Minor groove DNA alkylation directed by major groove triplex forming oligodeoxyribonucleotides
Minor groove DNA alkylation directed by major groove triplex forming oligodeoxyribonucleotidesEugeny A. Lukhtanov*, Alan G. Mills, Igor V. Kutyavin, Vladimir V. Gorn, Michael W. Reed and Rich B. Meyer
Epoch Pharmaceuticals Inc., 1725 220th Street SE, #104, Bothell, WA 98021, USA
Received July 15, 1997;Revised and Accepted October 29, 1997
ABSTRACT
We describe sequence-specific alkylation in the minor groove of double-stranded DNA by a hybridization-triggered reactive group conjugated to a triplex forming oligodeoxyribonucleotide (TFO) that binds in the major groove. The 24 nt TFOs (G/A motif) were designed to form triplexes with a homopurine tract within a 65 bp target duplex. They were conjugated to an N5-methyl-cyclopropapyrroloindole (MCPI) residue, a structural analog of cyclopropapyrroloindole (CPI), the reactive subunit of the potent antibiotic CC-1065. These moieties react in the DNA minor groove, alkylating adenines at their N3 position. In order to optimize alkylation efficiency, linkers between the TFO and the MCPI were varied both in length and composition. Quantitative alkylation of target DNA was achieved when the dihydropyrroloindole (DPI) subunit of CC-1065 was incorporated between an octa(propylene phosphate) linker and MCPI. The required long linker traversed one strand of the target duplex from the major groove-bound TFO to deliver the reactive group to the minor groove. Alkylation was directed by relative positioning of the TFOs. Sites in the minor groove within 4-8 nt from the end of the TFO bearing the reactive group were selectively alkylated.
Sequence-specific triple helix formation is an appealing method for controlling gene expression by oligonucleotides. Triplex forming oligonucleotides (TFOs) have been shown to interfere with DNA replication (1 ) and transcription (2 ) and can mediate site-directed mutagenesis (3 ). They can inhibit integration of the DNA copy of a viral RNA genome into the host cell DNA in vitro (4 ).
In general, however, reversible binding of TFOs to a DNA target is not tight enough to strongly affect gene function in cells. In order to enhance the efficacy of TFOs, a number of improvements have been utilized. Covalent attachment of duplex- (5 ) and triplex-specific (6 ) intercalating agents, minor groove binders (7 ) or polycationic molecules (8 ) has been shown to stabilize DNA triplexes. Alternatively, TFO backbone modifications, such as peptide nucleic acids (9 ) or N3' -> P5' phosphoramidates (10 ), can also be utilized for this purpose.
Covalent attachment of the TFO to its specific binding site contributes more permanence to the binding effect. Toole and co-workers demonstrated that addition of covalent bonding of the TFO to a target DNA site was required to permanently block transcription elongation (11 ). Covalent modification of the TFO binding site may also inactivate the targeted gene if it is not repaired correctly (3 ). A variety of reactive groups, most noticeably psoralen (12 ), bromoacetyl (13 ), nitrogen mustard (14 ) residues or transplatin adducts (15 ), have been conjugated with TFOs and shown to react with target sites.
We recently introduced a new class of alkylating oligodeoxyribonucleotides (ODNs) bearing cyclopropapyrroloindole (CPI) (16 ) or N5-methyl-CPI (MCPI) (17 ) alkylating groups. These agents were derived from the reactive subunit of the potent antibiotic CC-1065. DNA alkylation by bound CPI normally occurs in the minor groove of A/T-rich regions at the N3 position of adenines. The stability of these moieties in complex biological systems (18 ) and their hybridization-triggered reactivity (19 -22 ) separate them from the majority of other alkylating agents. We have shown that MCPI-ODN conjugates alkylate DNA duplexes adjacent to their complementary single-stranded binding sites with rates and efficiencies that depended on both the target sequence and the reactive group type (17 ). Under optimal conditions the efficiency of this alkylation reaction may reach 100% within minutes.
In order to apply the advantages of this class of reactive moieties to triplex-directed alkylation of DNA we have investigated their conjugation to TFOs. Topologically this necessitates connecting the major groove binding TFO with the minor groove binding CPI via a linker that crosses over one of the DNA strands. Here we report methods for application of this new type of reactive ODN for triplex-directed alkylation of double-stranded DNA.
ODN syntheses were carried out on an Applied Biosystems Model 394 DNA synthesizer utilizing the 1 µmol coupling program supplied by the manufacturer. A phosphoramidite for introduction of the 6-aminohexyl residue was purchased from Glen Research (Sterling, VA). 6-Aminohexyl and 6-hydroxyhexyl tails were introduced at the 3'-end of ODNs on the DNA synthesizer using modified CPG supports (23 ). Tetra(ethyleneglycol) phosphate and propylene phosphate linkers were introduced using Spacer 18 and Spacer C3 phosphoramidites from Glen Research. After ammonia deprotection the ODNs were HPLC purified, detritylated and precipitated as described previously (24 ). MCPI-ODN and MCPI-DPI-ODN conjugates were synthesized as described previously (17 ).
One strand of the target duplex (see Table 1 ) was 5'-32P-labeled using polynucleotide kinase and [[gamma]-32P]ATP and then mixed with a 2-fold excess of unlabeled complementary strand in a buffer containing 140 mM KCl, 10 mM MgCl2, 1 mM spermine and 20 mM HEPES, pH 7.2. The mixture was heated for 1 min at 95°C and incubated at 37°C for at least 30 min before addition of the reactive TFO. Final concentration of the 32P-labeled duplex was 20 nM and reactive TFO was added in 10- or 100-fold excess. Each reaction mixture was incubated at 37°C for 18-20 h. Aliquots were mixed (1:4 v/v) with loading solution (80% formamide, 0.25% bromophenol blue, 0.25% xylene cyanol FF) and analyzed by 8% denaturing PAGE at 50-55°C. The gel was transferred onto paper, dried and radioimaged on a BioRad Molecular Imager using the manufacturer's provided computer software. The cross-linked products were identified as slower (relative to unreacted radiolabeled duplex strand) moving bands (Figs 2 and 3 ). The efficiency of the cross-linking reaction (%) was calculated as a ratio between the radioactivity in the cross-linked products (cross-linked band, smear under the cross-linked band and products of partial spontaneous cleavage) and the total radioactivity in the lane. To determine the bases that were alkylated reaction mixtures were heated in boiling water (30 min), then in 10% aqueous piperidine (15 min), dried in vacuo, precipitated in ethanol to remove salts, dissolved in loading solution and analyzed by denaturing PAGE. Random cleavage of the 32P-labeled strand at the purines (A+G reaction, 60% formic acid, 37°C, 12 min, followed by treatment with piperidine as above) was used as a control lane to assign the cleavage products (Fig. 4 ). The kinetics of triplex-directed MCPI alkylation were determined by accumulation of cleavage products over time. Aliquots were withdrawn periodically and subjected to heat and piperidine treatment.
A gel retardation assay (25 ) was employed to determine the dissociation constant for selected TFOs. In order to ensure good resolution of triplex and duplex states a duplex target comprising d(AGCCTGGAAGAGAAGGAGAGAGGAGAGGAAAGA) and its complement was used for the precursors of TFOs 1, 9 and 21-24. For the precursors of TFOs 12 and 16 the duplex target was d(AGAAGGAGAGAGGAGAGGAAAGAGGAGACAA) and its complement. These are portions of the 65mer DNA target in Table 1 . The pyrimidine-rich strand of the duplex was 5'-32P-labeled and mixed in reaction buffer with a 2-fold excess of the complementary purine-rich strand. The mixture was heated for 1 min at 95-100°C. After incubation at 37°C for 30 min the labeled duplex was mixed with TFO solution. The final concentration of 32P-labeled duplex in the mixture was 200 pM. The concentration of TFO in the reaction mixture was changed stepwise from 20 µM to 2.1 nM in 2-fold decrements. The reaction buffer used for this experiment was the same as in the alkylation studies, but spermine was omitted because of the precipitation of oligonucleotide at low temperature during gel analysis. The prepared mixtures were incubated at 37°C for 18-24 h, loading solution (15% Ficoll, 0.25% bromophenol blue, 0.25% xylene cyanol FF, 5 mM MgCl2, 10% of reaction volume) was then added and samples were analyzed in non-denaturing 8% polyacrylamide gel. The gel buffer contained 90 mM Tris-borate, pH 8.3, 2 mM EDTA and 5 mM MgCl2. The temperature of the gel during analysis did not exceed 11-12°C. After electrophoresis the gel was transferred onto paper, dried and imaged. The ratio between triplex and duplex states in each sample was measured. The dissociation constant was determined as being equal to the concentration of TFO at which the duplex:triplex ratio was 0.5.
A 65 bp DNA fragment of the corrected sequence (26 ) of the human class II major histocompatibility antigen HLA DQ[beta]1× 0302 gene (GenBank accession no. K01499; 27 ) was employed as a double-stranded DNA target for the cross-linking studies. This fragment contains a 33mer long homopurine run and four A/T-rich sites, potential targets for MCPI-alkylation (28 ,29 ). Sequences of the target duplex and reactive TFOs are shown in Table 1 . The structures of the linkers and reactive groups used in TFO design are shown in Figure 1 . TFOs were designed in all but one case (TFO 6) as antiparallel 24mer purine motif ODNs (30 -32 ) that could form triplexes with the target homopurine run via Hoogsteen-type hydrogen bonding. TFO 6,in addition to the 24mer homopurine core, had a nonsense octapyrimidine sequence at the 5'-end employed as a linker arm. MCPI or MCPI-DPI moieties, both (+) and (-) enantiomers, were conjugated to either end of the TFOs. Since cell culture testing was ultimately planned for the TFOs, a hydroxyhexyl phosphate moiety was incorporated at the unmodified end to enhance stability of the ODN conjugates against exonucleases (23 ). TFOs 1-3, 7-9, 13 and 17-24 contained the same 24mer sequence, varying only in the structure of the reactive residue and the linker at the 5'-end. Others (4, 5, 10-12 and 14-16), while being the same length, were designed to position their 5'-reactive moiety toward target sites 1-3. TFOs 21 and 22 were 3'-MCPI-DPI conjugates, targeting site 4 at the 3'-end of the hybridized TFO. TFOs 23 and 24 were bis-alkylating conjugates, bearing an alkylating residue at both ends. Non-complementary ODN conjugates 25 and 26 had a scrambled G/A-containing sequence for use as sequence specificity controls in the cross-linking experiments.
The buffer employed in this study (140 mM KCl, 10 mM MgCl2, 1 mM spermine, 20 mM HEPES, pH 7.2) mimics the intracellular ionic environment (33 ). Polyvalent cations such as Mg2+ and spermine are known to stabilize G/A triple-stranded helixes, while potassium cations may compromise triplex stability (30 ,34 ,35 ). In order to show triplex formation under these conditions the dissociation constants (Kd) for selected precursor ODNs were determined. The precursor to TFOs 1 and 9, with the 5'-X8 linker and a 3'-hexanol tail but no reactive moiety, revealed a Kd of 1 × 10-8 M. When the X8 linker was moved to the 3'-end of this ODN (precursor of TFOs 21 and 22) the Kd was measured as 2 × 10-8 M. This same ODN with X8 linkers at both ends (the precursor of TFOs 23 and 24) had a Kd of 5 × 10-8 M. When the target sequence was shifted eight bases to the 3' side of the homopurine tract (precursor to TFOs 12 and 16) the Kd was 1 × 10-7 M. Triplex formation, therefore, was at least 50-80% complete at the 200 nM concentration of the TFOs used in cross-linking experiments. Since spermine, a potent triplex stabilizing agent (34 ,35 ), was omitted in the gel retardation study due to ODN precipitation at low temperature, the triplex stability was likely more favorable in the alkylation experiments. Within limits, however, it appears that the affinity of the TFOs for their binding site on the duplex target is not affected by the linker to a great degree.
The major focus of this study was to find linkers that could bridge from the major groove binding site of the TFO to the minor groove binding site of the MCPI moiety. Molecular modeling showed that relatively long linkers are required. The linkers have to traverse or wrap one of the DNA strands, overcoming steric and, in some cases, electrostatic effects. Not surprisingly, TFO 7, which carried a (+)-MCPI-DPI residue attached via linker H (the shortest linker employed in the study), showed no alkylation, as shown in Figure 2 . The longer linker X4 (TFO 8), which consists of four propylene phosphates terminated with an H tail, was much more effective, alkylating the purine strand of the target duplex with an efficiency of 78%. Additional extension of the linker to eight propylene phosphates (linker arm X8 in TFO 9)resulted in a quantitative reaction.
Linkers Y and Y2 were tested to determine whether the electrostatic repulsion between the negatively charged linkers X4 or X8 and the traversed duplex strand affects the cross-linking efficiency. Reduction in the number of negative charges might allow shortening of the linker. Linker Y (TFO 17) is similar in length to linker X4 (30 and 34 atoms) but has only one internal phosphodiester bond in a hexaethylene glycol chain, compared with four in X4. Analogously, Y2 (TFO 18) is comparable with X8 in length (50 and 58 atoms) but is less charged (two versus eight internal phosphodiester groups). Surprisingly, we found almost no benefit from using less charged linkers. Substitution of linker X4 (TFO 8) by linker Y (TFO 17) resulted in a modest increase in efficiency of target alkylation (from 78 to 87%), whereas linker Y2 did not show any advantage over linker X8 (Table 1 ). TFO 6, with a linker comprising eight pyrimidine nucleotide residues, had the same sequence and number of linker atoms as TFO 9 but showed only modest reactivity (15%), in contrast to quantitative alkylation by TFO 9.
(+)-MCPI and both enantiomers of MCPI-DPI were compared for triplex-directed alkylation of the minor groove. The (-)-MCPI analog was not examined since previous work (17 ) showed it to be inefficient for DNA alkylation.
TFOs 1-5 (Table 1 ) carried the (+)-MCPI residue at the 5'-end and varied in length and composition of the linker arms as well as in juxtaposition of the triplex relative to target sites 2 and 3. None of these (+)-MCPI-TFO conjugates showed satisfactory alkylation of the double-stranded DNA target (Table 1 and Fig. 3 ). The maximum efficiency (22%) of sequence-specific alkylation at site 2 was observed for TFO 4, which had octa(propylene phosphate) linker X8. Among the three TFOs 1-3 directed to site 3, TFO 2, which had a relatively short Y linker, was almost inactive, while TFOs 1 and 3, with longer linker arms X8 and Y2, showed modest alkylation. Although the low efficiency of reaction of the (+)-MCPI-TFO conjugates could be due to competition between the MCPI residue and spermine for the minor groove binding site, elimination of spermine from the reaction decreased the efficiency for TFO 1 (data not shown).
The addition of the DPI minor groove binding subunit to the (+)-MCPI moiety dramatically improved TFO-directed target alkylation (Fig. 3 ). Almost quantitative reactions (>90%) were achieved for most TFOs carrying the (+)-MCPI-DPI moiety (Table 1 ). The TFOs bearing (-)-MCPI-DPI showed an intermediate reactivity between the (+)-MCPI and (+)-MCPI-DPI conjugates, averaging 30% efficiency.
The sequence specificity of alkylation by CPI dictates that the TFO carrier must hybridize near a favored site. Prediction of the optimal triplex location was complicated, however, by the unknown behavior of the long `wrap around' linkers. TFOs 9-12, with a (+)-MCPI-DPI attached at the 5'-end via the X8 linker, were designed with the 5'-terminus progressively closer to site 1 or 2 to probe the regiospecificity of this method of linking reactive group to TFO. The difference in site specificity of TFOs 9-12 was first detected by denaturing gel analysis of corresponding reaction mixtures (Fig. 3 ). The pattern change between TFO 9 and TFO 12 suggested that the conjugates likely alkylate the target duplex at different, and in most cases multiple, sites. This was confirmed by analysis of heat/piperidine cleavage products (Fig. 4 B). TFO 9 efficiently reacted only with site 3, situated immediately next to the 5'-end of the TFO. A four base shift (TFO 10) significantly diminished site 3 alkylation in favor of site 2. An additional two base shift (TFO 11) resulted in targeting of site 2, leaving only traces of site 3 alkylation. Site 1 became accessible only for TFO 12, although with lower overall efficiency of alkylation (69%). The lower efficiency may be due in part to the somewhat weaker binding of this TFO as determined by gel retardation (see above). A similar dependence was found for the analogous set of TFOs (13-16) bearing the (-) enantiomer of the MCPI-DPI alkylator (Figs 3 and 4 B). Even TFO 16 was not able to alkylate site 1, reflecting its incapacity to bind in the correct manner to react with this site.
TFOs 21 and 22 had respectively (+)- and (-)-MCPI-DPI attached at their 3'-end for reaction at site 4 (Table 1 ). In contrast to the 5'-conjugates, where strand orientation favored alkylation of the purine strand by conjugated (+)-MCPI-DPI, 3'-conjugates of the (+) enantiomer preferably alkylated adenines on the pyrimidine target strand in the 5'-CTTAAG-3' sequence. Indeed, analysis of the cross-linking products (Fig. 3 ) revealed a very efficient modification (88%) of the pyrimidine strand by TFO 21. TFO 22, which carried the (-) enantiomer of the same agent, demonstrated moderate (36%) alkylation of the purine strand. Alkylation of different DNA strands by these two conjugates is a direct consequence of their opposing specificity for reaction with only one side of the groove. The shorter tethers should favor only one binding orientation for each of these enantiomers, but this becomes less relevant when long and flexible linkers are employed. This was exactly the case when TFO 13 [(-)-MCPI-DPI conjugate] alkylated the same strand as TFO 9 [(+)-MCPI-DPI conjugate]. The required different binding modes for each tethered enantiomer are shown in Figure 5 A and B.
TFOs 9 and 21, with the same sequence but with the (+)-MCPI-DPI residue attached via the X8 linker at the 5'- or 3'-end respectively, gave very efficient alkylation of the target but on opposite strands. TFO 23, which had both 3'- and 5'-ends modified with (+)-MCPI-DPI, alkylated both strands of the target duplex with >85% efficiency (Table 1 ). A diagram of this cross-link is illustrated in Figure 5 C. The cross-linked product was detected by gel analysis as a very slow migrating band (Fig. 2 ). Cleavage product analysis showed that the alkylation sites were the same as for TFOs 9 and 21 (Fig. 4 ).
The exact bases of target alkylation were determined by denaturing PAGE analysis of heat/piperidine-treated reactions with a 32P-radiolabeled duplex strand (Fig. 4 ). As expected (36 ), only adenine residues were alkylated. Within that binding site base specificity of alkylation is controlled by stereochemistry of the MCPI cyclopropane group. For example, TFOs carrying (+)-MCPI (1-3) (data not shown) or (+)-MCPI-DPI (8, 9, 17 and 18) reacted with the adenine closest to the 3'-end of site 3 (3'-GAAAG-5', where underlining indicates the alkylated base) regardless of the type or length of linker arm (Fig. 4 A). Conjugated (-)-MCPI-DPI (TFOs 13, 19 and 20), on the other hand, binds in the reverse orientation (Fig. 5 ) and has a different alkylation specificity than the (+) enantiomer (29 ,37 ). It preferentially alkylates adenine in the middle of this site (3'-GAAAG-5'), as do the free (-) enantiomers (37 ). The 3'-terminal adenine was also alkylated by (-)-MCPI-DPI conjugates, but only half as efficiently.
Site 2 was selectively modified with the same regiospecificity by TFOs that bound closest to that site (TFOs 10 and 11). The (+)-MCPI-DPI-TFOs modified the 3'-terminal adenine residue of site 2, while their (-) enantiomers 14 and 15 preferred to alkylate the adenine residue in the middle. TFO 12 was able to reach, although not very efficiently, the more remote site 1 (3'-TACAT-5' sequence in the purine strand). Both adenines were alkylated, with some preference for the 3'-terminal one (Fig. 4 B). When the minor groove alkylating agent was conjugated to the 3'-end of TFOs 21 and 22 site 4 was alkylated with the same preferences shown above. The (+)-MCPI-DPI on TFO 21 reacted with the most remote 3'-terminal adenine of site 4 (5'-CTTAAG-3') in the pyrimidine strand. The slight reaction with the adjacent adenine (Fig. 4 C) is due to `sliding' of the (+)-MCPI-DPI residue at site 4, which has four A/T bases instead of the three in sites 2 and 3. As expected, the (-)-MCPI-DPI residue in TFO 22 alkylated the adenine located nearer the middle of site 4 (3'-GAATTC-5') in the purine target strand.
The time course of reaction was determined for two of the TFOs (9 and 10) found to be effective in sequence-specific target alkylation. As shown in Figure 6 , at 200 nM both TFOs modified the target duplex relatively quickly. These reactions were 50% complete after 25 (TFO 9) and 31 min (TFO 10) at 37°C. A 10-fold increase in the concentration of TFO 9 to 2 µM significantly accelerated the reaction (t1/2 = 5 min). Since the concentrations used for TFOs 9 and 10 are well above the determined Kd (10 nM for TFO 9) of the triplex, it is reasonable to suggest that the rate of triplex formation limits the subsequent alkylation step at the lower TFO concentration. As a measure of the rate of DNA alkylation by (+)-MCPI-DPI in the absence of TFO binding the rate of reaction of scrambled control TFO 25 was much slower at 2 µM concentration (curve 4 in Fig. 6 ). This non-TFO-directed reaction continued to increase over time, however, as discussed below.
CPI and particularly CPI-DPI bind and react in the DNA minor groove (37 ), as do the MCPI analogs used herein. This could lead to alkylation events specific for the A/T region preferences of MCPI but not related to TFO binding. We have previously observed that conjugation of a reactive minor groove binding moiety to an ODN significantly masks this non-ODN-specific binding ability (17 ), presumably due to repulsion between the negatively charged DNA and ODN strands. At a concentration of <= 200 nM non-TFO-specific reactions were not consequential. This was directly tested with two scrambled sequence TFOs, 25 and 26, which could not form triplexes with the target duplex but which carried (+)- and (-)-MCPI-DPI residues. At 200 nM only traces of non-specific target alkylation (<5%) were detected for the scrambled (+)-MCPI-DPI conjugate after 18 h (Fig. 2 ). Predictably, TFO 21 with conjugated (+)-MCPI-DPI showed very low levels of non-specific alkylation of the purine strand at site 3 (Fig. 4 A) compared with very efficient alkylation of the pyrimidine strand at the expected site 4 (Fig. 4 C). At 2 µM TFO 25, however, the non-TFO-specific reaction level increased to ~80% after 18 h incubation (data not shown). As expected, this alkylation was entirely specific for the preferred binding sites of this minor groove binder, being distributed among the A/T-rich sites 1-4. In contrast, the less reactive (-)-MCPI-DPI and (+)-MCPI conjugates did not have this non-TFO-specific `anchor' effect at any tested concentrations (200 nM and 2 µM).
In order to bind in the minor groove a conjugated residue must be linked to the TFO by a tether that traverses one of the duplex strands. Molecular modeling indicates that for 5'-conjugated TFOs the linker likely crosses the pyrimidine-rich strand because of the right hand turn of the helix. In most of the alkylated complexes studied here both the TFO and alkylating residue interact with the purine-rich strand. This leaves the pyrimidine strand topologically arrested (wrapped) by the linker (Fig. 5 A). This creates an interesting phenomenon whereby the pyrimidine strand cannot readily separate from the alkylated complex, even under conditions of denaturing gel analysis (8 M urea). At a gel temperature of 40-45°C we found almost identical bands from the reaction mixtures of 5'-tailed TFOs, whether the purine or the pyrimidine target strands had been 5'-32P-labeled. Raising the gel temperature to 50-55°C or preheating the loading samples in boiling water for 1 min eliminated this band, particularly when the pyrimidine strand was labeled.
This study demonstrates triplex-directed covalent modification of a minor groove site by a major groove binding TFO. The long flexible linkers used here traverse one of the target duplex strands and allow simultaneous binding of the TFO and the alkylator in opposite DNA grooves. The length was most critical, with the best result achieved using 50-58 atom linkers. Composition of the linker has less effect on alkylation efficiency. Interestingly, in earlier reports of TFO-minor groove binder (TFO-MGB) chimeras much shorter linkers were utilized. Dervan and co-workers used a C12 linker to link the 5'-end of a T/C motif TFO with a polyamide minor groove binding residue (38 ). In our study linker H in combination with a succinyl residue from the reactive group (Fig. 1 ) was analogous in length and composition, but was apparently too short to allow any minor groove interaction by the reactive group. McLaughlin and co-workers showed that Hoechst 33258 conjugated through a hexa(ethyleneglycol) linker (18 atoms in length) significantly stabilized a pyrimidine motif triplex (7 ). In this case the MGB residues had positive charges and the linker structure was not optimized. Since these studies were with MGBs that can bind only in a reversible manner, the possibility that the Hoechst dye can fold back into what remains of the triplex major groove cannot be excluded (7 ). While we did consider that possibility, for these tethered MCPI agents the base specificity and speed of quantitative alkylation of the adjacent duplex regions, along with the requirement for a long linker arm, support the `wrap around' mechanism of binding and alkylation of the duplex minor groove.
At low concentrations ( <= 200 nM) of cross-linking TFO the sequence specificity of alkylation is controlled by the TFO and not by the tethered MGB. TFO binding delivers the alkylating moiety to the proximity of an A/T-rich site located near the end of the triplex region. For (+)-MCPI-DPI conjugates alkylation within the triplex is very fast. The rate limiting step for these reactions was tentatively ascribed to triplex formation. The site of reaction could be as far as 4-8 bp away from the TFO terminus or it could be within the triplex homopurine run itself. The long linkers, like X8, seemed to allow the reactive group to find its most preferred site within that region. Additionally, the `flip around' orientation (Fig. 5 B) required for specific alkylation by the (-)-MCPI-DPI conjugates could be accommodated by these long linkers.
The efficiency of the alkylation reaction is an important characteristic of a reactive conjugate. The (+)-MCPI-DPI-tailed TFOs have the fastest rate of reaction we have yet observed and are powerful reagents for site-directed modification of DNA in purified systems. As noted above, however, the non-TFO-directed reactions of this potent reactive group can reach unacceptable levels in the micromolar concentration range of conjugate. It is likely that this would make these unsuitable for use in cells, where there are many A/T-rich sites with which the reactive MGB may react. Additionally, nuclease degradation of the ODN and consequent release of toxic (+)-MCPI-DPI (39 ) could be a serious problem.Ultimately, these minor groove-reactive conjugates should find their greatest utility if the DPI could be eliminated without sacrificing reaction speed or efficiency. This would provide a reactive group with little, if any, non-specific interactions on its own, but with the full power of its hybridization-triggered reactivity directed by TFO binding. Such projects are now in progress.
F. Nagatsugi, S. Sasaki, P. S. Miller, and M. M. Seidman Site-specific mutagenesis by triple helix-forming oligonucleotides containing a reactive nucleoside analog
Nucleic Acids Res.,
March 15, 2003;
31(6):
e31 - e31.
[Abstract][Full Text][PDF]
CiteULike 
Connotea 
Del.icio.us What's this?
