Site-Specific Cleavage of RNA by a Metal-Free Artificial Nuclease Attached to Antisense Oligonucleotides - Oligonucleotides

Site-Specific Cleavage of RNA by a Metal-Free Artificial Nuclease Attached to Antisense Oligonucleotides
Received February 13, 2006
Web Release Date: May 25, 2006
Claudio Gnaccarini, Sascha Peter, Ute Scheffer, Stefan Vonhoff, Sven Klussmann, and Michael W. G?bel*
J. Am. Chem. Soc.
ACS Publications
Copyright ? 2006 American Chemical Society
Contribution from the Institute for Organic Chemistry and Chemical Biology, Goethe University Frankfurt, Marie-Curie-Str. 11, D-60439 Frankfurt am Main, Germany, and NOXXON Pharma AG, Max-Dohrn-Str. 8-10, D-10589 Berlin, Germany
m.goebel@chemie.uni-frankfurt.de
Abstract:
RNA cleaving tris(2-aminobenzimidazoles) have been attached to DNA oligonucleotides via disulfide or amide bonds. The resulting conjugates are effective organocatalytic nucleases showing substrate and site selectivity as well as saturation kinetics. The benzimidazole conjugates also degrade enantiomeric RNA. This observation rules out contamination effects as an alternative explanation of RNA degradation. The pH dependency shows that the catalyst is most active in the deprotonated state. Typical half-lifes of RNA substrates are in the range of 12-17 h. Thus, conjugates of tris(2-aminobenzimidazoles) can compete with the majority of metal-dependent artificial nucleases.
--------------------------------------------------------------------------------
Tris(2-aminobenzimidazoles) 1 and 2 (Chart 1) have been recently identified as powerful cleavers of ribonucleic acids.1 The reaction involves nucleophilic attack by the 2'-hydroxy groups at phosphorus and the formation of 2',3'-cyclic phosphates. Thus, catalysts 1 and 2 are completely specific for RNA and do not hydrolyze DNA.1 Unfortunately, tris(2-aminobenzimidazoles) tend to aggregate in a pH-dependent way, thereby preventing deeper mechanistic insight from pH-rate correlations and similar experiments. For future applications, it is also required to determine the catalytic efficiency in the nonaggregated state. This issue has not yet been fully resolved. Here we describe the conjugation of catalyst 2 with a series of DNA oligonucleotides. The resulting conjugates 4-8 efficiently cleave complementary RNA strands at submicromolar concentrations in the expected positions. They do not aggregate and allow to complete the characterization of benzimidazole derived RNA cleavers.
-------------------------------------------------------------------------------
Chart 1. Sequences of Oligonucleotides 4-13. For Complete Linker Structures, See Scheme 1
--------------------------------------------------------------------------------
A standard procedure to synthesize DNA conjugates involves acylation of amino linkers by active esters in aqueous buffer. When applied to the charged benzimidazoles 2 or 3, however, complete precipitation of the oligonucleotide prevented coupling. To avoid the aggregation of polyionic species, we used two conjugation methods utilizing protected DNA still bound to the solid support (Scheme 1). In the first method, a trityl-protected mercapto linker was attached to the 5'-end as a phosphoramidite building block (14). After removal of trityl, the support was incubated with compound 15. A disulfide linkage of sufficient stability to survive the subsequent deprotection steps was formed (for details, see Supporting Information).2 Fractions of 60-70% of conjugated versus nonconjugated strands could be typically obtained. Pure conjugates 4 and 6-8 were finally isolated by polyacrylamide gel electrophoresis and characterized by mass spectroscopy. Alternatively, we attached a monomethoxytrityl-modified amino linker to the DNA strand that was deprotected and coupled with acid 3 in the presence of DIC and HOBt. The resulting peptide bond of conjugate 5 was stable under the conditions of DNA deprotection (coupling yield ~ 70%).
--------------------------------------------------------------------------------
Scheme 1. Synthesis of Oligonucleotide Conjugates
--------------------------------------------------------------------------------
All cleavage experiments were run with the RNA substrates 9-11. The fluorescent label allows convenient detection and quantification of products by gel electrophoresis in a DNA sequencer. Note that the sequence of oligonucleotide 10 is analogous to that of 9. However, it is a hybrid formed of DNA (natural configuration) with an embedded part of enantiomeric RNA.
To exclude the existence of molecular aggregates of aminobenzimidazole conjugates with noncognate oligonucleotides, the dye-labeled DNA 12 (25 nM, diluted with 175 nM unlabeled DNA 13) was studied by fluorescence correlation spectroscopy (FCS).1,3 The diffusion time around 150 s at 24 C is consistent with the size of 12. Upon addition of conjugates 4-8 (1.5 M), no change in diffusion times and no signs of aggregation could be observed. Substrate 9 (200 nM) was then mixed with complementary conjugates 4-7 (1.5 M), again without producing significant effects. The gain in molecular mass by hybridization was not expected to be sufficient to influence the diffusion time of 9. Further experiments with substrate 11 and conjugate 8 confirmed the general view that higher aggregates of oligonucleotides beyond the stage of hybridization will not occur in the experiments shown below.
When substrate 94 was incubated with the 15mer conjugate 5 (1.5 M) for 20 h, strong cleavage occurred that could be assigned to nucleotides 13, 14, and 15 by comparison with the hydrolysis pattern of 9 (Figure 1, lanes a and e). Almost identical results were obtained in analogous experiments using the disulfide conjugate 4 (lane b). In contrast, the preferred cleavage sites of the 17mer conjugate 6 and 20mer 7 are nucleotides 17 and 19/20, respectively (lanes c and d). The reaction site thus correlates with the position of the benzimidazole moiety in each substrate-catalyst duplex. The preferential formation of 14mer fragments in the reactions of 4 and 5 may be a consequence of fraying due to the weak terminal A-T base pair. No significant turnover could be observed when 4 (150 nM) was incubated with a 10-fold excess of substrate 9. This is a common observation seen with most RNA cleaving DNA conjugates when the catalyst is attached to either the 5'- or 3'-ends. Addition of EDTA (1 mM) does not change the kinetics and cleavage patterns of these reactions. A mechanistic role of contaminating metal ions bound to the benzimidazole moiety therefore can be ruled out.
--------------------------------------------------------------------------------
Figure 1 Cleavage of RNA 9 (150 nM) induced by conjugates 4-7 at concentrations of 1.5 M (50 mM Tris-HCl, pH 8.0, 37 C, 20 h). Lane e: hydrolysis pattern of substrate 9 (Na2CO3).
--------------------------------------------------------------------------------
The influence of pH on catalyst efficiency was studied next (see Table 1). Since ionic interactions might play a major role in the benzimidazole phosphate interaction, experiments initially were conducted in Tris-HCl buffer (50 mM) without adding inert salts. Upon variation of pH under such conditions, ionic strength is not constant. Critical experiments therefore were repeated in the presence of 100 mM NaCl. Cleavage yield increases with pH and levels off at pH 8. At high ionic strength, a slight drop at pH 9 was observed. Since the pKa of related benzimidazoles comes close to 7, the catalyst should be mostly deprotonated under such conditions. Thus, general base catalysis must be an important mechanistic aspect.
--------------------------------------------------------------------------------
Figure 2 Cleavage of RNA substrates 9 and 11 as a function of conjugate concentrations (150 nM 9 or 11, 20-1500 nM of conjugates, 50 mM Tris-HCl, pH 8.0, incubation for 20 h). Data points are connected by lines for the sake of clarity.
--------------------------------------------------------------------------------
An important requirement for artificial nucleases is substrate specificity. Figure 3 shows a cross reactivity experiment with two nonhomologous RNA substrates 9 and 11 challenged by the conjugates 4 and 8. Cleavage fully depends on complementarity. While both catalysts transform their cognate substrates (lanes a and c), they do not react with noncomplementary RNA (lanes b and d). All other combinations of substrates 9-11 and catalysts 4-8 have been tested as well. The results consistently fit into the picture (data not shown).
--------------------------------------------------------------------------------
Figure 3 Substrate specificity of conjugates 4 and 8 (150 nM substrate, 1.5 M catalyst, 50 mM Tris-HCl, pH 8.0, incubation for 20 h). Lane a: conjugate 4 and substrate 9. Lane b: conjugate 4 and substrate 11. Lane c: conjugate 8 and substrate 11. Lane d: conjugate 8 and substrate 9. Lanes e and f: hydrolysis patterns of 9 and 11 (Na2CO3).
--------------------------------------------------------------------------------
Cleavage kinetics was studied in detail for the reaction of conjugate 8 with substrate 11 (pH 8, high ionic strength). Without catalyst, the substrate proved almost stable for several days (Figure 4, lane b; Figure 5; Figure S8). In contrast, conjugate 8 cleaves significantly within a few hours. The substrate is almost completely degraded after 56 h. A small percentage, however, does not react even after 1 week (Figure 5). This may result from structural damages in chemically synthesized RNA preventing hybridization with conjugate 8 (e.g., 2',5' linkages, residual protective groups, etc.). Interestingly, a more defined cleavage pattern arises after longer incubation times (Figure 4, lane d; see also lane c in Figure 3 and Figure S7). Upon standing, secondary cleavage events may cut off all conformationally mobile ribonucleotides protruding out of the duplex with conjugate 8.
---------------------------------------------------------