Chromosome Targeting at Short Polypurine Sites by Cationic Triplex-forming Oligonucleotides*

Triplex-forming oligonucleotides (TFOs) bind specifically to duplex DNA and provide a strategy for site-directed modification of genomic DNA. Recently we demonstrated TFO-mediated targeted gene knockout following systemic administration in animals. However, a limitation to this approach is the requirement for a polypurine tract (typically 15–30 base pairs (bp)) in the target DNA to afford high affinity third strand binding, thus restricting the number of sites available for effective targeting. To overcome this limitation, we have investigated the ability of chemically modified TFOs to target a short (10 bp) site in a chromosomal locus in mouse cells and induce site-specific mutations. We report that replacement of the phosphodiester backbone with cationic phosphoramidate linkages, either N,N-diethylethylenediamine orN,N-dimethylaminopropylamine, in a 10-nucleotide, psoralen-conjugated TFO confers substantial increases in binding affinity in vitro and is required to achieve targeted modification of a chromosomal reporter gene in mammalian cells. The triplex-directed, site-specific induction of mutagenesis in the chromosomal target was charge- and modification-dependent, with the activity of N,N-diethylethylenediamine >N,N-dimethylaminopropylamine ≫ phosphodiester, resulting in 10-, 6-, and <2-fold induction of target gene mutagenesis, respectively. Similarly,N,N-diethylethylenediamine andN,N-dimethylaminopropylamine TFOs were found to enhance targeting at a 16-bp G:C bp-rich target site in a chromatinized episomal target in monkey COS cells, although this longer site was also targetable by a phosphodiester TFO. These results indicate that replacement of phosphodiester bonds with positively chargedN,N-diethylethylenediamine linkages enhances intracellular activity and allows targeting of relatively short polypurine sites, thereby substantially expanding the number of potential triplex target sites in the genome.

The recognition of double-stranded DNA by a single-stranded TFO 1 provides a technique for site-specific genome modification in living cells (1). Felsenfeld et al. (2) described the first example of the formation of triple helical nucleic acid structures in 1957. A decade later, transcription inhibition of Escherichia coli RNA polymerase by an RNA third strand was reported by Morgan and Wells (3), demonstrating a potential biological role of triplex structures. Subsequently, synthetic oligonucleotides designed to form triplexes have been used in a variety of applications to manipulate genes and gene function both in vitro and in vivo. TFOs have been used to inhibit protein binding to DNA (4 -7), to inhibit gene expression (8 -12), to inhibit the replication of DNA (13,14), to direct sitespecific DNA damage (15)(16)(17)(18)(19)(20)(21), to enhance recombination (1,(22)(23)(24)(25)(26), and to induce mutagenesis (17,(27)(28)(29)(30)(31). Recently, we have demonstrated (32) TFO-induced mutagenesis in transgenic mice following intraperitoneal injection of oligonucleotides designed to bind to a 30-bp polypurine site within a chromosomally integrated reporter gene.
The binding specificity afforded by TFOs is obtained from the Hoogsteen or reverse Hoogsteen hydrogen bonds formed in the major groove between the TFO and the purine-rich strand of the underlying target duplex (33). G-rich purine stretches are preferentially bound by purine-rich (GA or GT) TFOs, which are oriented anti-parallel to the purine strand of the duplex (8,34,35). In a purine triplex motif, guanines in the TFO pair with guanines in the target duplex forming G:GC triplets, whereas adenines or thymines bind to adenines forming A:AT or T:AT triplets. Whereas purine TFOs form stable triplexes at physiological pH, purine triplex formation can be inhibited by physiological levels of potassium (36,37), perhaps due to self-aggregation and the stabilization of guanine quartets or related structures. Several modifications have been incorporated into TFOs to diminish the potassium-mediated inhibition, including the base modifications, 6-thioguanine (36 -38) and 7-deazaxanthine (39,40), and the phosphoramidate backbone modifications, N,N-dimethylaminopropylamine (DMAP), N,N-diethylethylenediamine (DEED), and methoxyethylamine (41)(42)(43).
In order to modify effectively a gene in cells via triplex formation, high affinity, specific binding is required. Calorimetry measurements have demonstrated similar energies for Hoogsteen (Ϫ2.6 kcal/mol per base triplet) as compared with Watson-Crick (-1.3 kcal/mol per bp) hydrogen bonds, consistent with stable third strand binding (44). However, the requirement for relatively long polypurine sequences limits the number of potential target sites for triplex formation.
Several strategies have been used to enhance TFO binding affinity at sites where unmodified phosphodiester TFOs would be ineffective in cells. These include the covalent attachment of the duplex intercalator, pyrene, to the end of the oligonucleotide (along with sugar modification at the 2Ј position) or the incorporation of a phosphoramidate backbone consisting of substitution of a nitrogen for the 3Ј-bridging oxygen (31,45). Another approach to stabilize triplexes (and to aid entry into cells) is the use of polyamines either covalently attached to the TFOs or used in conjunction with the oligonucleotides during transfection (46,47).
The work reported here describes studies of triplex-mediated, targeted mutagenesis using short psoralen-linked TFOs with chemically altered backbones designed to enhance their binding affinities in vivo. We have compared the effects of backbone composition and charge on triplex formation in vitro under physiologic conditions, and we have examined the ability of the modified TFOs to target mutations within a chromosomal supF reporter gene at a 10-bp homopurine site and within an episomal reporter at a 16-bp site. We report that mutations were induced in the supF target gene at a frequency ϳ10-fold above background with a cationic TFO containing the DEED modification (pAG10DEED) and 6-fold above background with a TFO made cationic with DMAP modification (pAG10DMAP), whereas the phosphodiester TFO (pAG10PD) did not significantly induce mutagenesis above background levels. At the 16-bp episomal site, the PD TFO was somewhat effective (as expected, since this site affords high affinity binding by the standard DNA TFO), but the DEED and DMAP substitutions still enhanced the activity of the TFO. These results demonstrate that DEED-and DMAP-modified, cationic TFOs have greater intracellular activity than negatively charged phosphodiester TFOs of the same sequence. In addition, the results show that a homopurine site of only 10 bp can be a sufficient target for TFO-mediated genome targeting with appropriate chemical modification. Collectively, these results lend additional support to the potential application of triplex technology in vivo.
Binding Studies-Electrophoretic mobility shift assays were performed to determine apparent dissociation constants (K d values). For the 10-bp target site in the supF gene, oligonucleotides (24 nucleotides) corresponding to a region of the supF sequence were annealed to form a synthetic 24-bp target duplex. For the 16-bp target site in the variant supF10 gene, a synthetic 29-bp duplex was used. Duplexes were radiolabeled on the 5Ј end using T4 polynucleotide kinase and [␥-32 P]ATP, gel-purified, and incubated ϳ18 h at 37°C with increasing concentrations of TFO in a buffer containing 10 mM Tris-HCl, pH 7.6, 1 mM spermine, 10% glycerol, and additional cations (MgCl 2 and KCl) as indicated. Samples were then subjected to polyacrylamide gel electrophoresis in 12% native gels containing 89 mM Tris, 89 mM boric acid, pH 8.0, and 10 mM MgCl 2 for ϳ4 h at 60 V.
Chromosome Mutagenesis Protocol-Mouse fibroblasts (LN12 cells (48)) with multiple integrated copies of a vector DNA containing the supF mutation reporter gene were maintained and treated in media (Dulbecco's modified Eagle's medium supplemented with 20% fetal calf serum) containing G418 at 0.2 mg/ml (Life Technologies, Inc.). Oligonucleotides were added to the cells (plated at 1 ϫ 10 6 cells/10-cm 2 culture dish) at a final concentration of 1 M, via GenePorter transfection reagent (Gene Therapy Systems, San Diego, CA) according to manufacturer's suggestions. Cells were incubated at 37°C for 2 h, and UVA irradiation was administered at a dose of 1.8 J/cm 2 . Cells were collected for shuttle vector rescue and analysis 2-4 days later. Genomic DNA was isolated from the cells and incubated with in vitro packaging extracts for shuttle vector rescue as described (30). The packaged phage were adsorbed to E. coli C lacZ 125(am) and plated on LB plates containing 5-bromo-4-chloro-3-indolyl-␤-D-galactoside (X-Gal) at 1.6 mg/ml and isopropyl-␤-D-thiogalactoside at 1.3 mg/ml.
Episome Mutagenesis Protocol-An SV40-based shuttle vector plasmid pSupF10, similar to pSupFG1 (17), was used as a target. It was obtained from Dr. Michael Seidman (National Institute on Aging, Baltimore, MD). This vector contains a modified supF gene, supF10, containing a 16-bp G:C-rich target site at the 3Ј end of the gene. Monkey COS cells were transfected with the pSupF10 vector DNA using electroporation as described previously for similar vectors (22). Twenty four hours later, the cells were electroporated with the oligonucleotides (22), followed by UVA irradiation at a dose of 1.8 J/m 2 2 h later. Forty eight hours after irradiation, the cells were lysed for preparation of episomal vector DNA, as described (17). The rescued vector DNA was used to transform indicator bacteria to ampicillin resistance, and supF10 gene function was determined based on colony color, as described (17).

Effect of TFO Charge on Triplex
Formation-Previously, we reported a 5-fold TFO-directed induction of mutagenesis in transgenic mice (3340) containing a modified supF gene (supFG1) incorporating a 30-bp polypurine target site (32). Additionally, we demonstrated a 10-fold induction in a fibroblast cell line established from these mice (30). Mutagenesis was achieved using a specific 30-mer TFO, AG30, that (with a standard phosphodiester backbone) was found to bind to the target site with high affinity (K d ϳ 3 ϫ 10 Ϫ9 M). However, in the same cell culture study, efforts to target mutations to a shorter chromosomal polypurine site (10 bp) in the unmodified supF gene in mouse LN12 cells (48) were unsuccessful. Neither the specific 10-mer TFO (AG10), a related 13-mer (AG13), nor the 30-mer (AG30), all containing unmodified phosphodiester backbones, were able to induce mutagenesis in the target supF gene (30). We reasoned that the lack of mutagenesis was due to low affinity binding by the unmodified TFOs to the 10-bp supF polypurine target (K d values Ն10 Ϫ6 M). In all cases, nuclease resistance was provided by 3Ј-propylamine substitution (49).
In previous work, it was shown that the binding affinity of purine TFOs could be improved by replacing the phosphodiester backbone with either of two different cationic modifications, DMAP or DEED, both of which are protonated at physiological pH (41)(42)(43). In the present study, we asked whether DMAP or DEED modification of the AG10 TFO could provide sufficient increases in binding to allow targeting of the 10-bp homopurine site in the supF gene in LN12 cells. The structures of the DMAP and DEED modifications incorporated into the oligonucleotides used in this study are depicted in Fig. 1A. The sequence of the 10-mer TFO, AG10, designed to bind in an anti-parallel fashion to the supF target duplex is shown in Fig.  1B, along with the scrambled sequence control (SCR10).
Gel mobility shift assays were used to determine the binding affinity of pAG10PD, pAG10DMAP, and pAG10DEED to the supF triplex target site. Under the conditions used (10 mM Tris-HCl, pH 7.6, 10 mM MgCl 2 , 1 mM spermine, and 10% glycerol) there was essentially no detectable binding of pAG10PD to the supF target site ( Fig. 2A). However, the binding affinity was substantially increased when the phosphodiester backbone linkages were replaced with either the DMAP or DEED phosphate derivatives ( Fig. 2A). This result is consistent with previous work using 17-mer TFOs with the DEED modification targeted to the enhancer of the GS17 gene of Xenopus laevis (41). Hence, both of the cationic modifications showed substantially higher affinity triplex formation compared with the unmodified TFO. The simplest explanation is that the increased affinity is due to the reduction of electrostatic repulsion between duplex target and the TFO phosphates.
Magnesium Dependence of Triplex Formation-Magnesium is known to stabilize triplex structures but is not necessarily required for triplex formation to occur (37). In order to test whether magnesium was required for the DEED and DMAP TFOs to bind their target duplex, a range of MgCl 2 concentrations was used in gel mobility shift assays. When the MgCl 2 concentration was reduced to 0.1 mM, triplex formation again was not detected with pAG10PD at concentrations up to 10 Ϫ5 M, whereas triplex formation with pAG10DMAP and pAG10DEED was still detected (Fig. 2B). In fact, the extent of triplex formed by the modified TFOs was not significantly diminished in low MgCl 2 . These results suggest that the charge neutralization afforded by magnesium is not necessary for high affinity binding by the cationic TFOs. Triplex Formation in the Presence of Potassium-The binding of G-rich TFOs can be inhibited by physiological levels of potassium, most likely due to self-association during the formation of G-quartets and related structures. The DEED modification has been shown to reduce the inhibitory effect of potassium on triplex formation by a 17-mer TFO (41). In order to test whether the short, 10-mer TFOs used in this study modified with DEED or DMAP, could bind in the presence of physiological levels of potassium, we performed gel mobility shift assays following incubation of the TFOs with duplex DNA in buffer containing 140 mM KCl (Fig. 2C). The binding affinities for both the DMAP and DEED-modified TFOs were somewhat reduced in the presence of physiological levels of potassium, with apparent K d values in the range of 10 Ϫ7 M.
Targeted Mutagenesis in the Chromosomal supF Gene Is Dependent on TFO Charge-In order to determine the mutagenic potential of the DMAP-and DEED-modified TFOs on a chromosomal target, we transfected LN12 mouse cells containing the supF reporter gene carrying the 10-bp triplex target site with the TFOs. Two hours after transfection, the cells were irradiated with UVA (1.8 J/cm 2 ) to photoactivate the psoralen moiety covalently linked to the 5Ј end of the oligonucleotides. The results indicated an oligonucleotide charge-and structuredependent induction of mutagenesis (Table I). The specific TFO, pAG10PD, containing a negatively charged phosphodiester backbone, did not significantly induce mutations in supF above the background frequency (Table I), consistent with our previous work (30). Despite the very similar target site binding by the two cationic modifications in vitro, there was a small but detectable difference when the intracellular efficacy was examined. The DMAP modification (pAG10DMAP) induced mutagenesis ϳ6-fold above the background mutation frequency, whereas the DEED TFO, pAG10DEED, induced mutagenesis 10-fold above the background, based on the compiled results of  two independent experiments using separate preparations of oligonucleotides (Table I).
To confirm that the effects seen with pAG10DMAP and pAG10DEED were sequence-specific and triplex-mediated, we treated cells with a control oligonucleotide, pSCR10DEED, that contained the cationic DEED-modified backbone structure and the same base composition as AG10 but a scrambled sequence. Mutagenesis in the supF gene was not induced in cells treated with this control oligonucleotide. These data show that the DEED-and DMAP-modified TFOs can specifically induce mutations in the target gene in mouse cells, whereas a scrambled control oligonucleotide, even with the cationic DEED modification, cannot.
Mutagenesis Is Not Induced in the cII Gene Lacking the Triplex Target Site-To confirm further that the mutagenesis was triplex-dependent and to examine the specificity of gene targeting, we screened the cII gene in the same vector in LN12 cells for possible nonspecific induction of mutagenesis by the PD, DMAP, and DEED TFOs. The cII gene serves as a specificity control since it does not contain the 10-bp triplex target site. Although the background mutation frequency in the cII gene is higher than that of the supF gene, mutations in the cII gene were not significantly induced in cells treated with any of the 10-mer oligonucleotides (Table II). This result provides additional support for a specific triplex-directed effect in the supF gene with the DMAP-and DEED-modified 10-mers.
Mutation Spectra in the supF Gene of LN12 Cells-Phage plaques containing mutated supF genes were purified and subjected to DNA sequencing analysis. Other than an apparent nonspecific hot spot (or pre-existing mutation) at position 106, the majority of mutations induced by pAG10DMAP and pAG10DEED were T:A to A:T transversions at bp 167 within the targeted psoralen intercalation site (bp 166 -167) at the triplex-duplex junction (Fig. 3A). This result is consistent with the pattern of mutations to be expected from triplex-targeted psoralen adducts, and so it provides unambiguous evidence for specific TFO-induced mutagenesis on the chromosome of LN12 cells. In contrast, the mutations induced by the control oligonucleotides were more dispersed throughout the supF gene and were not found either within the triplex-binding site or at the bp 166 -167 psoralen intercalation site (Fig. 3B), further illustrating the specificity of the TFO-directed mutagenesis. Interestingly, in the pAG10DEED-treated sample, we did detect a 1-bp slippage event in the 5-bp G:C run in the triplex target site. This finding is consistent with our previous observation of frequent Ϫ1 and ϩ1 deletions and insertions induced in the longer 8-bp G:C run in the supFG1 polypurine target site by the AG30 TFO in 3340 cells (30).
TFO-directed Mutagenesis in a 16-bp Target-To extend the comparison to a second target sequence, we tested psoralen-TFO-mediated targeting of a 16-bp G:C-rich site at the 3Ј end of a modified supF gene, supF10. This gene was carried in an SV40-based shuttle vector, pSupF10, similar to pSupFG1 used in previous work from our group (17). The 16-bp target site in supF10 is shown in Fig. 1C, along with the sequence of the specific TFO, AG16. Binding analyses indicated that pAG16PD can bind with high affinity to the target site, even in the presence of high K ϩ and low Mg 2ϩ (Fig. 4). High affinity binding was also found in the case of pAG16DMAP and pAG16DEED, although under the conditions of the assay some self-association of the DMAP and DEED TFOs prevented full shifting of the target duplex. A targeting protocol was carried out in which monkey COS cells were transfected by electroporation with the pSupF10 vector DNA. After 24 h to allow chromatinization of the vector (50), the cells were transfected by electroporation with selected psoralen-linked TFOs, pAG16PD, pAG16DMAP, and pAG16DEED, followed by UVA irradiation 2 h later. After 48 h to allow repair and replication, the vector DNA was harvested for supF10 gene analysis via transformation into indicator bacteria. The background of spontaneous mutagenesis in the assay was 0.06% (Table III). pAG16PD was able to induce mutagenesis above this level, similar to the activity of a 30-mer PD TFO (pAG30) targeting a related gene in the same type of assay (17). However, both pAG16DMAP and pAG16DEED showed enhanced levels of targeted mutagenesis, yielding supF10 mutations at frequencies of 0.83 and 0.87%, respectively. These data again show the utility of the DMAP and DEED modifications, even for a high affinity TFO target site.

Triplex-induced Mutagenesis Requires High Affinity
Binding-Triplex technology offers a promising strategy to modify a mammalian genome. However, in order to apply this technology, TFOs must be capable of binding chromosomal DNA sites in living cells. In previous work, we demonstrated triplexmediated mutagenesis on plasmid DNA (27) using pAG10PD to target the short triplex target site (10 bp) in the supF gene. In that study, the plasmid DNA was preincubated in vitro with the psoralen-modified TFO and UVA-irradiated prior to delivery into cells; therefore, intracellular binding was not required to induce mutagenesis. Subsequently, we showed that the 10-bp triplex target site in the supF gene was not a successful target with PD TFOs in either an episomal or a chromosomal context within cells (17,30). However, when either the episomal or chromosomal target site was modified to contain a 30-bp polypurine target to afford high affinity TFO binding, we observed a 10 -60-fold induction in TFO-mediated mutagenesis with a 30-base PD TFO (17,30). We concluded that the inability to target the short supF site on a chromosomal locus was most likely due to the low binding affinity of pAG10PD in vivo.
In another study, we found that dimeric, clamp-forming peptide nucleic acids (PNAs) designed to bind to a 10-bp region in the supFG1 gene were capable of mediating targeted mutagenesis of the reporter gene in mouse fibroblasts in culture (51). Unlike TFOs, the PNAs were designed such that one segment could strand invade and bind to the purine strand of the target duplex by Watson-Crick interactions, with a second PNA segment attached by a flexible linker designed to bind to the resulting PNA/DNA duplex as a third strand by Hoogsteen bonding to form a PNA/DNA/PNA triplex, with the other strand of the target duplex displaced. The dimeric PNA has a neutral charge and binds with high affinity to the 10-bp target site, much higher than the affinity of the PD TFO, AG10, to the same site. Hence, although the dimeric PNA forms a very different structure at the target site than does a TFO, the PNA study suggested that DNA-binding molecules of altered charge and/or enhanced binding affinity could mediate intracellular targeting at relatively short sites that might otherwise be refractory to targeting by PD TFOs.
In the present study, we modified pAG10PD by replacing the phosphodiester linkages with cationic phosphoramidate linkages. The modified TFOs (pAG10DMAP and pAG10DEED) both bound the 10-bp target site with high affinity (K d ϳ10 Ϫ8 M), and were capable of targeting the supF gene in cells, as assessed by mutagenesis. The supF mutation frequency was specifically enhanced 10-fold by pAG10DEED and 6-fold by pAG10DMAP in the mouse cells. These results demonstrate that a genomic site only 10 bps in length can provide a successful target for triplex-mediated genome modification.
In addition, comparison of targeting by a series of 16-mer TFOs at a 16-bp site in the episomal pSupF10 vector also showed increased activity conferred by the DEED and DMAP modifications. This enhancement was seen even though the 16-bp target site provides for high affinity binding by the PD 16-mer TFO.
Targeted Mutagenesis Is Enhanced by Using Cationic TFOs-Our results indicate that TFO-mediated mutagenesis can be enhanced using positively charged TFOs. The two different cationic modifications tested both increased mutation rates compared with unmodified TFOs. The elevated mutagenesis with the modified TFOs most likely reflects their high affinity binding to the target site under physiological conditions, as both modifications showed enhanced binding to their duplex target when assayed in vitro under conditions reflecting physiologic Mg 2ϩ and K ϩ concentrations. In addition, the cationic oligonucleotides may have improved intracellular stability as compared with the phosphodiester oligonucleotides.
A small difference was detected when comparing the intracellular effectiveness of the two modifications at the chromosomal supF target. The greater mutagenic effect at the chromosomal target site in LN12 cells seen with the DEED modification compared with DMAP may reflect yet to be determined differences in competition with histones or other DNAbinding proteins, or differences in the recognition of the triplex by the DNA repair machinery. However, the DEED-and DMAP-modified TFOs showed similar activity in targeting the episomal vector in COS cells, and so further study is needed to confirm whether the DEED modification is superior to DMAP.
Specificity of Triplex-induced Mutagenesis-Since the oligonucleotides used in this study contained psoralen on the 5Ј end, we wanted to rule out the possibility that the cationic TFOs were simply serving as nonspecific, positively charged carriers of psoralen. Therefore, we compared mutation frequencies in the supF gene containing the 10-mer polypurine site to that of the cII gene lacking the triplex target site. Although the mutation frequency was induced ϳ10-fold in the supF gene with the specific TFO (pAG10DEED), induced mutagenesis was not detected in the cII gene following treatment with pAG10DEED. In addition, a control psoralen-oligonucleotide containing the DEED modification, but having a scrambled sequence, was not capable of inducing mutagenesis in either the supF gene or the cII gene. These data rule out a nonspecific mutagenic effect of the tethered psoralen or some nonspecific effect of the DEED modification on DNA metabolism in the absence of high affinity triplex formation.

TABLE III
Targeted mutagenesis of the chromosomal supF10 gene within an episomal SV40 vector in COS7 monkey cells by DMAP and DEEDmodified psoralen TFOs Cells were transfected via electroporation with the pSupF10 vector followed 24 h later by transfection with the indicated oligonucleotides and incubated at 37°C for 2 h. Cells were then irradiated with 1.8 J/cm 2 UVA. After 2 days of repair and replication, the cells were collected for vector rescue and transformation into indicator bacteria. The frequency of mutations in the supF10 gene was calculated by dividing the number of colorless mutant colonies by the total number of colonies counted. The data represent the combined results of two independent experiments using separate preparations of oligonucleotides, with the S.E. given. Sequence Analysis-The mutation pattern induced by the specific TFOs in the chromosomal supF gene target is consistent with a triplex-mediated process since the majority of mutations were found in the triplex binding site or at the targeted psoralen intercalation site at the triplex-duplex junction. The mutations in the control samples show a more scattered pattern, providing additional evidence that the specific TFOs were capable of binding to the triplex target site, whereas the scrambled control oligonucleotides were not. The T:A to A:T transversions seen at the triplex-duplex junction are in keeping with the pattern expected to arise from psoralen adducts targeted by third strand binding, thereby providing clear evidence for triplex formation at the chromosomal target site. Interestingly, the TFO-targeted mutations included a 1-bp insertion within the polypurine target site. This type of mutation is consistent with the pattern of frequent Ϯ1-bp insertions and deletions in the 8-bp run in the supFG1 gene induced by AG30 (30,32). The lower frequency of such slippage mutations in the present work likely reflects the difference in stability and in propensity for strand dislocation and slippage between the 5-bp G:C run in supF and the 8-bp G:C run in supFG1. Dramatic differences in slippage mutation frequency based on mononucleotide repeat length have been documented in yeast (52).
In conclusion, we have demonstrated TFO-mediated targeted mutagenesis of the supF gene in mouse cells at a short 10-bp triplex target site. This effect was dependent on the cationic nature of the TFO. By replacing the phosphodiester linkages with phosphoramidate DMAP or DEED linkages, we have enhanced the binding affinity of a 10-mer by more than 100-fold, indicating that the limitation of triplex technology to long polypurine sites can be overcome to some extent. In addition, the enhancement of targeting at the 16-bp supF10 polypurine site suggests that these cationic modifications may have utility even for reagents directed at high affinity binding sites. Although a 10-bp site will occur frequently in the genome and so is not a unique target, the work reported here serves to demonstrate the ability of selectively modified TFOs to modify short, otherwise suboptimal polypurine sites in the genome, thereby enhancing the utility of triplex technology for sitespecific genome modification.