A triple-helix forming oligonucleotide targeting genomic DNA fails to induce mutation

doi:10.1093/mutage/ges037 Mutagenesis vol. 27 no. 6 pp. 713–719, 2012 Advance Access publication 21 August 2012 A triple-helix forming oligonucleotide targeting genomic DNA fails to induce mutation Reshat Reshat1, Catherine C. Priestley2 and Nigel J. Gooderham,1,* Biomolecular Medicine, Imperial College, London, SW7 2AZ, UK and Genetic Toxicology, AstraZeneca, Macclesfield, Cheshire, SK10 4TG, UK 1 2 *To whom correspondence should be addressed. Biomolecular Medicine, Imperial College London, Sir Alexander Fleming building, London SW7 2AZ, UK. Tel: +0207 594 3188; Fax: +020 7594 3050; Email: Received on May 30 2012; revised on 5 15 2012; accepted May 16 2012 Purine tracts in duplex DNA can bind oligonucleotide strands in a sequence specific manner to form triple-helix structures. Triple-helix forming oligonucleotides (TFOs) targeting supFG1 constructs have previously been shown to be mutagenic raising safety concerns for oligonucleotide-based pharmaceuticals. We have engineered a TFO, TFO27, to target the genomic Hypoxanthine-guanine phosphoribosyltransferase (HPRT) locus to define the mutagenic potential of such structures at genomic DNA. We report that TFO27 was resistant to nuclease degradation and readily binds to its target motif in a cell free system. Contrary to previous studies using the supFG1 reporter construct, TFO27 failed to induce mutation within the genomic HPRT locus. We suggest that it is possible that previous reports of triplex-mediated mutation using the supFG1 reporter construct could be confounded by DNA quadruplex formation. Although the present study indicates that a TFO targeting a genomic locus lacks mutagenic activity, it is unclear if this finding can be generalised to all TFOs and their targets. For the present, we suggest that it is prudent to avoid large purine stretches in oligonucleotide pharmaceutical design to minimise concern regarding off-target genotoxicity. Introduction Purine tracts in duplex DNA can bind third oligonucleotide strands in the major groove to form a triple-helix, or triplex, where the interactions are stabilised by Hoogsteen hydrogen bonds (1,2). Additionally, guanine-based sequences are capable of folding into DNA quadruplex structures (3). To the best of our knowledge, a third oligonucleotide strand has not been shown to facilitate quadruplex formation in duplex DNA. Triplex forming oligonucleotides (TFOs) can be designed to be in the purine motif with an identical sequence and antiparallel orientation relative to the purine target sequence in the duplex. Those in the pyrimidine motif have a complementary sequence and are parallel in orientation relative to the purine target sequence [reviewed in Ref. (1)]. Using engineered reporter constructs, several studies have reported that such triplex structures can induce mutation (4–7). In particular, a 30 nucleotide purine motif TFO targeting a supFG1 based reporter construct was reported to be capable of inducing mutation in the order of 0.27% (13-fold above the spontaneous mutant frequency) (6). A corresponding 20 nucleotide TFO was also reported to be biologically active, but to a lesser extent (3.7-fold increase). Sequencing mutant supFG1 clones revealed mainly point mutations and some small deletions around the triplex target sequence (6). Triplex formation, in the supFG1 based system, was also capable of inducing DNA repair activity and producing truncated transcripts. In accordance with their results, Wang et al. (6) proposed that triplex formation at duplex DNA blocks the progression of a transcription fork triggering gratuitous and error-prone repair. This gratuitous repair would lead to repeated attempts in transcription generating re-iterative repair patches where such hyperactivity may then introduce mutation (8). Repeated cycles of triplex formation, inhibition of transcription and stimulation of repair may increase the probability of a mutagenic event. Triplex formation has also been reported to induce recombination between two tandem supF genes (7). In particular, TFOs that could bind to the triplex target sequence with high affinity and were DNA based were found to be biologically active (7). Those that were RNA based were found to lack biological activity even if they could bind to the triplex target sequence with high affinity. This variation was proposed to be due to the greater helical distorting ability of DNA-based TFOs which may provoke the repair pathway in an attempt to remove the DNA lesion (9). Although antisense oligonucleotides appear to be powerful tools for sequence-specific manipulation of gene expression (10,11), the studies described above present a concern for oligonucleotide-based therapies; specifically, they are suggestive that triplex formation at genomic DNA may lead to heritable sequence alterations. Indeed, the European Medicines Agency (EMA) has raised concern as to whether such oligonucleotide-based pharmaceuticals could present with off-target genotoxicity through triplex formation (12). Furthermore, chemical modifications adopted in oligonucleotide pharmaceutical design to facilitate nuclease resistance, target binding affinity or in vivo distribution may facilitate triplex formation at genomic DNA (11,13–15). We have therefore examined the hypothesis that triplex formation within the coding region of genomic DNA is mutagenic. To test this, we have engineered a TFO, TFO27, to target the genomic coding region of the hemizygous HPRT locus in human lymphoblastoid TK6 cells. We report that TFO27 can bind to its target motif with high affinity in a cell free system and was resistant to nuclease degradation. In contrast to previous studies employing reporter systems, TFO27 failed to induce mutation at its genomic target. We also report that the previously employed triplex target sequence within the supFG1 reporter has the potential for DNA quadruplex formation in the presence of the targeting oligonucleotide and suggest that this latter property, rather than triplex formation, may be the mechanism whereby the reporter system is mutated. Materials and methods All cell culture reagents were obtained from Invitrogen (Paisley, UK) and cell culture consumables from VWR (Leicestershire, UK). All other chemicals/ reagents were obtained from Sigma Aldrich (Poole, UK), unless stated otherwise. © The Author 2012. Published by Oxford University Press on behalf of the UK Environmental Mutagen Society. All rights reserved. For permissions, please e-mail: . 713 R. Reshat et al. Oligonucleotides All oligonucleotides were DNA based and obtained from Sigma Genosys (Haverhill, UK). Those used in mutation studies were chemically modified to contain four terminal phosphorothioate linkages and reverse phase purified. Oligonucleotide sequences are listed in Table I. Electrophoretic mobility shift assay The triplex forming ability of oligonucleotides was assessed using a 35 bp duplex containing the HPRT triplex target sequence. The quadrup (...truncated)