Specific Mutations Induced by Triplex-Forming Oligonucleotides in Mice

See allHide authors and affiliations

Science  20 Oct 2000:
Vol. 290, Issue 5491, pp. 530-533
DOI: 10.1126/science.290.5491.530


Triplex-forming oligonucleotides (TFOs) recognize and bind to specific duplex DNA sequences and have been used extensively to modify gene function in cells. Although germ line mutations can be incorporated by means of embryonic stem cell technology, little progress has been made toward introducing mutations in somatic cells of living organisms. Here we demonstrate that TFOs can induce mutations at specific genomic sites in somatic cells of adult mice. Mutation detection was facilitated by the use of transgenic mice bearing chromosomal copies of the supF and cII reporter genes. Mice treated with a supF-targeted TFO displayed about fivefold greater mutation frequencies in the supF gene compared with mice treated with a scrambled sequence control oligomer. No mutagenesis was detected in the control gene (cII) with either oligonucleotide. These results demonstrate that site-specific, TFO-directed genome modification can be accomplished in intact animals.

By inducing site-specific mutations in the genome, heritable changes can be achieved in gene function and expression. Triplex formation, through recognition of a specific region of duplex DNA by a single-stranded TFO, offers an attractive strategy to modify a mammalian genome. TFOs have so far been used successfully to modify genes and gene function in cells in vitro (1). With a view toward a therapeutic application, we have investigated the potential of a triplex approach to target specific mutations in somatic cells of mice in vivo. Transgenic mice (C57BL/6 mice containing multiple copies of a chromosomally integrated λsupFG1 vector, designated 3340) were previously generated (2) containing a 30–base pair (bp) triplex-target site within thesupFG1 mutation–reporter gene. We previously demonstrated TFO-targeted mutations in the supFG1 gene in vitro on intracellular plasmid targets and on a chromosomal locus in a fibroblast cell line established from these mice (3–5). The results revealed 10- to 100-fold induction of site-specific mutations in cells treated with the specific TFO. Although psoralen conjugation to the TFO, coupled with UVA (ultraviolet, long wave) irradiation, generally increased the frequency of mutations, targeted mutagenesis was seen even without psoralen photoproduct generation, suggesting a substantial triplex-mediated process of mutagenesis. On the basis of these results, the oligonucleotides in this work were not conjugated to psoralen, and no other DNA-damaging agent (or mutagen) was required to induce mutagenesis. This affords an advantage in that UVA activation is not required, and there is no apparent toxicity to the animals.

Mice were given daily intraperitoneal (i.p.) injections with 1 mg day−1 of either AG30 or SCR30 for five consecutive days. AG30 is the specific TFO designed to bind to the polypurine site in thesupFG1 gene, whereas SCR30 is a control oligonucleotide with the same base composition as AG30 but with a scrambled sequence that differs at 12 positions (6). Using gel mobility shift assays and deoxyribonuclease I (DNase I) footprinting, we found that AG30 binds with high affinity (equilibrium dissociation constant ∼10−9 M) and specificity to the target site, whereas SCR30, having mismatches in the triplex binding code, does not. Ten days after injection, mouse tissues were collected for mutation analysis (7). The combined mutation frequencies in a variety of tissues from AG30-treated mice were increased by about fivefold compared with tissues from SCR30-treated mice (Table 1). The mean mutation frequencies in liver, skin, kidney, colon, small intestine, and lung from AG30-treated mice were significantly higher than those from SCR30-treated mice (P values calculated by the Student's t test are listed in Table 1). These data demonstrate an oligonucleotide-specific induction of mutagenesis in mice. Additionally, these data indicate efficient tissue uptake and distribution of oligonucleotides in mice after i.p. injections.

Table 1

Targeted mutagenesis of the supFG1 gene in 3340 mice by TFOs. Mice were injected i.p. with 1 mg of TFO per day for 5 days (∼50 mg kg−1 day−1). AG30 is the specific TFO, SCR30 is the control TFO. Values of mutation frequency for skin, liver, and kidney are reported as the mean ± SD. For colon, small intestine, and lung, values are reported as the mean ± the difference between the two values. P values were calculated by the Student's t test.

View this table:

In a previous study, the tissue uptake and distribution after i.p. administration of G-rich, propylamine end-capped, phosphodiester oligonucleotides (similar to AG30 and SCR30) was examined (8). This study demonstrated substantial uptake in a number of tissues except for the brain. Because oligonucleotides do not efficiently cross the blood-brain barrier after i.p. injection, mutagenesis in brain tissue was measured as an internal control. Consistent with this, mutagenesis was not induced above background levels in brain tissues analyzed from mice treated with either AG30 or SCR30 (Fig. 1A). In contrast, all other tissues tested from AG30-treated animals showed an average fivefold increased induction in mutagenesis compared with tissues from SCR30-treated animals (Fig. 1A).

Figure 1

(A) Targeted mutagenesis of the supFG1 gene in somatic tissues of 3340 mice. The combined totals from all tissues tested, except brain, are displayed in the first set, with brain only in the second set. The number of mutants/total plaques is listed above each bar. UT, untreated mice. (B) Mutations in the supFG1 gene in tissues from mice treated with AG30. The 30-bp triplex target site is underlined. The supFG1 gene sequences were determined as described (5). Base changes are indicated above the sequence. (+) and (−) represent single-base insertions or deletions.

The levels of mutagenesis obtained from mice injected with SCR30 were similar to those observed in mice injected with phosphate-buffered saline (PBS) only (Fig. 1A), showing that SCR30, even at a dose of 1 mg day−1 for 5 days, does not have any detectable mutagenic effect in the animals. The values for both the SCR30- and PBS-injected animals, furthermore, were essentially the same as those observed in untreated animals in the same 3340 lineage (9) (Fig. 1A). These results not only provide additional evidence that AG30-mediated mutation induction is occurring through a sequence-specific, triplex-mediated mechanism, but also indicate that nonspecific oligonucleotides are not generally mutagenic in animals.

To further confirm that the induced mutagenesis obtained from AG30 treatment resulted from a triple-helix–dependent event and not from some hypothetical, nonspecific effect of AG30 on DNA metabolism, we tested the effect of AG30 and SCR30 in the Mutamouse mutagenesis model (10). We chose to use the lambda cII gene as a mutation reporter, because this locus has been well studied (11) and does not contain the AG30 triplex target site. ThecII control gene showed no induction of mutagenesis in animals treated with either AG30 or SCR30 compared with background levels (Table 2). These results rule out any nonspecific mutagenic effect of AG30 and are consistent with a gene-specific, triplex-mediated effect of AG30 in inducing an increased level of mutagenesis in the supFG1 gene of 3340 mice.

Table 2

Lack of TFO-induced mutagenesis of the cIIgene in Mutamice. Mice were injected i.p. with 1 mg of oligonucleotide (either AG30 or SCR30) per day for 5 days (∼50 mg kg−1 day−1) or with PBS.

View this table:

Mutant plaques were isolated from a variety of tissues from AG30-treated animals, and the supFG1 genes were analyzed by DNA sequencing. The sequences of supFG1 mutations from a number of AG30-treated mouse tissues are listed in Fig. 1B. The results are consistent with a TFO-directed effect with the majority (85%) of the mutations concentrated within the 30-bp triplex target site. Nearly 40% of the mutations are single-base insertions or deletions in a stretch of eight contiguous Gs. It is possible that these mutations resulted from slippage errors during TFO-induced repair, because this site has been shown previously to be prone to slippage events (2, 5).

Indeed, the genome sites conducive to high-affinity triple-helix formation (consisting of polypurine sequences with segments of mononucleotide repeats) are precisely the kinds of sequences that are hot spots for insertion and deletion mutagenesis in a template dislocation and slippage mechanism, as proposed by Streisinger (12). As such, the 30-bp polypurine triplex target site in supFG1 is an “at-risk motif” (13) that is a hot spot for both spontaneous and induced mutagenesis. Even in control animals, the 30-bp, G-rich polypurine site is a hot spot for mutations, with about 50% of spontaneous mutations occurring in that site, and even more in the absence of DNA mismatch repair (2). However, specific TFO treatment increases the proportion of mutations at this site and also increases the absolute mutation frequency at this site more than fivefold, in a highly sequence-specific manner. Hence, the hypermutability of G-rich (or A-rich) sites is a property that actually enhances the utility of TFOs that preferentially bind to such sites in efforts directed at somatic gene knockout.

In a previous study of directed mutagenesis by psoralen-conjugated TFOs within the hprt gene in Chinese hamster ovary cells (14), mostly large deletions rather than single-base insertions or deletions were seen. However, in that study, the polypurine target site was situated within an intron, and so the type of slippage mutations we observed would have had no effect on the coding region and so would not have been detectable. In another study where psoralen-conjugated TFOs were used to target mutations in yeast, only single-base substitution mutations within a specific ochre codon were seen, rather than slippage events (15). However, the yeast study made use of an assay requiring specific reversion of the ochre stop codon, instead of a forward mutation assay, and so neither slippage mutations nor base substitutions elsewhere in the target site would have been observable.

We have investigated the potential of a triplex-based strategy to target mutations to a specific gene in the somatic tissues of mice. The results demonstrate substantial induction of mutagenesis in a variety of tissues with a TFO (AG3O) designed to bind to the chromosomal triplex target site in transgenic animals. We detected no induction of mutagenesis with either AG30 or SCR30 in the supFG1 gene in brain tissue, or in the cII control gene lacking the AG30 triplex target site. Additionally, no induced mutagenesis was detected in the target gene in any of the tissues tested with the control oligonucleotide, SCR30. These findings demonstrate the successful application of a site-specific, DNA-binding reagent to specifically induce genome modifications on a chromosomal target in intact animals.

Although the mechanism of the triplex-directed mutagenesis is not fully established, evidence suggests that the triplex structure can invoke the nucleotide excision repair machinery, either directly or by transcription-coupled repair (16). This possibility is consistent with results of previous studies that demonstrate a lack of triplex-induced mutagenesis or recombination in nucleotide excision repair–deficient cells (4, 17). The single-base insertions and deletions that were detected in the triplex target site in mice may therefore be a result of template misalignment in association with TFO-induced repair of this “at risk” site (13).

The ability to induce mutations in living animals by means of TFOs affords a powerful strategy to modify the genome. In addition, recent developments in oligonucleotide technology have advanced the potential of gene medicines for other approaches, such as specifically binding to RNA [antisense and ribozymes (18,19)], DNA [antigene TFOs, peptide nucleic acids, polyamides, and RNA/DNA chimeras (1, 20–22)], and protein [aptamer oligomers (23)] targets. Our data showing a lack of mutagenesis by the scrambled sequence control (SCR30) are promising for these other nucleic acid–based strategies because nonspecific oligonucleotides do not appear to be mutagenic.

In addition, for gene-targeting technologies to provide a therapeutic benefit, oligonucleotides must be shown to be capable of binding specifically to chromosomal targets in intact animals. The in vivo distribution of oligonucleotides has been studied with a variety of modified oligonucleotides in mice (8,24, 25). These studies established that small DNA molecules can be administered by i.p. or intravenous injections and gain access to tissues (outside the central nervous system) and to cell nuclei. The work presented here extends these studies by demonstrating that chromosomal DNA throughout the somatic tissues of an animal can be targeted by nucleic acids. This finding has important implications for the ability to use small molecules to modify gene structure and function in animals.

Although the overall efficiency of mutagenesis demonstrated with AG30 is low, our data provide the initial evidence that sequence-specific, DNA-modifying reagents may prove useful in intact animals. These observations should provide motivation to further enhance triplex technology to improve the efficiency of gene targeting, thereby allowing the targeted modification of genes in whole animals.

  • * To whom correspondence should be addressed. E-mail: peter.glazer{at}


View Abstract

Navigate This Article