 30 lm (0.1 A 260 unit) oligonucleotide, 1 lL (10 lCi) of [c- 32 P]ATP and T4 polynucleotide kinase (6 units) was incubated for 30 min at 37 °C, and then heat denatured and stored at ) 20 °C. Enzymatic assay The substrate cleavage reactions were performed under single-turnover conditions with the DNA enzyme in 100- fold excess over the substrate. The 5¢-labeled substrate (0.1 lm) was incubated with deoxyribozyme (10 lm)in 20 mm Tris ⁄ HCl (pH 7.5) containing 100 mm NaCl, and 3mm MgCl 2 or 3 mm MnCl 2 ,at37°C. After various time intervals, 10 lL aliquots were withdrawn, and the cleavage reaction was stopped by addition of 50 mm EDTA and by cooling on ice. Before electrophoresis, 8 lL of formamide containing 0.03% bromophenol blue and 0.03% xylene cyanol was added to each sample, and the cleavage products were separated from noncleaved sub- strate by electrophoresis in 20% polyacrylamide gel under denaturing conditions. The amount of product was deter- mined by autoradiography with PhosphorImager (Molecu- lar Dynamics, Sunnyvale, CA), and the observed rate constants (k obs ) were calculated from a pseudo-first-order reaction equation, Y ¼ [EP] [1 ) exp(– k obs t)], where Y is the percentage of the cleaved product at time t, and EP is the endpoint, showing the percentage of cleaved product at the plateau of reaction. Reactions were carried out near to completion. Endpoints between 80% and 90% were used in kinetic analyses. In all cases, good fits to the appropriate kinetic model were obtained, with R 2 > 0.96. The k obs values for cleavage of the substrate by modified deoxyribozymes represent mean values of at least three independent experiments, and errors indicate deviations between individual experiments. The error bars in Figs 2, 3 and 4 were calculated in the following manner. The rel- ative k-values (k rel ) were calculated as a ratio of the k obs values for the modified and unmodified enzyme. The upper limits for k rel were calculated as a ratio of (k obs M+SD M ) ⁄ (k obs U ) SD U ), where k obs M and SD M , and k obs U and SD U , are the mean reaction rates and SD errors for the modified and unmodified enzymes, respectively. Similarly, lower limits for k rel were calculated from the equation (k obs M ) SD M ) ⁄ (k obs U+SD U ). To ensure that the substrate was completely saturated by the deoxyribozyme, the rate constants at concentrations of the deoxyribozyme increasing from 1 to 30 lm were measured (data not shown). The rate of cleavage was inde- pendent of the concentration of the deoxyribozyme above 10 lm, indicating that the chemical step within the deoxy- ribozyme-assisted substrate cleavage was a rate-limiting step. The ‘thio effect’ was calculated as a ratio of k Mg obs of the reference unmodified enzyme to k Mg obs of the particular N N N NH O NH 2 O P O O O O O O P O O A O O O T O Mg 2+ O O T O O P O O O A O ? ? pro-R P P5 P9 G 6 A 5 pro-R P pro-S P T 4 A 9 T 8 Fig. 7. Model for the metal-binding site in the catalytic core of deoxyribozyme 10–23. No clear evidence is given concerning whe- ther these coordinations are to the same or different magnesium ions. Metal-binding site in deoxyribozyme 10–23 B. Nawrot et al. 1070 FEBS Journal 274 (2007) 1062–1072 ª 2007 Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences Journal compilation ª 2007 FEBS modified enzyme, and the ‘rescue effect’ was calculated as a ratio of k Mn obs to k Mg obs of modified enzyme. Acknowledgements The authors thank Professor J. Connolly of Glasgow University for critical reading of the manuscript and valuable suggestions. This work was supported by the Ministry of Science and Higher Education (Poland) through the Centre of Molecular and Macromolecular Studies, Polish Academy of Sciences, under Decision 70⁄ E-63 ⁄ SN-014 ⁄ 2006 and ICGEB project CRP ⁄ POL04-01. References 1 Santoro SW & Joyce GF (1998) Mechanism and utility of an RNA-cleaving DNA enzyme. Biochemistry 37 , 13330–13342. 2 Santoro SW & Joyce GF (1997) A general purpose RNA-cleaving DNA enzyme. Proc Natl Acad Sci USA 94, 4262–4266. 3 Dass CR (2004) Deoxyribozymes: cleaving a path to clinical trials. Trends Pharmacol Sci 25, 395–397. 4 Silverman SK (2005) In vitro selection, characterization, and application of deoxyribozymes that cleave RNA. Nucleic Acids Res 33, 6151–6163. 5 Dass CR (2006) A DNA enzyme that cleaves RNA. Preclinical anticancer activity of DNA-based cleavage molecules. Drug Dev Ind Pharm 32, 1–5. 6 Breaker RR & Joyce GF (1994) A DNA enzyme that cleaves RNA. Chem Biol 1, 223–229. 7 Nowakowski J, Shim PJ, Prasad GS, Stout CD & Joyce GF (1999) Crystal structure of an 82-nucleotide RNA– DNA complex formed by the 10–23 DNA enzyme. Nat Struct Biol 6, 151–156. 8 Nowakowski JP, Shim J, Stout CD & Joyce GF (2000) Alternative conformations of a nucleic acid four-way junction. J Mol Biol 300 , 93–102. 9 Sugimoto N, Okumoto Y & Ohmichi T (1999) Effect of metal ions and sequence of deoxyribozymes on their RNA cleavage activity. J Chem Soc Perkin Trans 2, 1381–1386. 10 Okumoto Y & Sugimoto N (2000) Effects of metal ions and catalytic loop sequences on the complex formation of a deoxyribozyme and its RNA substrate. J Inorg Biochem 82, 189–195. 11 Zaborowska Z, Furste JP, Erdmann VA & Kurreck J (2002) Sequence requirements in the catalytic core of the ‘10–23’ DNA enzyme. J Biol Chem 277, 40617–40622. 12 Schubert S, Gul DC, Grunert HP, Zeichhardt H, Erdmann VA & Kurreck J (2003) RNA cleaving ‘10–23’ DNAzymes with enhanced stability and activity. Nucleic Acids Res 31 , 5982–5992. 13 Zaborowska Z, Schubert S, Kurreck J & Erdmann VA (2005) Deletion analysis in the catalytic region of the 10–23 DNA enzyme. FEBS Lett 579, 554–558. 14 Frey PA & Sammons RD (1985) Bond order and charge localization in nucleoside phosphorothioates. Science 228, 541–545. 15 Waring RB (1989) Identification of phosphate groups important to self-splicing of the Tetrahymena rRNA intron as determined by phosphorothioate substitution. Nucleic Acids Res 17, 10281–10293. 16 Slim G & Gait MJ (1991) Configurationally defined phosphorothioate-containing oligoribonucleotides in the study of the mechanism of cleavage of hammerhead ribozymes. Nucleic Acids Res 19, 1183–1188. 17 Knoll R, Bald R & Furste JP (1997) Complete identifi- cation of nonbridging phosphate oxygens involved in hammerhead cleavage. RNA 3, 132–140. 18 Scott EC & Uhlenbeck OC (1999) A re-investigation of the thio effect at the hammerhead cleavage site. Nucleic Acids Res 27 , 479–484. 19 Yoshinari K & Taira K (2000) A further investigation and reappraisal of the thio effect in the cleavage reac- tion catalyzed by a hammerhead ribozyme. Nucleic Acids Res 28 , 1730–1742. 20 Shan SO & Herschlag D (2000) An unconventional ori- gin of metal-ion rescue and inhibition in the Tetrahy- mena group I ribozyme reaction. RNA 6, 795–813. 21 Pearson RG (1968) Hard and soft acids and bases, HSAB, part I: fundamental principles. J Chem Educ 45, 581–587. 22 Piccirilli JA, Vyle JS, Caruthers MH & Cech TR (1993) Metal ion catalysis in the Tetrahymena ribozyme reac- tion. Nature 361, 85–88. 23 Padgett RA, Podar M, Boulanger SC & Perlman PS (1994) The stereochemical course of group II intron self-splicing. Science 266, 1685–1688. 24 Podar M, Perlman PS & Padgett RA (1995) Stereochemi- cal selectivity of group II intron splicing, reverse splicing, and hydrolysis reactions. Mol Cell Biol 15, 4466–4478. 25 Weinstein LB, Jones BC, Cosstick R & Cech TR (1997) A second catalytic metal ion in group I ribozyme. Nature 388, 805–808. 26 Warnecke JM, Furste JP, Hardt WD, Erdmann VA & Hartmann RK (1996) Ribonuclease P (RNase P) RNA is converted to a Cd(2+)-ribozyme by a single Rp-phos- phorothioate modification in the precursor tRNA at the RNase P cleavage site. Proc Natl Acad Sci USA 93, 8924–8928. 27 Warnecke JM, Held R, Busch S & Hartmann RK (1999) Role of metal ions in the hydrolysis reaction cat- alyzed by RNase P RNA from Bacillus subtilis. J Mol Biol 290, 433–445. 28 Zhou DM, Kumar PKR, Zhang LH & Taira K (1996) Ribozyme mechanism revisited: evidence against direct B. Nawrot et al. Metal-binding site in deoxyribozyme 10–23 FEBS Journal 274 (2007) 1062–1072 ª 2007 Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences Journal compilation ª 2007 FEBS 1071 [...]...Metal-binding site in deoxyribozyme 10–23 29 30 31 32 33 coordination of a Mg2+ ion with the proR oxygen of the scissile phosphate in the transition state of a hammerhead ribozyme-catalyzed reaction J Am Chem Soc 118, 8969–8970 Zhou DM, He QC, Zhou JM & Taira K (1998) Explanation by a putative triester-like mechanism for the thio effects and Mn2+ rescues in reactions catalyzed by... 363–377 37 Nawrot B, Rebowska B, Cieslinska K & Stec WJ (2005) New approach to the synthesis of oligodeoxyribonucleotides modified with phosphorothioates of predetermined sense of P-chirality Tetrahedron Lett 46, 6641–6644 38 Suzumura K, Takagi Y, Orita M & Taira K (2004) NMR-based reappraisal of the coordination of a metal ion at the proRp oxygen of the A9 ⁄ G101 site in a hammerhead ribozyme J Am Chem... dependence of in vitro selected deoxyribozymes which cleave DNA ⁄ RNA chimeric oligonucleotides J Mol Biol 269, 188–202 Ota N, Warashina M, Hirano K, Hatanaka K & Taira K (1998) Effects of helical structures formed by the binding arms of DNAzymes and their substrates on catalytic activity Nucleic Acids Res 26, 3385–3391 1072 B Nawrot et al 34 Ferrari D & Peracchi A (2002) A continuous kinetic assay... RNA-cleaving deoxyribozymes, exploiting ethidium bromide as an extrinsic fluorescent probe Nucleic Acids Res 30, e112–e120 35 Breaker RR, Emilsson GM, Lazarev D, Nakamura S, Puskarz LJ, Roth A & Sudarsan N (2003) A common speed limit for RNA-cleaving ribozymes and deoxyribozymes RNA 9, 949–957 36 Brautigam CA & Steitz TA (1998) Structural principles for the inhibition of the 3¢-5¢ exonuclease activity of. .. J Am Chem Soc 126, 15504– 15511 39 Xu YZ, Zheng Q & Swamm PF (1992) Synthesis by post-synthetic substitution of oligomers containing guanine modified at the 6-position with S-, N-, O-derivatives Tetrahedron 48, 1729–1740 FEBS Journal 274 (2007) 1062–1072 ª 2007 Centre of Molecular and Macromolecular Studies of the Polish Academy of Sciences Journal compilation ª 2007 FEBS ... 431, 154–160 Suzumura K, Yoshinari K, Tanaka Y, Takagi Y, Kasai Y, Warashina M, Kuwabara T, Orita M & Taira K (2002) A reappraisal, based on (31)P NMR, of the direct coordination of a metal ion with the phosphoryl oxygen at the cleavage site of a hammerhead ribozyme J Am Chem Soc 124, 8230–8236 Zon G & Stec WJ (1991) Oligonucleotides and Analogues: a Practical Approach (Eckstein F, ed.), pp 87– 108 IRL . modification on the catalytic activity of deoxyribozyme 10–23 The functional role of the individual phosphate groups in the catalytic core of deoxyribozyme 10–23. Substrate–enzyme binding occurs via the Watson–Crick mode of base-pairing. The arrow indicates the cleavage site. The positions of the phosphate groups of the catalytic