3′-Modified oligonucleotides by reverse DNA synthesis

Abstract

Reverse DNA oligonucleotide synthesis (i.e. from 5′→3′) is a strategy that has yet to be exploited fully. While utilized previously for the construction of alternating 3′-3′- and 5′-5′-linked antisense oligonucleotides, the use of nucleoside 5′-phosphoramidites has not generally been used for the elaboration of (modified) oligonucleotides. Presently, the potential of reverse oligonucleotide synthesis for the facile synthesis of 3′-modified DNAs is illustrated using a phosphoramidite derived from tyrosine. The derived oligonucleotide was shown to have chromatographic and electrophoretic properties identical with the modified oligonucleotide resulting from the proteinase K digestion of the vaccinia topoisomerase I–DNA covalent complex. The results confirm the nature of the structure previously assigned to this product, and establish the facility with which proteinase K is able to complete the digestion of the polypeptide backbone of the DNA oligonucleotide-linked topoisomerase I.

INTRODUCTION

The chemical synthesis of DNA based upon the phosphoramidite approach developed by Caruthers (1) proceeds in a 3′→5′ direction as a consequence of the use of nucleoside building blocks activated as 3′-O-phosphoramidites. The success of this strategy has contributed importantly to the phenomenal growth in studies that employ synthetic DNA oligonucleotides. In contrast, ‘reverse’ oligonucleotide synthesis (i.e. in a 5′→3′ direction) has not been utilized to nearly the same extent, even though this approach offers a facile route to 3′-modified DNA oligonucleotides. Beaucage and co-workers have used 5′-O-phosphoramidites for the formation of oligonucleotides having alternating 3′-3′ and 5′-5′ linkages (2). More recently, however, with the increasing use of DNA chip technology (3,4), interest has focused upon the synthesis of support-bound, fully deprotected oligonucleotides (5). Such molecules are accessible through the use of 2-(4-nitrophenyl)ethyl/[2-(4-nitrophenyl)ethoxy]carbonyl (npe/npeoc) protecting groups (6).

Our present application of reverse oligonucleotide synthesis centers upon the characterization of the reaction products of topoisomerase I-mediated DNA cleavage. The topoisomerases are ubiquitous cellular enzymes that are responsible for the topological state of chromosomal DNA (7). Through the introduction of transient single-strand breaks in the phosphodiester backbone (8), topoisomerase I facilitates relaxation of supercoiled DNA, thus allowing replication and transcription to proceed (Fig. 1). DNA nicking involves a nucleophilic tyrosine residue in the active site of the enzyme and proceeds by nucleophilic attack at the scissile phosphate (9,10) to afford a transient covalent intermediate. While the accumulation of this DNA–enzyme covalent complex can be quite toxic in vivo (7,11,12), formation and subsequent purification of the complex in vitro by the use of ‘suicide’ substrates has proven to be quite useful in characterization of the mechanism of action of topoisomerase I (13,14).

Figure 1.

Topoisomerase I-mediated DNA cleavage and digestion of the covalent binary complex by proteinase K. Transesterification occurs via the nucleophilic attack of Tyr274 at the phosphodiester backbone of the sequence 5′-(C/T)CCTT↓ to afford an enzyme–DNA adduct linked through a 3′-phosphotyrosine bond (9,10). A free 5′-OH terminating DNA strand is produced concomitantly. Subsequent digestion of the covalent binary complex by the non-specific serine protease proteinase K results in digestion of covalently linked topoisomerase I, putatively affording oligonucleotide tyrosine adduct 1.

In our ongoing efforts to characterize the chemical mechanism of topoisomerases I, we have focused on the products resolved upon PAGE analysis of the topoisomerase I–DNA covalent binary complex following digestion with the non-specific serine protease proteinase K. While it has reasonably been assumed that proteinase K digestion ultimately results in an oligonucleotide terminating at the 3′ end with a tyrosine residue (1), definitive chemical evidence is lacking. Presently, we describe the facile chemical synthesis of the 3′-modified oligonucleotide 1 by reverse DNA oligonucleotide synthesis, and report its identical electrophoretic and chromatographic behavior with enzyme-derived 1.

MATERIALS AND METHODS

General methods

Experiments requiring anhydrous conditions were performed using flame-dried glassware and under a nitrogen atmosphere. Solvents were purchased from Fisher and were used without further purification unless otherwise indicated. Dichloro methane was dried over calcium hydride, distilled and stored over 4 Å molecular sieves and used within several days. ¹H- and ¹³C-NMR spectra were recorded at 300 and 75 MHz, respectively, on either a General Electric QE-300 or a Varian Gemini 300 NMR spectrometer. All 500 MHz NMR spectra were recorded on a Varian Unity 500 MHz NMR spectrometer. All δ values are given in p.p.m. and are relative to tetramethylsilane; J-values are recorded in Hz. ³¹P-NMR spectra were referenced to an external standard of 85% H₃PO₄ in the appropriate deuterated solvent. Thin layer chromatography (TLC) was performed using Merck silica gel F₂₄₅ pre-coated plates with spots visualized using UV or by dipping the plates in a vanillin-staining reagent (15). Silicycle ultra pure silica gel, mesh size 35–75 µm, was used for flash column chromatography. High-resolution mass spectral data were obtained at the Michigan State University Mass Spectrometry Facility supported, in part, by grant DRR-00480 from the Biotechnology Research Technology Program, National Center for Research Resources, National Institutes of Health.

Unmodified oligonucleotides were purchased from Integrated DNA Technologies Inc. (Coralville, IA). The synthetic oligonucleotide was purified by C₁₈ reversed-phase HPLC. Vaccinia topoisomerase I was a generous gift from Dr Stewart Shuman. T4 polynucleotide kinase was purchased from New England Biolabs (Beverly, MA). Proteinase K was obtained from Invitrogen (Carlsbad, CA). [γ-³²P]ATP (3000 Ci/mmol) was purchased from Amersham Biosciences Corp. (Piscataway, NJ).

Synthesis of 3′-O-tyrosyl oligonucleotide 1

N^α-(9-Fluorenylmethoxycarbonyl)-O-monochlorotrityl-l-tyrosine methyl ester (3). One milliliter of TMS-diazomethane (2.0 M solution in hexanes) was added dropwise to a stirred solution containing 1.14 g (1.67 mmol) of N^α-(9-fluorenylmethoxycarbonyl)-O-monochlorotrityl-L-tyrosine (2) (Nova biochem) in 25 ml of CH₂Cl₂, at room temperature, resulting in a clear, yellow reaction mixture. After a period of 24 h, the excess reagent was quenched by the addition of 5 ml of acetic acid until a clear, colorless solution resulted. This solution was immediately concentrated under diminished pressure, co-evaporated with two 25-ml portions of toluene and purified by flash chromatography on a silica gel column (40 × 4 cm). Elution with 3:7→ 1:1 ethyl acetate–hexanes afforded the desired methyl ester 3 as a colorless foam: yield 473 mg (0.681 mmol, 41%); silica gel TLC R_f 0.24 (3:7 ethyl acetate–hexanes); ¹H-NMR (CDCl₃) δ 2.90 (m, 2H), 3.62 (s, 3H), 4.20 (t, 1H, J = 6.9 Hz), 4.35–4.58 (m, 3H), 5.09 (d, 1H, J = 7.7 Hz), 6.58–6.68 (m, 4H), 7.20–7.45 (m, 17 H), 7.56 (t, 2H, J = 6.2 Hz), 7.72–7.81 (m, 3H); ¹³C-NMR (CDCl₃) δ 14.2, 37.4, 47.2, 52.2, 54.7, 66.7, 120.0, 120.8, 125.0, 125.8, 127.1, 127.6, 127.7, 128.2, 128.5, 129.0, 129.4, 132.1, 141.3, 142.8, 142.9, 155.0, 171.8; mass spectrum (FAB) m/z 694.2360 (M + H)⁺ (C₄₄H₃₇NO₅Cl requires 694.2360).

N^α-(9-Fluorenylmethoxycarbonyl)-l-tyrosine methyl ester (4). 100 µl of trifluoroacetic acid was added dropwise to a stirred solution containing 450 mg (0.648 mmol) of tritylated tyrosine derivative 3 in 10 ml of CH₂Cl₂ and 500 µl of triisopropylsilane at room temperature. The reaction mixture was stirred for 25 min. The reaction mixture was concentrated under diminished pressure and then purified by flash chromatography on a silica gel column (40 × 4 cm). Elution with 4:6→ 6:4 ethyl acetate–hexanes afforded the desired phenol 4 as a colorless foam: yield 250 mg (0.599 mmol, 92%); silica gel TLC R_f 0.50 (6:4 ethyl acetate–hexanes); ¹H-NMR (CDCl₃) δ 3.05 (t, 2H, J = 5.4 Hz), 3.74 (s, 3H), 4.21 (t, 1H, J = 6.9 Hz), 4.35 (dd, 1H, J = 10.8, 6.9 Hz), 4.45 (dd, 2H, J = 10.8, 7.3 Hz), 4.64 (dt, 1H, J = 8.5, 5.0 Hz), 4.98 (br s, 1H), 5.25 (d, 1H, J = 8.1 Hz), 6.74 (d, 2H, J = 8.5 Hz), 6.95 (d, 2H, 8.5 Hz), 7.29–7.43 (m, 4H), 7.56 (dd, 2H, J = 7.3, 3.5 Hz), 7.77 (d, 2H, J = 7.3 Hz); ¹³C-NMR (CDCl₃) δ 14.2, 37.4, 47.1, 52.4, 54.9, 66.9, 115.5, 120.0, 125.0, 127.1, 127.7, 130.5, 141.3, 143.7, 143.8, 154.8, 155.6,172.1; mass spectrum (FAB), m/z 418.1655 (M + H)⁺ (C₂₅H₂₄NO₅ requires 418.1654).

O-[N^α-(9-Fluorenylmethoxycarbonyl)-l-phenylalanyl]-O-(2-cyanoethyl)-N,N-diisopropylaminophosphoramidite methyl ester (5). 400 µl of diisopropylethylamine followed by 417 µl (1.80 mmol) of β-cyanoethyl-N,N-diisopropylchlorophosphoramidite was added to a stirred solution of 251 mg (0.601 mmol) of 4 in 2 ml of anhydrous CH₂Cl₂. The reaction mixture was stirred under N₂ for a period of 30 min at room temperature. The reaction mixture was then applied directly to a silica gel column (35 × 3 cm) and washed with 1:1 ethyl acetate–hexanes + 1% pyridine to afford the desired activated tyrosine phosphoramidite as a colorless gum: yield 371 mg (0.600 mmol, 99%); silica gel TLC R_f 0.50 (1:1 ethyl acetate–hexanes); ¹H-NMR (CDCl₃) δ 1.16 (d, 6H, J = 6.9 Hz), 1.22 (d, 6H, J = 6.9 Hz), 2.64 (t, 2H, J = 6.6 Hz), 3.05–3.12 (m, 2H), 3.66–3.78 (m, 3H), 3.72 (s, 3H), 3.87–3.95 (m, 2H), 4.21 (t, 1H, J = 6.9 Hz), 4.31–4.47 (m, 2H), 4.62 (m, 1H), 5.24 (d, 1H, J = 7.7 Hz), 6.98 (dd, 3H, J = 11.9, 8.1 Hz), 7.29–7.35 (m, 2H), 7.38–7.44 (m, 2H), 7.55–7.61 (m, 2H), 7.77 (d, 2H, J = 7.7 Hz); ¹³C-NMR (DMSO-d₆) δ 14.1, 19.7, 19.8, 24.1, 24.2, 24.3, 35.8, 42.9, 43.1, 46.6, 51.9, 55.6, 58.6, 58.9, 65.6, 118.9, 119.4, 120.1, 123.9, 125.2, 127.0, 127.6, 130.2, 131.5, 140.7, 143.7, 149.6, 152.5, 152.6, 155.8, 172.3; ³¹P-NMR (121 MHz, DMSO-d₆) δ 146.4; mass spectrum (ESI), m/z 618.0 (M + H)⁺ (theoretical 618.7).

Oligonucleotide substrates

Oligonucleotide 1 was synthesized on an Applied Biosystems ABI 391 DNA synthesizer using phosphoramidite chemistry (1,5) with protocols optimized for the instrument. Ancillary reagents were purchased from TransGenomic Inc.; controlled pore glass (CPG) support material and reverse (5′) phosphoramidite monomers were purchased from ChemGenes Inc. The synthesis was carried out on a 1 µmol scale in a 5′→3′ direction (Fig. 2) with the 5′-phosphoramidite monomers diluted to a concentration of 0.1 M in anhydrous CH₃CN. Stepwise coupling yields were estimated to be 93% or greater based upon the characteristic trityl absorbance assay in CH₂Cl₂ at 498 nm (ε 22 868 M^–1 cm^–1). Coupling of the tyrosine phosphoramidite (5) was carried out at 0.4 M and was estimated to proceed in a yield of 60% based upon the absorbance at 300 nm after treatment with 5 ml of 20% piperidine in DMF, when compared with the final trityl group absorbance value. Methyl ester saponification was performed by the addition of 250 µl of 1 N NaOH (aq.) to the solid support at room temperature for 2 h. Subsequently, oligonucleotide cleavage from the solid support and base deblocking was accomplished by exposure to 1 ml of NH₄OH (aq.) for 12 h at 60°C. After filtration to remove the CPG and subsequent filtrate concentration by vacuum centrifugation, the synthetic oligonucleotide was purified by reversed-phase HPLC (10 µm Econosphere C₁₈ in a 250 × 10 mm column, using a gradient elution of 95:5→ 70:30, 0.05 M NH₄OAc-acetonitrile at a flow rate of 4.0 ml/min over a period of 30 min) and the DNA was recovered by lyophilization after desalting (Millipore microconcentrator). Analytical re-injections were performed to ensure oligonucleotide purity. MALDI-TOF mass spectral analysis of 1 indicated a m/z of 3831.30 (M^–) (theoretical 3829.65).

Figure 2.

Synthesis of oligonucleotide 1 by reverse DNA oligonucleotide synthesis.

Preparation of 5′ ³²P-end-labeled oligonucleotides

Each synthetic oligonucleotide (10 pmol) was 5′ ³²P-labeled in a reaction mixture (25 µl total volume) of 70 mM Tris–HCl, pH 7.6, containing 10 mM MgCl₂, 5 mM dithiothreitol and 0.01 mCi of [γ-³²P]ATP. The reaction was initiated by the addition of 10 U of T4 polynucleotide kinase (New England Biolabs). The reaction mixture was incubated at 37°C for 2 h and terminated by heat treatment at 65°C for 20 min to inactivate T4 polynucleotide kinase.

Hybridization of substrates

Oligonucleotides were hybridized in a reaction mixture (50 µl total volume) containing 50 mM Tris–HCl, pH 7.5. The solution was heated to 80°C for 5 min and cooled slowly to room temperature under ambient conditions (∼3 h). Due to the low DNA concentration, hybridization mixtures contained 400 fmol of the labeled strand and a 100-fold excess of the unlabeled (non-cleaved) strand to ensure complete hybridization of the labeled DNA.

Oligonucleotide cleavage by topoisomerase I

The 5′ ³²P end-labeled DNA duplex (16 fmol) was treated with 25 ng of vaccinia topoisomerase I in a 20-µl reaction mixture containing 50 mM Tris–HCl, pH 7.5. The reaction mixtures were incubated at 37°C for 1 h, then aliquots were quenched by treatment with proteinase K (1 mg/ml containing 1% SDS, 37°C) for varying lengths of time and analyzed by 20% denaturing PAGE.

Electrophoresis

Denaturing polyacrylamide gel [19% (w/v) acrylamide, 1% (w/v) N,N-methylenebisacrylamide, 7 M urea] electrophoresis was carried out at 500 V for 2–3 h. All gels were run in buffer (89 mM Tris-borate, pH 8.3, containing 1 mM EDTA). The denaturing polyacrylamide gel loading solution contained formamide, 20 mM EDTA, 0.05% bromophenol blue and 0.05% xylene cyanol. Gels were visualized by autoradiography and quantified utilizing a Molecular Dynamics Phosphorimager SP with ImageQuant version 3.2 software.

RESULTS

The characterization of DNA cleavage by topoisomerase I often involves the initial digestion of the DNA–enzyme covalent binary complex with proteinase K to afford a product amenable to analysis by PAGE (Fig. 1). Digestion of the complex has long been assumed to result ultimately in an oligonucleotide having a single amino acid (i.e. tyrosine) attached at the 3′ end of the DNA oligonucleotide. Methods for the synthesis of such 3′-modified oligonucleotides exist, but have largely relied upon the derivatization of a solid support with the tyrosine adduct for subsequent DNA synthesis in the 3′→5′ direction, rather than utilization of the more direct reverse oligonucleotide synthesis (16–18). Here, we report the use of reverse DNA oligonucleotide synthesis as a facile means of constructing an oligonucleotide modified covalently with tyrosine at the 3′ end. The tyrosine–oligonucleotide conjugate was also used to verify the nature of the chemical linkage between topoisomerase I and DNA, the assignment of which to date has been based on the detection of O⁴-phosphotyrosine following acidic hydrolysis of the enzyme–DNA complex (10).

The synthesis of the desired oligonucleotide with a tyrosine bound to the 3′ end followed the protocol shown in Figure 2 and 3. Commercially available Fmoc tyrosine derivative 2 was converted to its methyl ester by exposure to TMS-diazomethane in CH₂Cl₂ for 24 h, affording 3 in 41% yield. The chlorotrityl group was removed from the phenol by treatment with 1% trifluoroacetic acid for 0.5 h to afford 4 in 92% yield. Activation of the phenol with β-cyanoethyl-N,N-diisopropylchlorophosphoramidite in methylene chloride and Hünig’s base for 30 min yielded the desired tyrosine phosphoramidite 5 in 99% yield. This tyrosine phosphate precursor was then successfully appended to the 3′ end of a growing oligonucleotide elaborated by reverse DNA synthesis (Figs 2 and 3) using monomers and CPG support purchased from ChemGenes Inc.

Figure 3.

Pathway used for the synthesis of tyrosine phosphoramidite 5.

To avoid the isolation of the oligonucleotide tyrosine carboxamide as a consequence of the aqueous ammonia treatment required to effect oligonucleotide cleavage from the solid support (16), the resin-bound oligomer was first treated with 1 N NaOH for a period of 2 h to effect tyrosine demethylation prior to oligonucleotide deblocking and cleavage. This strategy has been utilized previously in our laboratory (19) and was also successful in this case as evidenced by the appropriate signal displayed by MALDI-TOF mass analysis. Reversed-phase HPLC purification of the NH₄OH-treated mixture resulted in the isolation of a single peak whose composition was identical with that of the tyrosine-linked oligonucleotide 1 as evidenced by MALDI-TOF mass analysis and subsequent PAGE (Fig. 4B, lane 9).With the authentic 3′-tyrosine-linked oligonucleotide in hand, we sought to verify the identity of the digestion product resulting from the treatment of the vaccinia topoisomerase I–DNA covalent binary complex with proteinase K. Figure 4 depicts the substrate used for the analysis and the autoradiogram of the 20% denaturing polyacrylamide gel used to analyze the incubation mixture with aliquots taken at the time intervals indicated. The multiple bands in lanes 2–8 all contain dodecanucleotide cleaved by topoisomerase I at the single cleavage site, and having covalently attached proteolytic fragments of the topoisomerase. The dodecanucleotide containing the smallest enzyme fragments migrated farthest on the gel. It is apparent that the oligonucleotide containing a single tyrosine was not observed until at least 4 h of incubation with the serine protease. The amount of 1 accumulated steadily over a period of 3 days with intermediate bands still visible even after extended reaction times (data not shown). Co-migration of the digestion product and authentic synthetic standard on PAGE (Fig. 4) and reversed-phase HPLC support the assignment of the structure previously proposed (10) for the initially formed topoisomerase I–DNA covalent binary complex, as well as the product resulting from subsequent proteinase K digestion.

Figure 4.

Visualization of the products of proteinase K digestion of the vaccinia topoisomerase I–DNA covalent complex. (A) The 18-oligonucleotide/30-oligonucleotide partial duplex utilized for the vaccinia topoisomerase I-mediated cleavage reaction. An arrow indicates the cleavage site. The lower case c at the 3′ terminus of the scissile strand denotes the absence of a canonical Watson–Crick base pair. (B) Following digestion with proteinase K for the times indicated, aliquots of the reaction mixture were analyzed by a 20% denaturing polyacrylamide gel. Lane 1, DNA substrate alone; lane 2, 1 h proteinase K digestion; lane 3, 2 h digestion; lane 4, 4 h digestion; lane 5, 6 h digestion; lane 6, 24 h digestion; lane 7, 48 h digestion; lane 8, 72 h digestion; lane 9, synthetic oligonucleotide 1.

DISCUSSION

Vaccinia topoisomerase I has been employed as a model for eukaryotic topoisomerases IB as a consequence of its compact size (314 amino acids) (20) and well defined preferred cleavage sequence [5′-(C/T)CCTT↓] (9). As is true of other topoisomerases I, vaccinia topoisomerase controls the topology of supercoiled DNA and has the ability to relax helical stress by the introduction of transient breaks. Based on the detection of O⁴-phosphotyrosine following acidic hydrolysis of a eukaryotic topoisomerase I–DNA complex (10), it is believed that the mechanism by which this is achieved involves the nucleophilic attack of an active site tyrosine residue upon a DNA phosphate ester moiety. The reaction results in the formation of a transient enzyme–DNA covalent binary complex connected through a 3′-phosphotyrosine linkage. Also formed is a free DNA strand terminating in a 5′-OH group, which can be passaged around the unbroken DNA strand. Following strand passage, the DNA break is resealed through a mechanism that is nominally the reverse of the cleavage reaction (8). While the cleavage and ligation reactions are quite tightly coupled in vivo, thus preventing accumulation of potentially toxic lesions (11,12,21), much use has been made of the ability to uncouple the two half reactions in vitro. For instance, we examined the ability of human topoisomerase I to accommodate modified acceptors in the ligation reaction, using the purified enzyme–DNA covalent binary complex that had been produced by an admixture of topoisomerase I and a ‘suicide’ substrate (14) according to the procedure of Westergaard (13). The Shuman and Hecht laboratories have recently utilized this strategy in examining the cleavage kinetics of methylphosphonate-derived oligonucleotide substrates by vaccinia topoisomerase I (22).

For the investigation of DNA cleavage by topoisomerase I, it is often necessary to digest the enzyme–DNA covalent binary complex in order to facilitate analysis by denaturing PAGE. Typically, exposure of the enzyme–DNA covalent binary complex to the non-specific protease proteinase K results in digestion of the enzyme to afford an oligonucleotide bearing small peptide fragments attached through the 3′-phosphotyrosine linkage (23). To facilitate a more detailed chemical characterization of the topoisomerase I-mediated DNA cleavage product, we have prepared the putative product of complete proteinase K digestion, namely oligonucleotide 1. The vaccinia enzyme was chosen for this study due to its rapid reaction kinetics and well defined sequence specificity of DNA cleavage (24).

While the use of 5′-phosphoramidites has been restricted primarily to the synthesis of 5′-5′- and 3′-3′-linked oligonucleotides (2), recent efforts that exploit automated DNA synthesis in a 5′→3′ direction have gained momentum. DNA chip technology demonstrates the utility of the protocol through the use of the npe/npeoc base protection strategy (6); support-bound, fully deprotected oligonucleotide sequences are now readily available using this strategy. Wagner and Pfleiderer have recently reported the synthesis of a 3′-derivatized oligonucleotide in a similar manner (5). Through the generation of cholesteryl and α-tocopheryl O-phosphoramidites, the authors were able to elaborate 3′-modified oligomers in a fashion analogous to that described herein.

The synthesis of 12mer 1 is not the first 3′-phosphorotyrosine oligonucleotide reported. Sadowski and co-workers have reported the synthesis of a number of similarly functionalized oligonucleotides in their ongoing research centered upon FLP recombinases (18). Their major strategy involves initial derivatization of the synthesis support with a tyrosine derivative, followed by oligonucleotide synthesis in the usual (3′→5′) fashion (17,18). As demonstrated in this study, the tyrosine phosphoramidite required for reverse DNA oligonucleotide synthesis was readily accessible (Fig. 3) and was employed in the final coupling following elaboration of the requisite dodecanucleotide by reverse (i.e. 5′→3′) synthesis. The stepwise yield for reverse oligonucleotide synthesis was 93%, and the tyrosine was appended in 60% yield (Fig. 2). MALDI-TOF mass analysis confirmed the identity of synthesized oligonucleotide 1, permitting its utilization as an authentic standard for comparison of electrophoretic mobility against the proteinase K-mediated digestion of the vaccinia topoisomerase I-induced enzyme–DNA covalent binary complex.

Examination of the time course of proteinase K digestion revealed two interesting results. First, proteinase K digests the final peptide fragment attached to the DNA oligonucleotide rather slowly. Indeed, only after 4 h of proteinase K treatment could a species that co-migrated with the authentic synthetic oligonucleotide 1 be observed (Fig. 4). Interestingly, while 1 could be resolved from the same dodecanucleotide having a 3′-OH group, the difference in migration was not pronounced, and the two species could not be separated by C₁₈ reversed-phase HPLC. Apparently, the zwitterionic 3′ functionality in 1 has only a limited effect on the migration of the oligonucleotide.

In summary, we have demonstrated the utility of reverse oligonucleotide synthesis as a rapid and facile protocol for the construction of 3′-modified DNA fragments. We have synthesized a standard identical in chromatographic and electrophoretic behavior with the digestion product resulting from the exposure of the vaccinia topoisomerase I–DNA covalent binary complex to the non-specific serine protease proteinase K. This has provided more direct evidence for both the chemical structure of the linkage in the topoisomerase I–DNA covalent binary complex, and for the structure of the product resulting from proteinase K treatment. Finally, it may be noted that this synthetic strategy can be employed for the preparation of oligonucleotide analogs having a series of modifications at the 3′ end by utilizing portions of a single protected oligonucleotide preparation with different phosphoramidites in the final coupling step (Fig. 2), thus obviating the need for the parallel synthesis (18) of such species.