| Nucleic Acids Research | Pages |
Linkers designed to intercalate the double helix greatly facilitate DNA alkylation by triplex-forming oligonucleotides carrying a cyclopropapyrroloindole reactive moiety
Introduction
Materials And Methods
6-(5-Azidopentoxy)-3-(5-hydroxypentoxy)pyridazine (5)
Pyridazine phosphoramidite (8)
8-(5-Azidopentoxy)-4-(5-hydroxypentoxy)quinoline (9)
Quinoline phosphoramidite (10)
4-(5-N-Phthalimidopentoxy)-9-thioacridine (11)
Acridine phosphoramidite (12)
16-Azido-6,11-dioxahexadecane-1-ol (13)
16-Trifluoracetamido-6,11-dioxahexadecane-1-ol (14)
16-Trifluoracetamido-6,11-dioxahexadecane-1-(cyanoethyl N,N-diisopropylphosphoramidite) (15)
Oligonucleotide synthesis and derivatization
Preparation of a 203 bp DNA target and reactions with the (+)CPI-TFO conjugates
Results And Dicussion
Conclusion
References
Linkers designed to intercalate the double helix greatly facilitate DNA alkylation by triplex-forming oligonucleotides carrying a cyclopropapyrroloindole reactive moiety
Received March 8, 1999; Revised and Accepted May 31, 1999
DDBJ/EMBL/GenBank accession no. AF154273
ABSTRACT Triplex-forming oligonucleotides (TFOs) bind sequence-specifically in the major groove of double-stranded DNA. Cyclopropapyrroloindole (CPI), the electrophilic moiety that comprises the reactive subunit of the antibiotic CC-1065, gives hybridization-triggered alkylation at the N-3 position of adenines when bound in the minor groove of double-stranded DNA. In order to attain TFO-directed targeting of CPI, we designed and tested linkers to `thread' DNA from the major groove-bound TFO to the minor groove binding site of CPI. Placement of an aromatic ring in the linker significantly enhanced the site-directed reaction, possibly due to a `threading' mechanism where the aromatic ring is intercalated. All of the linkers containing aromatic rings provided efficient alkylation of the duplex target. The linker containing an acridine ring system, the strongest intercalator in the series, gave a small but clearly detectable amount of non-TFO-specific alkylation. An equivalent-length linker without an aromatic ring was very inefficient in DNA target alkylation.
INTRODUCTION
Triplex-directed covalent modification of specific nucleotides in genes can be a means of accomplishing site-directed mutagenesis in living cells with high precision (1,2). This approach to gene-directed therapeutics faces hurdles associated with efficient triplex formation, which is limited to polypurine tracts (3), and restrictions imposed by the sequence specificity of the covalent modifying group. To date, most work on mutagenesis directed by triplex-forming oligonucleotides (TFOs) has used a conjugated photoactivatable psoralen as the reactive group. This moiety reacts predominately with 5[prime]-TpA sites adjacent to the triplex region (4).
We have been investigating methods for specific covalent modification of DNA directed by a TFO bearing a hybridization-triggered crosslinking group. These reactive groups, such as the cyclopropapyrroloindole (CPI; Fig. 1) subunit of the potent anti-tumor antibiotic CC-1065, do not react significantly with other biological nucleophiles, but become highly reactive when bound to A/T-rich regions in the minor groove of double-stranded DNA (dsDNA) (5,6). For specific targeting on dsDNA, we had previously attached CPI derivatives to a TFO via long (>50 atom) linkers to reach from the major to the minor groove by wrapping around one DNA strand (7). Site-directed alkylation of A/T-rich regions adjacent to the triplex region was found to be rapid and efficient only if the CPI bore an additional binding subunit (7,8) to enhance affinity for the minor groove. This dimeric structure gave, however, an unacceptable amount of non-TFO-specific alkylation because of the relatively high affinity of the binding subunit for any A/T-rich site. Therefore, a means of anchoring the CPI monomer directly into the minor groove after TFO binding was required in order to achieve efficient and specific TFO-directed alkylation.
Figure 1. (A) Proposed mode of binding of TFO with threading linker to a CPI bound in the minor groove. (B) Structures of the threading linkers and CPI. (C) Sequence of the TFO used and the target duplex in the vicinity of TFO binding, showing the interaction of the linker and CPI (dark symbol).
It is well established that conjugated intercalating moieties can stabilize both DNA duplexes (9-12) and triplexes (3,13-16). In the case of triplex stabilization, intercalating residues are usually attached to either end of the TFO via a 5-6 atom flexible linker (13,14) which is of sufficient length to allow for effective interaction with the triplex structure, or with the nearest base pairs at the duplex/triplex junction, depending on the structure and binding preference of the intercalator.
Certain naturally occurring intercalating agents (17,18), particularly the pluromycins (19), are known to `thread' the DNA structure placing carbohydrate residues into both grooves. The synthetic amsacrine-4-carboxamides (20) are known to thread the DNA double helix so as to intercalate the acridine chromophore, leaving the carboxamide moiety and methanesulfonanilino group positioned within the minor and major grooves of the double helix, respectively. The concept of inserting bulky residues or side chains through the DNA helix as a result of intercalation is also well known (21-23) and many so-called `threading' intercalators have been described (24-28).
We sought to use this `threading' effect to deliver the CPI moiety into the minor groove from the end of a TFO already bound to a specific site in the major groove, as shown in Figure 1A. Linkers with mono- (2), bi- (3) and tri-cyclic (4) aromatic rings, inserted in the middle of an aliphatic chain, were designed and used to prepare CPI-TFO conjugates. Providing direct evidence for placement of CPI in the minor groove by the `threading' mechanism (i.e. NMR solution structure) is beyond the scope of the current study. This report describes the site-specific and nearly quantitative alkylation of targeted DNA by the newly designed TFO derivatives and discusses the indirect evidence that supports a putative `threading' mechanism.
MATERIALS AND METHODS
All air and water sensitive reactions were carried out under a slight positive pressure of argon. Anhydrous solvents were obtained from Aldrich (Milwaukee, WI). Enzymes were obtained from Sigma (St Louis, MO). Flash chromatography was performed on 230-400 mesh silica gel. Thin-layer chromatography (TLC) was run on silica gel 60 F-254 (EM Reagents) aluminum-backed plates. UV-visible absorption spectra were recorded in the 200-400 nm range on a Lambda 2 (Perkin Elmer, CT) spectrophotometer. 1H NMR were obtained at 300 MHz on a Varian VXR-300 spectrometer. The chemical shift values are expressed in [delta] (parts per million) relative to tetramethylsilane as an internal standard. Elemental analyses were performed by Quantitative Technologies Inc. (Boundbrook, NJ).
6-(5-Azidopentoxy)-3-(5-hydroxypentoxy)pyridazine (5)
Sodium metal (4.95 g, 0.225 mmol) was dissolved in 45 ml of 1,5-pentanediol at 100°C. 3,6-Dichloropyridazine (5.0 g, 33.56 mmol) was added and the solution was stirred for 3 h at 100°C and then poured into 300 ml of water and extracted with 500 ml of ethyl acetate. The extract was washed with water (4 × 200 ml), dried over sodium sulfate and evaporated. The residue was crystallized from methylene chloride affording 3.7 g (39% yield) of the bis-substitution product which was reacted with p-toluenesulfonyl chloride (2.41 g, 12.66 mmol) in 75 ml of anhydrous pyridine at 25°C for 12 h. The solution was evaporated and the residue was purified by silica gel chromatography (20% hexane in ethyl acetate) giving 1.1 g (20% yield) of the mono-tosylated product as a homogenous oil. To a solution of this product in 20 ml of dimethylformamide (DMF) was added sodium azide (0.816 g, 12.55 mmol) and ammonium chloride (0.816 g, 15.26 mmol). The mixture was stirred for 4 h at 75°C and then cooled to room temperature, filtered and evaporated. The residue was dissolved in 200 ml of ethyl acetate, washed with 2% sodium bicarbonate solution (3 × 175 ml), dried over sodium sulfate and evaporated affording 5 as a homogenous oil: 0.70 g (90%) yield; TLC (ethyl acetate), Rf = 0.56; 1H NMR (CDCl3) [delta] 6.90 (2H, s, aromatic), 4.41 (4H, m, ArO-CH2's), 3.68 and 3.30 (4H, 2 × t, J = 6.2 and 6.7 Hz, respectively, -CH2-OH and CH2-N3, no assignments made), 1.80-1.52 (12H, 2 × m, methylene protons).
Pyridazine phosphoramidite (8)
A mixture of 5 (0.695 g, 2.25 mmol) in 50 ml of methanol containing 75 mg of 10% pd/C was shaken under 35 psi of hydrogen gas for 2.5 h and then filtered through Celite. The filtrate was evaporated affording the amino derivative as a white solid. To a solution of the amine (0.695 g, 2.25 mmol) in 8 ml of ice-cold anhydrous pyridine was added triethylamine (0.47 ml, 3.38 mmol) followed by trifluoroacetic anhydride (0.48 g, 3.38 mmol). The solution was stirred for 15 min at room temperature and then diluted with 30 ml of 5% sodium bicarbonate solution. Stirring was continued for an additional 20 min before the mixture was diluted with 100 ml of water and extracted with ethyl acetate (2 × 200 ml). The combined extracts were dried over sodium sulfate and evaporated. The residue was purified by silica gel chromatography eluting with 20% hexane in ethyl acetate. The trifluoroacetamide derivative 7 was precipitated from ethyl acetate/hexane: 0.46 g (54% yield). To a solution of 7 (0.39 g, 1.04 mmol) in 20 ml of anhydrous methylene chloride, containing 0.54 ml of N,N-diisopropylethylamine, was added 2-cyanoethyl diisopropylchloro-phosphoramidite (0.40 ml, 1.76 mmol). The solution was stirred under argon for 30 min and then treated with 2 ml of methanol, diluted with 300 ml of ethyl acetate and washed with 200 ml of 5% sodium bicarbonate solution. The organic solution was dried over sodium sulfate, filtered and evaporated. The residual oil was purified by silica gel chromatography eluting with a 20-60% gradient of ethyl acetate in hexane. The pure fractions were evaporated affording 8 as a homogenous oil: 0.43 g, (72%) yield; TLC (1:1, ethyl acetate/hexane), Rf = 0.65; 1H NMR (DMSO-d6) [delta] 9.43 (1H, br t, trifluoroacetamido N-H), 7.13 (2H, s, aromatic), 4.30 (4H, overlapping triplets, ArO-CH2's-), 3.79-3.49 (6H, m, isopropyl CH's, -CH2-CH2CN and -CH2-OP), 3.19 (2H, m, -CCH2-NHCO), 2.75 (2H, t, J = 5.8 Hz, -CH2-CH2CN), 1.81-1.35 (12H, 2 × m, methylene protons), 1.11 (12H, overlapping doublets, isopropyl CH3's). Anal. Calcd for C25H41F3N5O5P: C, 51.81; H, 7.13; N, 12.08. Found: C, 52.01; H, 6.85; N, 11.86.
8-(5-Azidopentoxy)-4-(5-hydroxypentoxy)quinoline (9)
To a solution of 4,8-dihydroxyquinoline (5.2 g, 32.3 mmol) in 40 ml of 25% DMF in acetone was added potassium carbonate (4.4 g, 32.3 mmol) followed by 5-azido-1-iodo-pentane (7.7 g, 32.3 mmol). The reaction mixture was stirred at reflux for 5 h, filtered and evaporated to dryness. The 8-O-alkylation product was purified by silica gel chromatography eluting with 10% methanol in ethyl acetate: 3.76 g (43% yield); TLC (10% methanol in ethyl acetate), Rf = 0.35. This product was dissolved in 120 ml of phosphorous oxychloride and stirred at reflux for 1 h. The reaction solution was cooled to room temperature and then poured into 1.5 liters of ice-water. The resulting solution was neutralized (pH 6.5) with concentrated ammonium hydroxide and then extracted with ethyl acetate (4 × 400 ml). The pooled extracts were dried over sodium sulfate, filtered and evaporated. The chloroquinoline was purified by silica gel chromatography (eluting with ethyl acetate) giving an oil: 3.56 g (89%) yield; TLC (ethyl acetate), Rf = 0.74.
Sodium metal (0.35 g, 15.2 mmol) was dissolved in 3 ml of 1,5-pentanediol at 100°C. This solution was then added to the chloroquinoline (1.0 g, 3.45 mmol) and stirred at 100°C for 40 min. The mixture was cooled to room temperature and neutralized by addition of 0.86 ml of acetic acid. The mixture was suspended in 4 ml of methanol and crystallized by addition of water: 1.12 g (90%) yield; TLC (10% methanol in ethyl acetate), Rf = 0.44; 1H NMR (DMSO-d6) [delta] 8.66 (1H, d, J = 6.0 Hz, 2-H), 7.66 (1H, d, J = 8.4 Hz, 5-H), 7.43 (1H, t, J = 8.4 Hz, 6-H), 7.17 (1H, d, J = 7.8 Hz, 5-H), 7.02 (1H, d, J = 5.2 Hz, 3-H), 4.41 (1H, t, J = 3.0 Hz, -OH), 4.21 and 4.12 (4H, 2 × t, J = 6.3 Hz for both triplets, ArO-CH2's-), 3.38 (4H, m, -CH2-N3 and -CH2-OH), 1.86 and 1.60 (12H, 2 × m, methylene protons). Anal. Calcd for C19H26N4O3: C, 63.67; H, 7.31; N, 15.63. Found: C, 63.82; H, 7.12; N, 15.72.
Quinoline phosphoramidite (10)
Compound 9 was reduced and reacted with trifluoroacetic anhydride as described for the preparation of 7. The trifluoroacetimide product was precipitated from ethyl acetate/ether (72% yield) without column chromatography purification; TLC (10% methanol in ethyl acetate), Rf = 0.67. Phosphoramidite 10 was then prepared by the procedure described for the synthesis of compound 8. The crude product was purified by silica gel chromatography eluting with 20% hexane in ethyl acetate (2% triethylamine). The pure phosphoramidite fractions were evaporated affording 10 as a homogenous oil (76% yield); TLC (ethyl acetate), Rf = 0.51; 1H NMR (DMSO-d6) [delta] 9.48 (1H, br t, J = 5.5 Hz, trifluoroacetamide N-H), 8.65 (1H, d, J = 5.1 Hz, 2-H), 7.65 (1H, d, J = 7.5 Hz, 5-H), 7.41 (1H, t, J = 7.9 Hz, 6-H), 7.16 (1H, d, J = 7.0 Hz, 7-H), 7.02 (1H, d, J = 5.3 Hz, 3-H), 4.22 and 4.12 (4H, 2 × t, J = 6.3 Hz for both triplets, ArO-CH2's), 3.64 (6H, m, -CH2-OP, isopropyl CH's and -CH2-CH2CN), 3.22 (2H, m, -CH2-NHCO-), 2.74 (2H, t, J = 6.0 Hz, -CH2-CH2-CN), 1.87 and 1.60 (12H, 2 × m, methylene protons), 1.11 (12H, m, isopropyl CH3's). Anal. Calcd for C30H44F3N4O5P·0.3 H2O: C, 56.83; H, 7.09; N, 8.84. Found: C, 56.76; H, 6.84; N, 8.59.
4-(5-N-Phthalimidopentoxy)-9-thioacridine (11)
To a solution 4,9-dihydroxyacridine (4.88 g, 23.09 mmol) in 50 ml of 40% DMF in acetone was added potassium carbonate (3.51 g, 25.39 mmol) followed by N-(5-bromopentyl)-phthalimide (6.84 g, 23.09 mmol). The mixture was stirred at 75°C for 6 h and then diluted with 200 ml of water. The mixture was extracted with ethyl acetate (2 × 350 ml) and the combined extracts were dried over sodium sulfate and evaporated. The residue was triturated in boiling ether and the solid that formed (the 4-O-alkylation product) was filtered and rinsed with ether: 6.9 g (70%) yield; TLC (ethyl acetate), Rf = 0.83. This solid was dissolved in 225 ml of phosphorous oxychloride and stirred at reflux for 1 h. The solution was evaporated to dryness in vacuo affording the crude chloroacridine as a solid which was triturated in 300 ml of 10% sodium bicarbonate solution. The remaining solid was filtered, washed with water and dried. The chloroacridine was vigorously stirred in 250 ml of methanol, saturated with sodium hydrogen sulfide, for 20 min at room temperature. Water (100 ml) was added and the solid was filtered, rinsed with water and dried. The crude product was recrystallized from methanol affording 6.4 g of the 4-thioacridine derivative 11; TLC (50% hexane in ethyl acetate), Rf = 0.65; 1H NMR (DMSO-d6) [delta] 11.75 (1H, s, N-H thioacridone tautamer), 8.86 (1H, d, J = 8.5 Hz, 8-H), 8.42 (1H, d, J = 8.5 Hz, 1-H), 8.01 (1H, d, J = 8.7 Hz, 5-H), 7.79 (5H, m, 6-H and phthalimido protons), 7.37 (2H, m, 3- and 7-H), 7.27 (1H, t, J = 7.8 Hz, 2-H), 4.26 (2H, t, J = 6.5 Hz, ArO-CH2-), 3.63 (2H, t, J = 6.7 Hz, -CH2-N-), 1.94, 1.72 and 1.55 (6H, 3 × m, methylene protons). Anal. Calcd for C26H22N2O3S·0.4 H2O: C, 69.44; H, 5.11; N, 6.23. Found: C, 69.46; H, 4.95; N, 6.10.
Acridine phosphoramidite (12)
To a solution of 11 (2.0 g, 4.52 mmol) in 7.5 ml of anhydrous DMF was added triethylamine (2 ml, 14.4 mmol) followed by 6-bromohexanol (1.8 ml, 13.8 mmol). The reaction solution was stirred under argon at 90°C for 2.5 h and then evaporated to dryness. The crude thioether product was triturated in boiling methanol and filtered: 1.78 g (73%) yield; TLC (ethyl acetate), Rf = 0.5. Phosphitylation of this product followed the procedure used for the preparation 8. The crude phosphoramidite was purified by silica gel chromatography eluting with 35% ethyl acetate in hexane. The pure fractions were evaporated affording a homogenous oil: 0.46 g, (84%) yield; TLC (1:1, ethyl acetate/hexane), Rf = 0.67. 1H NMR (DMSO-d6): [delta] 8.70 (1H, d, J = 8.8 Hz, 8-H), 8.24 (1H, d, 8.9 Hz, 1-H), 8.11 (1H, d, J = 8.5 Hz, 5-H), 7.81 (5H, m, 6-H and phthalimido protons), 7.71 (1H, m, 7-H), 7.58 (1H, t, J = 7.8 Hz, 2-H), 7.19 (1H, d, J = 7.5 Hz, 3-H), 4.21 (2H, t, J = 6.5 Hz, ArO-CH2-), 3.64 (4H, m, CH2-N phthalimide and CH2-CH2CN), 3.46 (4H, m, CH2-OP and isopropyl CH's), 2.95 (2H, t, J = 6.3 Hz, ArS-CH2-), 2.72 (2H, t, J = 5.5 Hz, CH2-CH2-CN), 1.93, 1.75, 1.54, 1.34 and 1.13 (14H, 5 × m, methylene protons), 1.08 and 1.02 (12H, 2 × d, J = 6.8 Hz for both doublets, isopropyl -CH3's). Anal. Calcd for C41H51N4O5PS·0.75 H2O; C, 65.18; H, 7.00; N, 7.41. Found C, 65.48; H, 6.74; N, 7.00.
16-Azido-6,11-dioxahexadecane-1-ol (13)
1,4-Dibromobutane (3.00 g, 13.89 mmol) was converted to 13 by the three-step process described for the preparation of 5 with an overall yield of 18%. 1H NMR (CDCl3): [delta] 3.65 (2H, t, J = 6.6 Hz), 3.40 (8H, m), 3.27 (2H, t, J = 6.9 Hz), 1.60 (12H, m), (4H, m).
16-Trifluoracetamido-6,11-dioxahexadecane-1-ol (14)
Compound 13 (0.87 g, 3.02 mmol) was converted to the trifluoro-acetamide derivative by the general procedure described for the synthesis of 7 (85% overall yield). Crude 14 was purified by silica gel chromatography eluting with 20% hexane in ethyl acetate. 1H NMR (CDCl3): [delta] 3.65 (2H, t, J = 6.6 Hz), 3.40 (8H, m), 3.27 (2H, t, J = 6.9 Hz), 1.60 (12H, m), 1.43 (4H, m).
16-Trifluoracetamido-6,11-dioxahexadecane-1-(cyanoethyl N,N-diisopropylphosphoramidite) (15)
Compound 14 (0.85 g, 2.38 mmol) was phosphitylated by the procedure described for the preparation of 8. The crude phosphoramidite was purified by silica gel chromatography eluting with 40% hexane in ethyl acetate (2% triethylamine) affording a foam: 1.8 g (91%) yield; 31P NMR (DMSO-d6) [delta] 146.57 (singlet).
Oligonucleotide synthesis and derivatization
ODN syntheses were carried out on an Applied Biosystems Model 394 DNA synthesizer utilizing the 1 µmol coupling program supplied by the manufacturer. Standard reagents for the [beta]-cyanoethyl phosphoramidite coupling chemistry were purchased from Glen Research (Sterling, VA). Phosphoramidites 8, 10, 12 and 15 were used at the last step of automatic synthesis. After ammonia deprotection, the ODNs were HPLC purified, detritylated when it was necessary, and precipitated as described previously (29). (+)CPI-ODN conjugates were synthesized, purified and characterized as described in (8).
Preparation of a 203 bp DNA target and reactions with the (+)CPI-TFO conjugates
The target duplex used in the study was a 203 bp DNA fragments of 0302 allele of Human MHC class II HLA DQ[beta]*1 containing a 33mer homopurine run in an intron region (30). We found the sequence recorded in GenBank (accession no. K01499) to be incorrect. The correct sequence (accession no. AF154273) is shown below (insertions are bold and underlined; T to G substitutions are italicized and underlined):
...CTGGAAGAGAAGGAGAGAGGAGAGGAAAGAGGAGACAAAGTGTACATTTACTACCAGTGACAGGACAA-
AGTGAGCATGGGGTTATTTTTGAAGATACGAATTTCTCCAGAGACACAGCAGGATTTGTCATTTAGGCGTGC-
CCCAAGACTTTGCCTGGACTAAA...
The 203 bp DNA target was prepared by PCR using forward d(GCATTCCCATTTATCTTTTAGTGA) and reverse d(AC-GCCTAAATGACAAATCCTGCT) primers. The 5[prime]-end of either primer was blocked with a hexanol tail to enable strand selective 32P-labeling. The 100 µl PCR mixture included 10 µl of 10× Promega PCR buffer with no magnesium; 20 µl of 1 mM four dNTP mixture; 5 µl of 25 mM MgCl2; 5 µl of each gel-purified primer (10-5 M); ~0.5 µg of human genomic DNA extracted from HT 29 cells (obtained from ATCC and containing one copy of the 0302 allele of Human MHC class II HLA DQ[beta]* 1) and 0.7 U of a Taq DNA polymerase (Promega, WI). HT 29 DNA was extracted by using a Wizard Genomic DNA Purification Kit (Promega, WI). PCR was performed in an Ericomp Thermocycler. After 3 min of denaturation at 94°C, 30 cycles of 1 min denaturation at 94°C, 1.5 min of annealing at 55°C and 2 min of extension at 72°C were performed followed by 10 min of final extension at 72°C. The target amplicon was purified by a Microcon microconcentrator (Amicon, Inc., MA), precipitated in alcohol, washed with 70% alcohol, dried, and resuspended in water.
A 203 bp target duplex (2 × 10-8 M), 32P-labeled in one strand, was incubated with the CPI-TFO conjugates (2 × 10-7 M) in 30 µl of 140 mM KCl, 10 mM MgCl2, 1 mM spermine and 20 mM HEPES, pH 7.2 at 37°C. Aliquots (10 µl) were taken after 1, 3 and 18 h, incubated at 90-95°C for 15 and 20 min after addition of 10% piperidine (100 µl). Reactions were dried in vacuo, resuspended in a loading solution (8 M urea), and analyzed by 8% denaturing polyacrylamide gel (PAG). The gel was transferred on a Watman 3MM paper, dried and imaged with GS-250 Molecular Imager (Bio-Rad Laboratories, CA). The experiment was performed in duplicate and the main results shown in Figure 2 were confirmed a couple of times on a shorter, 65mer synthetic duplex target (data not shown).
Figure 2. Specific TFO-directed alkylation by the minor groove-reactive CPI. A 203 bp target duplex (2 × 10-8 M), 32P-labeled in the strand with the purine tract, was incubated at 37°C with 2 × 10-7 M of the indicated TFO in 140 mM KCl, 10 mM MgCl2, 1 mM spermine and 20 mM HEPES, pH 7.2 for 1, 3 and 18 h at 37°C. After heat/piperidine treatment, products of the target cleavage were analyzed by 8% denaturing PAG. Lane 1, no TFO (18 h incubation); lane 2, (dGdA)11 with CPI and linker 3 on the 5[prime]-end (non-sequence-specific control; 18 h incubation); other lanes have TFO with indicated linker for 1, 3 or 18 h incubation. The radioactivity in all lanes was adjusted such that all were 15% of the mean, except the ODN 3/18 h lane, which contained ~50% of the mean counts. The numbers under the prominent band(s) in the target site are the efficiency of cleavage, found by dividing the density of radioactivity in these band(s) by the total density in the lane. Arrows indicate sites of non-TFO directed alkylation with linker 1, which was ~1-2% of the total density.
RESULTS AND DICUSSION
The design of linkers that could potentially enable a threading mechanism incorporated a short (5-6 atoms) aliphatic chain on the 5[prime]-phosphate of the TFO. This linker connects to an aromatic ring that could intercalate adjacent to the TFO binding site (31,32). Another short alkyl chain attached to the opposite side of the aromatic ring serves to bridge to the (+)-CPI monomer, attached via an amide linkage (8). It was proposed that the aromatic ring system would facilitate DNA threading through the double helix by the process of intercalation. For solid phase synthesis, linkers with mono- (2), bi- (3) and tri-cyclic (4) aromatic rings were prepared (Fig. 1B), with a phosphitylated hydroxyalkyl chain on one side and a blocked aminoalkyl chain on the other. The synthesis of these aromatic linkers is shown in Scheme 1. An equivalent-length (16 atom) straight chain linker 1, not containing an aromatic ring, was also prepared. All linkers were attached to the 5[prime]-end of the TFO during solid phase synthesis. The (+)-CPI was introduced post-synthesis via the tetrafluorophenyl ester of succinyl-(+)-CPI as described earlier (8).
Scheme 1. (a) HO(CH2)5ONa; (b) TsCl, Pyr; (c) NaN3, DMF; (d) Pd/C-H2, MeOH; (e) Trifluoroacetic anhydride; (f) (iPr)2NP(Cl)OCH2CH2CN, CH2Cl2; (g) Br(CH2)5N3, K2CO3; (h) POCl3; (i) Br(CH2)5NPht, K2CO3; (j) NaHS, MeOH; (k) Br(CH2)6OH, TEA.
Each of the (+)-CPI-linker-ODN conjugates were evaluated for speed and efficiency of alkylation. The target was a 203 bp duplex which contained, in addition to the homopurine tract, multiple A/T-rich tracts to allow assessment of the sequence specificity of the alkylation. The TFO chosen was a G/A-containing 22 base oligodeoxyribonucleotide targeted to the homopurine tract shown in Figure 1C. This position was chosen because of an adjacent (dA)3 tract which could serve as an alkylation site for (+)-CPI (7,8).
To determine the alkylation site of the CPI attached to these linkers, the conjugated TFOs were incubated with the target duplex for 1, 3 or 18 h, followed by heat/piperidine cleavage of the target strand at the alkylation site (8). Figure 2 shows the results of the study. Regardless of the mechanism of the action, `threading' through the duplex base pairs or `wrapping around' one of the duplex strands (8), placement of CPI in the minor groove after TFO binding would afford alkylation of the adenines in the target A3 tract located near the duplex/triplex junction. Indeed, all TFO conjugates containing an aromatic ring system in their linkers (2-4) showed high efficiency of targeted alkylation (81-90%) at the expected adenine tract after 18 h. Linkers 2 and 3, containing mono- and bicyclic aromatic rings, respectively, afforded high target specificity and only the distal adenines from the TFO binding site were alkylated. Linker 4 was also effective in alkylating the targeted adenines, but additionally gave trace amounts of DNA alkylation in non-targeted regions (see arrows in Fig. 2) suggesting that the strong DNA binding affinity of the three-membered acridine ring system, to some extent, competes with the sequence-specific binding of the TFO.
As was expected from the design of CPI-TFO conjugates, only the strand which contains the homopurine run underwent alkylation (no alkylation products were observed when the opposite strand was labeled-data not shown). As the number of aromatic rings increases in the intercalating core, from 1 to 3 (i.e., pyridazine to acridine), the CPI moiety tends to shift its base specificity of alkylation. The CPI of linker 2 reacts predominantly with the most distant adenine in the A3 tract while an increasing amount of alkylation product at the central adenine is observed for linkers 3 and 4, respectively. The nature of this effect is not clear, but could be attributed, at least in part, to subtle differences in the binding mode of the different aromatic ring systems with DNA. In terms of the putative `threading' model, different intercalators bind within the base pair stacks of duplex DNA with different orientations so as to optimize the energetically favorable hydrophobic interaction. This effect may have a bearing on the positioning of the covalently attached CPI moiety within the A3 tract. In any event, the observation that the alkylating specificity within the adenine tract is influenced by the choice of intercalating agent is noteworthy and may potentially be employed in the design of more highly specific alkylating linkers.
The non-aromatic linker (1) showed only a modest amount of DNA alkylation, predominantly at the guanine nearest to the triplex/duplex junction. This last result was rather curious considering the high affinity of CPI for the minor groove A3 site. A low-level of CPI alkylation product has been observed at the minor groove N-3 position of guanine in cases where a high-affinity adenine tract was not available (33). In this study, however, linker 1 is approximately the same length as the aromatic linkers and should similarly be able to reach the high-affinity A3 site, assuming that the CPI of this conjugate can reach the minor groove of the duplex target. If the CPI of 1 is unable to access the minor groove, then covalent attachment of CPI to a TFO bound in the major groove of DNA will result in a high local concentration of the CPI alkylating agent in the vicinity of the N-7 nucleophiles of the duplex target. The N-7 position of guanine is more nucleophilic than adenine (34) and the major-groove reaction of CPI should favor formation of the G-adduct. The absence of any alkyation product at the same guanine residue, in addition to the high efficiency of alkylation afforded by the aromatic linkers, suggests that CPI alkyation reactions with linkers 2-4 are taking place in the minor groove.
Comparative analysis of the aromatic linkers described in this study with the aliphatic `wrap around' linkers described in our previous work (8) may provide mechanistic insight. The same TFO sequence and DNA target was employed in both studies. In our previous report, we found that a long aliphatic linker (52 atom) was required to reach from the major groove-bound TFO to the minor groove A3 site by wrapping around one DNA strand. While a high level of target specificity at the A3 site was obtained, only a low-level of alkylation product (8%) was observed. A shorter linker (32 atom) failed to produce a significant quantity of alkylation product. The aromatic linkers examined in this study are much shorter than the linkers previously studied (21-22 atoms bridging the CPI moiety to the TFO), yet nearly quantitative alkylation was observed at the targeted A3 site. These results argue strongly in favor of a threading mechanism of action for CPI alkylation with linkers 2-4.
CONCLUSION
Although the indirect evidence in this study clearly supports the threading mechanism of DNA target alkylation by the CPI-TFO conjugates employing the aromatic linkers, this mode of the action remains to be proven directly, perhaps by elucidation of the solution structure of the crosslinked product by NMR. Regardless of any uncertainty in mechanism, current design of these linkers allows us to reach almost quantitative modification of a target DNA while maintaining remarkable specificity of action for these CPI-TFO conjugates, particularly when employing linkers 2 and 3. Such reagents should prove valuable and find utility in the further development of many diagnostic and therapeutic applications, especially ones involving the targeting of genomic DNAs.
REFERENCES
*To whom correspondence should be addressed. Tel: +1 425 821 7535; Fax: +1 425 825 0306; Email: rdempcy{at}epochpharm.com
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  W. Pils and R. Micura Flexible non-nucleotide linkers as loop replacements in short double helical RNAs Nucleic Acids Res., May 1, 2000; 28(9): 1859 - 1863. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||