Synthesis, thermal stability and reactivity towards 9-aminoellipticine of double-stranded oligonucleotides containing a true abasic site

Effects of (5'S)-5',8-cyclo-2'-deoxyadenosine on the base excision repair of oxidatively generated clustered DNA damage. A biochemical and theoretical study

The presence of 5',8-cyclo-2'-deoxyadenosine (5'S)-cdA induces modifications in the geometry of the DNA duplex in the 5'-end direction of the strand and in the 3'-end direction of the complementary strand. As a consequence, the enzymes are probably not able to adjust their active sites in this rigid structure. Additionally, clustered DNA damage sites, a signature of ionising radiation, pose a severe challenge to a cell's repair machinery, particularly base excision repair (BER). To date, clusters containing a DNA base lesion, (5'S)-cdA, which is repaired by nucleotide excision repair, have not been explored. We have therefore investigated whether bistranded clusters containing (5'S)-cdA influence the repairability of an opposed AP site lesion, which is repaired by BER. Using synthetic oligonucleotides containing a bistranded cluster with (5'S)-cdA and an AP site at different interlesion separations, we have shown that in the presence of (5'S)-cdA on the 5'-end side, repair of the AP site by the BER machinery is retarded when the AP site is ≤8 bases from the (5'S)-cdA. However, if (5'S)-cdA is located on the 3'-end side with respect to the AP site, the effect on its repair is much weaker and totally disappears for distances ≥8 bases.

Improved measurement of dibenzo[a,l]pyrene-induced abasic sites by the aldehyde-reactive probe assay

Dibenzo[a,l]pyrene (DB[a,l]P) induces abundant amounts of depurinating adducts that spontaneously dissociate to form abasic sites in DNA. However, several previous studies that used the aldehyde-reactive probe (ARP) assay, could not verify abasic site formation by DB[a,l]P. Therefore, we examined whether a modification of the ARP assay would allow greater quantification of abasic sites. A previous study indicated that the abasic site quantification is improved by letting ARP trap the nascent abasic sites in cells, before extracting DNA for the assay. To test whether the addition of ARP to the DB[a,l]P-DNA adduct-forming reaction would improve abasic site quantification, we treated calf thymus DNA (0.625 mg/mL) with DB[a,l]P (80 microM) and 3-methylcholanthrene-treated rat liver microsomes with or without ARP (3 mM). The inclusion of ARP in the adduct-forming reaction resulted in significantly greater detection of abasic sites (62 lesions/10(6) bp versus 3.7 lesions/10(6) bp). DB[a,l]P also induces DNA strand breaks. The strand breaks may occur at abasic sites and by other mechanisms, such as oxidative damage. ARP/O-methoxyamine-abasic site conjugates are refractory to strand breakage, however, ARP or O-methoxyamine (3-10 mM) could only partially protect DB[a,l]P-induced DNA degradation, presumably by protecting the abasic sites, but not the other strand breaks. These results suggest that if DNA strand breakages occur at the abasic sites or at bases flanking them, and the fragments are lost during DNA extraction, abasic site estimation could be compromised. To obtain an independent line of evidence for abasic site formation in DB[a,l]P-treated cells, mouse Mbeta16 fibroblasts were treated with DB[a,l]P and O-methoxyamine. O-Methoxyamine is known to potentiate cytotoxicity of abasic site-inducing chemicals by forming abasic site conjugates, which partially inhibits their repair. O-Methoxyamine was found to increase DB[a,l]P cytotoxicity in these cells, supporting the idea that DB[a,l]P formed abasic sites. In summary, the inclusion of ARP in the DB[a,l]P-DNA adduct-forming reaction traps and protects the nascent abasic sites, allowing an improved quantification of abasic sites.

Peroxide-mediated desulfurization of phosphorothioate oligonucleotides and its prevention

Desulfurization at the internucleotide phosphorothioate linkage of antisense oligonucleotides (ASOs) in dermatological formulations has been investigated using strong ion exchange chromatography and mass spectroscopy. The formation of phosphate diester linkages appeared to arise from a reaction between the phosphorothioate oligonucleotide and a potent oxidizing agent. Screening of excipients used in the formulation indicated that the cause of desulfurization was related to the presence of polyethylene glycol-derived nonionic surfactants MYRJ 52 or BRIJ 58. Autoxidation of the polyethylene glycol chain is suggested as the probable origin for the observed incompatibility. The ability of various antioxidants to prevent oxidative degradation of ASO-1 in simple test systems and in oil-in-water emulsions is described. It is found that in test systems both lipophilic and hydrophilic antioxidants are effective. However, in cream formulation (oil-in-water emulsions) of ASO-1 the addition of hydrophilic antioxidants L-cysteine or DL-alpha-lipoic acid has been shown to be superior in protecting the oligonucleotide from desulfurization upon storage.

Preparation of oligonucleotides without aldehyde abasic sites

High-quality oligonucleotides are obtained by selective modification of sequences containing aldehyde apurinic sites with a new chromatographic tag followed by RP-HPLC separation. Hydroxylamine derivative 1 of a water soluble nonionic surfactant modifies oligonucleotides selectively at abasic sites leading to significantly increased retention.

Rapid routes of synthesis of oligonucleotide conjugates from non-protected oligonucleotides and ligands possessing different nucleophilic or electrophilic functional groups

Optimized methods are described for post-synthetic conjugation of non-protected oligodeoxyribonucleotides to different ligands. Methods for the terminal functionalization of oligonucleotides by amino, sulfhydryl, thiophosphate or carboxyl groups using different chemical reactions and linkers in both organic and aqueous media are described and compared. Experimental conditions for subsequent coupling of ligands containing aliphatic and aromatic amines, aromatic alcohols, carboxylic, sulfhydryl, alkylating, aldehydic and other reactive nucleophilic and electrophilic groups to oligonucleotides were established, including covalent linkage to other oligonucleotides.

Abasic DNA structure, reactivity, and recognition

Loss of a base in DNA, i.e., creation of an abasic site leaving a deoxyribose residue in the strand, is a frequent lesion that may occur spontaneously, or under the action of radiations and alkylating agents, or enzymatically as an intermediate in the repair of modified or abnormal bases. The abasic site lesion is mutagenic or lethal if not repaired. From a chemical point of view,the abasic site is an alkali-labile residue that leads to strand breakage through beta- and delta- elimination. Progress in the understanding of the chemistry and enzymology of abasic DNA largely relies upon the study of synthetic abasic duplexes. Several efficient synthetic methods have thus been developed to introduce the lesion (or a stable analogue) at defined position in the sequence. Physicochemical and spectroscopic examination of such duplexes, including calorimetry, melting temperature, high-field nmr and molecular modeling indicate that the lesion strongly destabilizes the duplex, although remaining in the canonical B-form with structural modifications strictly located at the site of the lesion. Probes have been developed to titrate the damage in DNA in vitro. Series of molecules have been devised to recognize specifically the abasic site, exhibiting a cleavage activity and mimicking the AP nucleases. Others have been prepared that bind strongly to the abasic site and show promise in potentiating the cytotoxic and antitumor activity of the clinically used nitrosourea (bis-chloroethylnitrosurea).

The acridine ring selectively intercalated into a DNA helix at various types of abasic sites: double strand formation and photophysical properties

The interactions between the intercalating agent and the three types of abasic sites: abasic frameshift, apurinic and apyrimidinic, were investigated. 9-amino-6-chloro-2-methoxyacridine (ACMA), whose spectroscopic properties are strongly perturbed by the environment, was selected as the intercalating agent. The optically pure threoninol derived from the reduction of L-threonine was used as an artificial abasic site mimicking the ring-opened natural ribose. In order to secure the selective intercalation to the adjacent abasic site, ACMA and the abasic site were connected through a tri- pentamethylene linker. These modified oligonucleotides covalently linked to an ACMA molecule at the internucleotide site having the same base-sequence were synthesized using the acridine-phosphoramidites. Although all the modified oligonucleotides lack a nucleobase at the intervening position, these double strands showed high thermal stability. The pentamethylene linker and the apyrimidinic systems were especially stabilized. At the same time, sharpness of the absorption spectra and a new fluorescent band of the acridine, due to the fixation of the environment around ACMA, were observed. Therefore, it is concluded that the acridine binds preferentially to the apyrimidinic site rather than the frameshift abasic site and that the surroundings of the acridine are strictly fixed at the microenvironmental level.

On the mechanism of preferential incorporation of dAMP at abasic sites in translesional DNA synthesis. Role of proofreading activity of DNA polymerase and thermodynamic characterization of model template-primers containing an abasic site

DNA polymerase preferentially incorporate dAMP opposite abasic sites (A-rule). The mechanism of the A-rule can be studied by analyzing three dissected stages of the reaction including (i) initial nucleotide insertion, (ii) proofreading excision of the inserted nucleotide and (iii) extension of the nascent primer terminus. To assess the role of the stage (ii) in the A-rule, kinetic parameters of the proofreading excision of primer terminus nucleotides opposite abasic sites were determined using E.coli DNA polymerase I Klenow fragment. The relative efficiency of the excision (Vmax/Km) revealed that removal of A was the least favored of the four nucleotides, but the differences in the efficiencies between excision of A and the other nucleotides was less than 2-fold. In addition, in an attempt to reconcile kinetic data associated with the stage (i) or (ii), the differences in free energy changes (delta delta G degrees) for the formation of model template-primer termini containing XN pairs (X = abasic site, N = A, G, C or T) were determined by temperature dependent UV-melting measurements. The order of delta delta G degrees was XG > XA = XC > or = XT, with delta delta G degrees being 0.5 kcal/mol for the most stable XG and the least stable XT. Based on these data, the role of the stage (ii) and energetic aspects of the A-rule are discussed.

High-field NMR and restrained molecular modeling studies on a DNA heteroduplex containing a modified apurinic abasic site in the form of covalently linked 9-aminoellipticine

Two-dimensional NMR methods were used to model the possible solution structure of an intercalative complex of 9-aminoellipticine (Aell), a polycyclic pyridocarbazolamine, covalently bound to an apurinic ring-opened deoxyribose site of a duplex DNA fragment in the reduced Schiff base form. The required oligonucleotide single strand containing covalently attached aminoellipticine was obtained by reductive amination in the presence of sodium cyanoborohydride. The combined NMR-energy minimization methods were employed to refine the model structures of two distinct forms, intrahelical and extrahelical, of a control 9-mer duplex DNA, d(CGTG.dr.GTGC).d(GCACTCACG), which contains an apurinic site positioned opposite a dT residue on the complementary strand. The model structure of an aminoellipticine conjugate with the same DNA sequence, derivatized via the aforementioned covalent attachment, was also obtained by incorporating intermolecular drug-DNA and intra- and internucleotide NOE-derived proton-proton distance estimates as restraints in energy minimization routines. The indole ring system of aminoellipticine, which is inserted at the apurinic site, intercalates between and is parallel to flanking GC base pairs. The pyridinic ring of aminoellipticine, in protonated form, also stacks between cytidine and thymidine bases on the complementary strand, which is consistent with the observation that the normal sequential NOE connectivity at the 5'-C13-T14 step is broken and indeed diverted through the ellipticine moiety, e.g., C13-Aell-T14 connectivities through the Aell-H4/C5Me protons. Interestingly, the partial stacking of the pyridinic ring is observed only between the 5'-CT step vs an adjacent 5'-TC step, owing to inherently weak stacking interactions associated with the former. In the absence of any potential groups that can participate in electrostatic or hydrogen-bonding interactions with the nucleic acid, pi-pi stacking and hydrophobic contacts at the intercalation site appear to be the important factors in determining stability and conformation of the aminoellipticine-DNA conjugate. Stacking interactions in such a bistranded intercalative complexation of aminoellipticine apparently govern the formation of a single intrahelical form of a right-handed B-type DNA duplex. The overall structural features lead us to propose working models for an enzyme-like DNA cleavage activity of 9-aminoellipticine and the observed inhibition of the AP endonuclease-dependent DNA excision-repair pathway.

Implication of uracil in spontaneous mutagenesis on a single-stranded shuttle vector replicated in mammalian cells

Almost all spontaneous point mutations found on a single-stranded shuttle vector after its transfection and replication in monkey cells were located at cytosine residues. In order to understand this very specific type of targeting we have studied the possible implication of uracil residues in the induction of these spontaneous mutations. The single-stranded shuttle vector pCF3A carrying the supF tRNA gene as a mutagenesis target has been allowed to replicate in mammalian COS7 cells, mutations being screened in bacteria using the beta-galactosidase assay. Progenies from untreated DNA and DNA treated with the uracil-DNA glycosylase prior to transfection were analyzed to determine the amount and classes of mutations. While spontaneous mutation frequency was 9.7 x 10(-4) for control DNA, single-stranded vector treated with the E. coli uracil-DNA glycosylase exhibited a reduced mutation frequency of about 30%. The abolished mutations were mainly confined to the cytosine to thymine transitions for which a decrease by a factor of 5 was indeed observed. This finding fits well with the fact that it is usually admitted that uracil pairs with adenine, indicating therefore that approximately 30% of spontaneous mutations observed in our experimental conditions and 80% of C to T transitions may be due to the presence of uracil instead of cytosine.

Participation of glutamic acid 23 of T4 endonuclease V in the beta-elimination reaction of an abasic site in a synthetic duplex DNA

T4 endonuclease V catalyzes the hydrolysis of the glycosyl bond of a thymine dimer in a DNA duplex and the cleavage of the 3'-phosphate by beta-elimination. We have previously identified a catalytic site for the first reaction (pyrimidine dimer-glycosylase activity) by systematic mutagenesis (Doi et al. Proc. Natl. Acad. Sci. USA 1992 in press) and by x-ray crystallography (Morikawa et al. Science, 256: 523-526, 1992). The results showed that replacement of Glu23 with either glutamine or aspartic acid completely abolished the glycosylase activity. We describe the investigation of the second reaction (apurinic/apyrimidinic endonuclease activity), using twenty two mutants of T4 endonuclease V plus a DNA mini duplex containing an abasic site. Replacement of Glu23 by glutamine abolished the second reaction, but replacement with aspartic acid did not. The pH optima of the mutant (23 Asp) and the wild type were found to be 5.0 and 5.5, respectively. We conclude that the carboxylate anion in position 23 may act as a general base in the beta-elimination reaction of the endonuclease.

Mutagenicity of a unique apurinic/apyrimidinic site in mammalian cells

Abasic sites are common DNA lesions produced either spontaneously or as a consequence of the action of some genotoxic agent. The mutagenic properties of a unique abasic site replicated in mammalian cells have been studied using a shuttle vector. A plasmid, able to replicate both in mammalian cells and in bacteria, carrying a unique abasic site chemically synthesized has been constructed. After replication in mammalian cells, plasmid DNA was recovered and used to transform bacteria. Mutants were screened without selection pressure by differential hybridization with a labelled oligonucleotide and their DNA was sequenced. A mutation frequency ranging from 1% to 3% was found, depending on the base originally inserted during the vector construction, opposite the abasic site. All the sequenced mutants correspond to single base-pair substitutions targeted at the abasic site. We observed a deficit in guanine incorporation opposite the abasic site, while the three other bases were incorporated with a similar efficiency. The mutational potency of abasic sites was observed without any voluntary preconditioning treatment of mammalian cells in order to induce "SOS" like conditions.

9-Aminoellipticine-derivatized alpha- and beta-oligodeoxyribonucleotides targeted to the cap of beta-globin mRNA: hybridization to natural and engineered mRNA, inhibition of translation, and improved effect of tandem chains

We studied the duplex stability and the antimessenger activity of 9-aminoellipticine-5'-functionalized alpha- and beta-anomeric DNA sequences complementary to the first 14 nucleotides of the rabbit beta-globin mRNA. The duplex formed by the beta-conjugate with the natural mRNA target possessed a marginally better stability to that of the duplex formed by the unfunctionalized compound, as measured by the thermal elution. The alpha-conjugate did not anneal to native mRNA, possibly due to the interference of the 9-aminoellipticine with the cap structure and, unlike the beta-adduct, was practically inactive as inhibitor of translation in a cell-free system. However, it did hybridize to an RNA construction containing the beta-globin mRNA plus an additional 50 bases in 5'. Surprisingly, translation from this construction was inhibited by the alpha-species in spite of the nonvicinity of the target to the cap. Both alpha and beta conjugates hybridized to a DNA 14-mer of the same sequence as that targeted onto the mRNA. Thermal denaturation and fluorescence spectroscopy showed that the drug brought no considerable stabilization to the duplex, the linker presumably being unfavorable to intercalation. An increased stability of the complex and a higher inhibitory effect on cell-free beta-globin translation were obtained with two contiguous beta-oligomers of which one was functionalized.

Molecular modelling of 9-aminoellipticine interactions with abasic oligonucleotides

We have used molecular mechanics to study the insertion of the DNA intercalating agent 9-aminoellipticine (9AE) into single and double stranded abasic oligonucleotides containing abasic sites in the aldose or furanose conformations. 9AE-abasic oligonucleotide complexes have been considered with 9AE bound at abasic sites as a covalent complex, a reversible complex or a Schiff base. Results are in good agreement with experimental data available on abasic oligonucleotides (melting temperature measurement, NMR results) and allow an analysis of different possible structures for 9AE-abasic oligonucleotide complexes. Hypotheses concerning the role of 9AE-abasic site complexes in enzymatic inhibition are formulated from these data.

Mutagenic properties of a unique abasic site in mammalian cells

The mutagenic properties of a true unique abasic site located opposite a guanine residue were studied. An oligonucleotide containing a chemically-produced abasic site was inserted into a shuttle vector able to replicate both in simian cells and in bacteria. Plasmid DNA was rescued from simian cells and screened in bacteria by differential hybridization with a labelled oligonucleotide probe. Mutations were easily detected and sequenced. Results showed that opposite a guanine the abasic site was error free repaired or replicated by mammalian cells with an efficiency of 99%. Point mutations occurred at a frequency of approximately 1% in control host cells and at more than 3% in UV-pre-irradiated host cells. Adenine, cytosine or thymine were found to have been inserted opposite the abasic site. No preferential insertion for a particular base was observed in contrast to that reported in bacteria.