Induction of RNase H Activity by Arabinose−Peptide Nucleic Acid

Bioconjugate Chem. 2007, 18, 421−430

421

Induction of RNase H Activity by Arabinose-Peptide Nucleic Acid Chimeras Be´ne´dicte M. Paˆtureau,† Robert H. E. Hudson,*,‡ and Masad J. Damha*,† Department of Chemistry, McGill University, Montreal, Quebec, Canada H3A 2K6, and Department of Chemistry, The University of Western Ontario, Ontario, Canada N6A 5B7. Received September 25, 2006; Revised Manuscript Received November 27, 2006

We report the syntheses of chimeras of peptide nucleic acid (PNA) with DNA and 2′-deoxy 2′-fluoroarabinonucleic acid (2′-FANA). Chimeric oligomers possessing a single central PNA insert were capable of forming hybrid duplexes with complementary RNA, although with diminished thermal stability in comparison to the unmodified oligomers. We subsequently determined the ability of the DNA and 2′-FANA oligomers of mixed-base composition to elicit human RNase H1 degradation of complementary RNA that was either unstructured or as a hairpin. In the case of the more rigid FANA strand, a PNA insert led to a higher ability of the chimera to direct the degradation of both types of RNA targets. Generally, the enhancement observed was greater for a butanediol linker than for a more rigid PNA linker. Along with previous work, these studies suggest that the general flexibility associated with an acyclic insert (e.g., butyl vs PNA)sand not necessarily the presence of local structural imperfections in the heteroduplexsis beneficial for RNase H1 activity. As well, there are implications to the charge nature of non-nucleotide inserts (neutral vs negative) and their ability to maintain RNase H activity that may serve to direct further design considerations. Together, these studies support the notion that flexibility of antisense oligonucleotide (AON)/RNA hybrids is essential for high RNase H catalysis, in which an enzyme-induced altered trajectory of the bound AON/RNA substrate could facilitate optimal interaction with the catalytic site of RNase H.

Chart 1a

INTRODUCTION The ability of antisense oligonucleotides (AONs) to specifically bind to a single-stranded RNA via Watson-Crick complementary hybridization represents an attractive and increasingly important practical strategy for inhibition of gene expression. AONs are short, chemically modified oligonucleotide sequences designed to possess good exo- and endonuclease resistance, sufficient cell penetration, and good affinity for the target mRNA (1). An attractive mechanism for suppression of gene expression involves specific mRNA hydrolysis via activation of the ubiquitous intracellular enzyme RNase H (2). Although the origins of the substrate specificity of RNase H are not well-defined yet, structural analysis of modified AONs capable of eliciting this enzyme highlights a common feature: an O4′-endo sugar pucker that leads to the formation of a helical conformation intermediate between pure A- and B-forms (3, 4). The discovery that [3.3.0]bc-ANA satisfies the conformational requirement for RNase H recognition and cleavage, without activating this enzyme, revealed the importance of flexibility within the AON strand (3, 5). We therefore hypothesized that both preorganization and flexibility are two important criteria for the ability of an AON to elicit RNase H activity and thus for their therapeutic potential (6). This prompted us to examine the properties of known RNase H-active AONs such as DNA and 2′-FANA containing acyclic 2′,3′-secouridine (Chart 1C) or butanediol (D) linkers (6). 2′-FANA AONs incorporating a butyl linker showed RNase H induction with comparable efficiency to the native DNA strands, a result that is further supported by the present study. Unfortunately, a substantial decrease in the thermal stability imparted by the flexible, abasic linker, reduces its practical utility in AONs. * Authors to whom correspondence should [email protected], [email protected]. † McGill University. ‡ The University of Western Ontario.

be

sent:

a Top: Chemical structure of modified oligonucleotides. A, DNA; B, 2′-FANA; C, 2′,5′-secouridine residue; D, butanediol linker; E, peptide nucleic acid. Bottom: AON design.

Thus, we next turned our attention to a peptide nucleic acid (PNA) linker (Chart 1E) (7), which, due to the presence of an amide bond, is rendered comparably less flexible than the hydrocarbon linkers previously examined by our group (6), and conceivably more likely to engage in interactions with the opposing RNA residue of the formed heteroduplex. As a PNA insert also possesses an acyclic structure, and therefore some inherent flexibility relative to the native furanose framework, we surmised that its introduction in place of an aliphatic linker within the AON strand might also result in elevated RNase H activity. In the present work, we evaluate the affinity of DNA and 2′-FANA oligonucleotides containing one or three PNA inserts toward complementary DNA and RNA and compare their ability to elicit RNase H1 to native and butyl-modified AON oligonucleotides. This allowed us to assess the charge nature of non-nucleotide inserts (neutral vs negative) and their impact on RNase H activity.

10.1021/bc060300r CCC: $37.00 © 2007 American Chemical Society Published on Web 01/18/2007

422 Bioconjugate Chem., Vol. 18, No. 2, 2007 Scheme 1 a

a Reagents: (a) ClCH2CO2H, NaOH/H2O, Reflux (85%); (b) ClCH2CO2H, 4 °C, (51%, 2 Steps); (c) MMT-Cl, CH2Cl2, NEt(i-Pr)2, Overnight (50%, after Chromatography); (d) DCC, 24 h (65%); (e) 1M NaOH, 0 °C, (75%).

Scheme 2 a

a Reagents: (a) PPh3, NaN3, CBr4, Dry DMF, Overnight (80%); (b) PPh3, Pyridine, H2O, 3 h; (c) MMT-Cl, 4-DMAP Pyridine, Overnight (30% over Two Steps); (d) Cl-P(OCE)N(i-Pr)2, DIPEA, THF, 2.5 h (80%).

EXPERIMENTAL PROCEDURES Synthesis of Monomers. The N-(2-aminoethyl)glycine-based thymine PNA monomer was prepared according to literature precedent with minor modifications to the procedures (Scheme 1) (8-12). 2′-FANA phosphoramidites were synthesized according to established procedures (13) and DNA phosphoramidites purchased from ChemGenes Corp. (Ashland, MA) were used as received. The 5′-amino-dC derivative required for the PNA insertion was synthesized as described in Scheme 2). All standard 5′-O-dimethoxytrityl-3′-O-(2-cyanoethyl)-N,N-diisopropylphosphoramidites of the various 2′-deoxy- and ribonucleosides, 5′-monomethoxytritylamino-2′,5′-dideoxythymidine3′-phosphoramidite, and the phosphitylation reagent used for derivatization [β-cyanoethyl-(N,N-diisopropylamino) phosphorchloridite] were purchased from ChemGenes Corp. (Ashland, MA). Synthesis of Thymin-1-yl Acetic Acid (1). Thymin-1-yl acetic acid was prepared as previously described (8). Briefly, thymine (10.0 g, 79.0 mmol) was dissolved in 100 mL of water, by heating to reflux, which contained 9.0 g (160 mmol) of KOH. To this solution was slowly added chloroacetic acid (7.6 g, 80 mmol) dissolved in 25 mL of water. The pH was maintained at approximately 10 while the solution was refluxed for 2 h. Upon acidification with concd HCl, the acid (1) precipitated (12.4 g, 85%) and was subsequently carefully recrystallized from water. Mp 260-262 °C. EI MS: calcd 184.15, found 184 (M+, 40%), 140 (-CO2, 50%), 96 (100%).1H NMR (400 MHz, DMSOd6): 13.10 (br s, 1H), 11.33 (s, 1H), 7.49 (s, 1H), 4.35 (s, 2H), 1.74 (s, 3H). 13C NMR (100 MHz, DMSO-d6): 169.9, 164.6, 151.3, 142.2, 142.0, 108.6, 48.7. Methyl N-(2-Aminoethylglycinate)dihydrochloride (2). N-(2Aminoethyl)glycine was prepared according to the method of

Paˆtureau et al.

Heimer et al. (9). Neat ethylenediamine (500 mL, 7.5 mol) was rapidly stirred while being cooled in an ice bath (4 °C). Solid chloroacetic acid (75 g, 0.79 mol) was added in ten portions allowing for complete dissolution between each addition. The reaction mixture was then stirred overnight at room temperature. The unreacted ethylenediamine was removed by evaporation under vacuum and later recycled. The remaining paste was triturated with DMSO and then filtered, and the solid cake was washed with DMSO and then diethyl ether. The white solid was of sufficient purity to carry on to the next step without recrystallization (51 g, 55%). Recrystallization may be done in methanol to give white plates, mp 159-161 °C. 1H NMR (400 MHz, D2O): 3.03 (s, 2H), 2.79 (app t, J ) 6.5 Hz, 2H), 2.66 (app t, J ) 6.5 Hz, 2H). 13C NMR (100 MHz, D2O): 177.9, 51.3, 46.2, 38.3. Compound 2 was prepared by Fischer esterification of the parent acid (11). N-(2-Aminoethyl)glycine (15.0 g, 127 mmol) was suspended in 700 mL of methanol that was saturated with gaseous, dry HCl. The mixture was refluxed for 12 h, and the product was collected at room temperature. Subsequent crops were obtained by cooling the mother liquor to -20 °C and again by reducing the volume and cooling. Yield: 24 g, 92%. Mp 190-192 °C. 1H NMR (400 MHz, DMSO-d6): 8.58 (br s, 3H), 4.05 (s, 2H), 3.68 (s, 3H), 3.26 (m, 2H), 3.18 (m, 2H). 13C NMR (100 MHz, DMSO-d6): 166.9, 52.8, 46.7, 44.0, 35.0. Methyl N-(2-Monomethoxytritylaminoethylglycinate) (3) (10, 11). Methyl N-(2-aminoethylglycinate) dihydrochloride (1.0 g, 4.9 mmol) was stirred in 20 mL of CH2Cl2 containing 1 mL of Hu¨nig’s base while being cooled in an ice bath (4 °C). To this slurry was dropwise added a solution of 1.2 equiv of monomethoxytrityl chloride (1.8 g, 5.8 mmol) in CH2Cl2. The mixture was stirred for 12 h at room temperature, after which the reaction was quenched by the addition of 5 mL CH3OH, diluted with 100 mL of CH2Cl2 and was extracted (3 × 100 mL sat NaHCO3, 3 × 100 mL brine), dried (anhydrous sodium sulfate), and the solvent was evaporated to yield a sticky offwhite foam. TLC analysis (2:1 diethyl ether, hexanes with 1% NEt3) was done with silica gel plates that had been pretreated with eluent and indicated the presence of monomethoxytritanol and methyl monomethoxytrityl ether, as well as unreacted started material as likely impurities. The desired product was visualized by UV shadowing, staining with HCl vapor and ninhydrin. Rf ) 0.16. Isocratic silica gel column chromatography was done (60:30:1 diethyl ether/hexanes/triethylamine) to yield a viscous yellow oil (50%), as previously reported (11). 1H NMR (400 MHz, CDCl3): 7.47-6.79 (m, 14H), 3.77 (s, 3H), 3.71 (s, 3H), 3.48 (s, 2H), 2.73 (t, J ) 5.6 Hz, 2H), 2.27 (t, J ) 5.6 Hz, 2H), 1.89 (br s, 2H). Methyl N-(2-Monomethoxytritylaminoethyl)-N-(thymin1-yl) Glycinate (4). Compounds 1 and 3 were condensed to give the PNA monomer ester. Equimolar amounts of 1 (166 mg, 0.9 mmol), DCC, and HOBt were dissolved in dry DMF (3 mL) and stirred for 30 min. After the appearance of a heavy white precipitate (DCU), a solution of 3 (125 mg, 0.3 mmol) in 2 mL of DMF which contained a catalytic amount of 4-DMAP was added dropwise. The suspension was stirred for 1 h; then, 2 equiv Hu¨nig’s base was added, and stirring was continued for an additional 12 h, after which time TLC analysis (as above) indicated complete consumption of 3. The mixture was then filtered to remove the DCU and diluted with 100 mL of CH2Cl2 and extracted (3 × 100 mL sat NaHCO3, 2 × 100 mL brine). After drying (Na2SO4), removal of the solvent by evaporation gave a crude product that was purified, without chromatography, by precipitation of an ethereal solution with hexanes (65% yield). Two rotamers were observed: major (ma.) and minor (mi.). 1H NMR (400 MHz, DMSO-d6): 11.32 (br s, 1H), 7.41-6.84 (m, 15H), 4.84 (ma.) and 4.53 (mi.) (s, 2H),

RNase H Activation by Arabinose−PNA Chimeras

4.46 (mi.) and 4.09 (ma.) (s, 2H), 3.72 (s, 3H), 3.69 (mi.) and 3.58 (ma.) (s, 3H), 3.49 (ma.) and 3.37 (mi.) (m, 2H), 2.15 (ma.) and 2.08 (mi.) (m, 2H), 1.76 (s, 3H). N-(2-Monomethoxytritylaminoethyl)-N-(thymin-1-yl)-glycine (5) (12). Compound 4 was hydrolyzed to the free acid by treatment for 1 h with equal volumes of 2 M NaOH and methanol at 4 °C. The pH of the solution was then adjusted to 4-5 by dropwise addition of aqueous NaH2PO4 and the title compound isolated by repeated extraction into ethyl acetate. The organic fractions were pooled, dried over anhydrous sodium sulfate, and the solvent was removed by evaporation to yield a white solid. The crude product was purified by repeated trituration with diethyl ether, and the purified product was consistent with previous reports and isolated in 75% yield (12). Two rotamers were observed: major (ma.) and minor (mi.). 1H NMR (400 MHz, DMSO-d6): 11.31 (s, 1H), 6.8-7.4 (m, 15H), 4.81 (ma.) and 4.50 (mi.) (s, 2H), 4.30 (mi.) and 3.92 (ma.) (s, 2H), 3.71 (s, 3H), 3.33 (ma.) and 3.15 (mi.) (m, 2H), 2.15 (ma.) and 2.09 (mi.) (m, 2H), 1.74 (s, 3H). Alternatively, the triethylammonium salt was prepared by saponification with 2 M aqueous NEt3, 48 h at room temperature. The water was removed by evaporation under vacuum with gentle warming, and the excess triethylamine was removed by repeated coevaporation with absolute ethanol, ethyl acetate, and then dichloromethane. Synthesis of 5′-Azido-2′,5′-dideoxy-N4-benzoylcytidine (6) (14). N4-Benzoyl-2′-deoxycytidine (1.0 g, 3.02 mmol), PPh3 (80 mg, 3.08 mmol), NaN3 (33 mg, 5.13 mmol), and CBr4 (1.02 g, 3.08 mmol) were separately dried in a vacuum desiccator over P2O5 overnight. The following were then combined and dissolved in 15 mL of anhydrous DMF: N4-benzoyl-2′-deoxycytidine, PPh3, and NaN3. CBr4 was promptly added, and the reaction was stirred under a nitrogen atmosphere for 26 h at room temperature. The reaction was quenched by the addition of 3 mL of MeOH, and stirring was continued for 1 h. The solvent was then removed under vacuum, and the crude material was flash purified by silica gel chromatography using 5% methanol/dichloromethane to afford the desired product 1 as a white foam with a yield of 70% (0.75 g). TLC: Rf ) 0.43 in 9:1 (v/v) dichloromethane/methanol. FAB-MS (NBA matrix): calcd [M + Na+] 379.34, found 379. 1H NMR (400 MHz, DMSO-d6) of 5′-azido-2′,5′-dideoxy-N4-benzoylcytidine: 11.3 (s, 1H, N-H), 8.2 (d, 1H, J5,6 ) 7.7 Hz, H-6), 8.0 to 7.4 (m, 6H, H-5 and Bz), 6.2 (dd, 1H, J1′,2′ ) 6.7 Hz, J1′,2′′ ) 6.4 Hz, H-1), 5.5 (d, 1H, JOH,3′ ) 4.2 Hz, OH), 4.2 (m, 1H, H-3′), 4.0 (m, 1H, H-4′), 3.6 (m, 2H, H-5′ and H-5′′), 2.3 (m, 1H, H-2′), 2.2 (m, 1H, H-2′′). Synthesis of 5′-Amino-2′,5′-dideoxy-N4-benzoylcytidine (7) (15). Dry 5′-azido-2′,5′-dideoxy-N4-benzoylcytidine (6; 356 mg, 1.0 mmol) was dissolved in 5 mL of anhydrous pyridine, and to the stirred solution was added 0.3 g PPh3. After 3 h, 1 mL of water was added and stirring was continued for 2 h, at which time another 1 mL aliquot of water was added and the resulting precipitate was filtered, removing the PPh3 and Ph3Pd O. The water/pyridine solution was extracted three times with 5 mL ethyl acetate, and the aqueous solution lyophilized to afford crude 7. For characterization purposes, the N4-benzoyl group was removed as follows (16). A small amount of 7 (220 mg, 0.6 mmol) was dissolved in a mixture containing 2.5 mL of pyridine and 2.5 mL of concentrated NH4OH, and stirred overnight at 60 °C, after which the solution was cooled on ice and evaporated. The residue was partitioned between 15 mL of H2O and 1.5 mL of EtOAc. The aqueous layer was extracted, washed with Et2O (3 × 3 mL) and EtOAc/Et2O (1:2 v/v, 2 × 5 mL), and lyophilized to afford 5′-amino-2′,5′-dideoxycytidine in 98% yield (190 mg). FAB-MS (NBA-matrix): calcd [M + Na+] 249.22, found 249.0. 1H NMR (270 MHz, DMSO-d6) of

Bioconjugate Chem., Vol. 18, No. 2, 2007 423

5′-amino-2′,5′-dideoxycytidine: 7.7 (d, 1H, J5,6 ) 7.4 Hz, H-6), 7.15 (br s, 1H, NH2), 6.1 (t, 1H, J1′,2′ ) 7.17 Hz, J1′,2′′ ) 6.43 Hz, H-1′), 5.7(d, 1H, J5,6 ) 7.17 Hz, H-5), 5.2 (br s, 3H, NH2 and OH), 4.1 (m, 1H, H-3′), 3.7 (m, 1H, H-4′), 2.7 (m, 2H, H-5′ and H-5′′), 2.1 (m, 1H, H-2′), 2.0 (m, 1H, H-2′′). Synthesis of 5′-MMTNH-2′,5′-dideoxycytidine-N4-benzoylcytidine (8) (8). Compound 7 (226 mg, 1 mmol) was coevaporated twice with 20 mL dry pyridine and dissolved in 5 mL dry pyridine. MMT-Cl (460 mg, 1.5 mmol) was added and the reaction stirred overnight. When the reaction appeared complete by TLC (Rf ) 0.6 in 1:9 v/v MeOH/CH2Cl2), the mixture was diluted with 50 mL dichloromethane, washed with 50 mL brine, and the organic layer dried (sodium sulfate), filtered, and evaporated to obtain the crude product. The crude product was dissolved in a minimal amount of dichloromethane and purified by flash silica gel chromatography (eluant: dichloromethane, then 19:1 (v/v) dichloromethane/methanol with 3-5 drops of TEA to prevent removal of the trityl group due to acidity of the silica) to afford the desired product 8 as white foam with a yield of 30% (180 mg). FAB-MS (NBA matrix): calcd [M + Na+] 625.68, found 625.3. 1H NMR (400 MHz, DMSO-d6) of 5′-MMTNH-2′,5′-dideoxycytidine-N4-benzoylcytidine: 11.5 (s, 1H, N-H), 8.1 to 7.1 (m, 16H, H-6, H-5 and MMTr), 6.1 (t, 1H, J1′,2′ ) 6.4 Hz, J1′,2′′ ) 6 Hz, H-1′), 5.2 (d, 1H, JOH,3′ ) 4.8 Hz), 4.1 (m, 1H, H-3′), 3.9 (m, 1H, H-4′), 3.7 (s, 3H, CH3O-), 2.7 (t, 2H, JNH,5′ ) 8 Hz, JNH,5′′ ) 8 Hz), 2.4 (m, 1H, H-5′′), 2.2 (m, 2H, H-5′ and H-2′′), 2.0 (m, 1H, H-2′). Synthesis of 5′-MMTNH-2′,5′-dideoxycytidine-N4-benzoylcytidine phosphoramidite (9). Compound 8 (73.2 mg, 0.12 mmol) was dissolved in 2 mL of anhydrous THF and the flask purged with argon. While stirring, 0.076 mL (0.44 mmol) double-distilled N-ethyl-N,N-diisopropylamine (DIPEA) was added, followed by the slow addition of 0.03 mL (0.14 mmol) 2-cyanoethyl-N,N-diisopropylchlorophosphoramidite. The reaction mixture was stirred until complete consumption of starting material was observed (2.5 h). The reaction progression was monitored by silica plate TLC which employed 1:19 (v/v) methanol/dichloromethane as the mobile phase. TLC: Rf ) 0.4 and 0.47 (corresponding to two isomers). After evaporating the mixture to dryness, the crude material was dissolved in a minimum amount of dichloromethane. The chromatography column was preconditioned with 20:80 (v/v) dichloromethane/ hexane and 1% TEA, and the pure compound was eluted following a stepwise gradient of dichloromethane/hexane until a 1:1 ratio was reached. The pure fractions of the two isomers were combined, concentrated on a rotatory evaporator, and dried under high vacuum to obtain a stable foam following coevaporation with diethyl ether. FAB-MS (NBA matrix): [M + Na+] calcd 825.9, found 825. 31P NMR (270 MHz, acetone-d6): 153.9 and 153.8. Oligonucleotide Synthesis. DNA and 2′-FANA oligonucleotides containing PNA inserts were prepared as previously described (17-19), purified by PAGE or anion exchange HPLC, and characterized by MALDI-TOF mass spectrometry (Kratos MS25RFA instrument) (Table 1). Sequences of all the oligonucleotides studied are listed in Table 2. UV Thermal Denaturation Measurements. Thermal denaturation profiles were obtained on a Varian CARY 300 UVvis spectrophotometer. Prior to experiments, equimolar solutions of AON and target DNA or RNA oligonucleotides were mixed in a buffer consisting of 140 mM KCl, 1 mM MgCl2, and 5 mM Na2HPO4 (pH 7.2). The duplex concentration was close to 2 µM. Samples were heated at 95 °C for 10 min and then cooled slowly to room temperature and finally stored overnight at 4 °C to permit annealing. The solutions were gradually heated (0.5 °C/min) and the UV absorbance measured at 260 nm after

424 Bioconjugate Chem., Vol. 18, No. 2, 2007 Table 1

Paˆtureau et al.

a

entry

sequences 5′ f 3′ b

calcd mass (g‚mol-1)

obsd mass (g‚mol-1)

obsd molecular ion

I II III VI VII XII XV

ttt ttt ttt ttt ttt ttt ttt ttt ttt ttt tPt*t ttt ttt ttt ttP t*tt ttt ttt tta tat ttt Pt*c ttt ccc tta tat tPP Pt*c ttt ccc att ccg tca Pc*g ctc ctc ATT CCG TCA Pc*G CTC CTC

5413.6 5678.6 5374.4 5332.4 5255.9 5306.5 5610.5

5412.3 5708.3 5369.0 5350.4 5315.8 5307.2 5613.3

(M-H)(M-H + Li + Na)(M-H)(M-H + 2Li)(M-H + K + Na)(M-H)(M-H)-

a Matrix: 1 µL of saturated ATT/spermine solution, 1 µL fucose solution, and 1 µL autoclaved deionized water. b Bold uppercase, 2′-FANA nucleotides; uppercase letters, ribonucleotides; lowercase letters, deoxynucleotides; P, peptide nucleic acid; t*, 5′-amino-2′,5′-dideoxythymidine; c*, 5′-amino-2′,5′dideoxycytidine. For characterization of oligomers XIII and XVI, see ref 6.

Table 2. Melting Temperatures (Tm) of the AON/RNA Heteroduplexesa entry

sequences 5′ f 3′ b (composition)

I II III IVc

ttt ttt ttt ttt ttt ttt ttt ttt ttt ttt Pt*t ttt ttt ttt ttP t*tt ttt ttt ttt ttt ttt Btt ttt ttt

V VI VII VIII IX Xc

tta tat ttt ttc ttt ccc tta tat ttt Pt*c ttt ccc tta tat tPP Pt*c ttt ccc tta tat ttt ctc ttt ccc tta tat tcc ctc ttt ccc tta tat ttt Btc ttt ccc

XI XII XIIIc

att ccg tca tcg ctc ctc att ccg tca Pc*g ctc ctc att ccg tca Bcg ctc ctc

XIV XV XVIc

ATT CCG TCA TCG CTC CTC ATT CCG TCA Pc*G CTC CTC ATT CCG TCA BCG CTC CTC

Tm (°C) RNA target

Tm (°C) DNA target

∆Tm (°C) DNA target

∆Tm (°C) RNA target

Homopolymeric 45 35.5 35

36.5 27 31

-9.5 -10

-9.5 -5.5 -6

Mixed Base 47.5 34.5 15.0 36.5 n.d.

48.5 38 38.5 40 28.5

-13 -32.5 -11 n.d.

-10.5 -10 -8.5 -20 -5

Ha-ras n.d n.d 2′-FANA n.d n.d

72 27.5, 55 n.d.c

n.d.

-17 -12

81 71.5 n.d.c

n.d.

-9.5 -10

a Lowercase letters, deoxynucleotides; uppercase, ribonucleotides; bold uppercase, 2′-FANA nucleotides; underlined, mismatch nucleotide. Target RNA sequences correspond to rA18 or 5′-GGG AAA GAA AAA AUA UAA-3′ or for Ha-ras sequences, 5′-GAG GAG CGA UGA CGG AAU-3′; target DNA sequences correspond to dA18 or 5′-ggg aaa gaa aaa ata taa-3′; P, thymine peptide nucleic acid; t*, 5′-amino-2′,5′-dideoxythymidine; c*, 5′-amino-2′,5′dideoxycytidine; B, butanediol linker. b Aqueous solutions of ∼2 × 10-6 M of each oligonucleotide, 140 mM KCl, 1 mM MgCl2, 5 mM Na2HPO4 buffer (pH 7.2); uncertainty of (0.5 °C for Tm; n.d. - not determined. c Values taken from ref 6.

temperature stabilization. Tm values were calculated using the baseline method and have an uncertainty of (0.5 °C. RNase H assays. RNase H1 degradation assays were carried out to monitor the efficiencies of modified duplexes to elicit activity. The temperature at which the assays were run was usually chosen to be 10 °C lower than the determined thermal melting temperature of the least stable hybrid in the series. Whenever possible, assays were performed at 37 °C to simulate intracellular conditions. Generally, 150 pmol of antisense oligonucleotide were mixed with 50 pmol of 5′-32P-labeled RNA in buffer containing 60 mM Tris-HCl (pH 7.8), 60 mM KCl, 2.5 mM MgCl2, and 2 mM dithiothreitol (500 µL final volume). Samples were heated at 90 °C for 10 min and slowly cooled to room temperature to allow the duplexes to form. They were then transferred to a Plexiglass sample holder and left overnight in a freezer at -20 °C. 9 µL aliquots of the solutions were placed in sterile microtube vials and incubated at the assay temperature for at least 1 h prior to initiating enzymatic cleavage reactions. Determination of RelatiVe Enzyme Acceleration Profiles. Optimal conditions for enzymatic assays were evaluated from a series of preliminary RNase H dilution assays that were prepared as follows: 1 µL aliquot of purified stock enzyme solution was transferred to a microtube and diluted with varying amounts of double-distilled water. When assays were done at suboptimal temperatures (i.e., T < 37 °C), less dilution of the enzyme was necessary. Reactions were initiated in parallel at the specified temperature by addition of purified human RNase

H1 (1 µL) in sterile double-distilled water. The resulting mixture was gently mixed by pipeting action and centrifuged to ensure reaction homogeneity and consistency between vials. The enzyme was handled over ice to avoid any loss of activity. Trial runs were performed on two representative test substrates, usually the first and last points (e.g., 5 and 30 min) in the presence of three enzyme dilutions. At the end of the incubation time, samples were quenched with an equal volume of denaturing loading buffer (98% deionized formamide, 10 mM EDTA, 1 mg/mL bromophenol blue (BPB), and 1 mg/mL xylene cyanole (XC)) and vortexed to mix the contents and ensure complete deactivation of the enzyme. The samples were then heated at 90 °C for 5 min to denature the strands and were subsequently loaded onto a 16% polyacrylamide sequencing gel containing 7 M ultrapure urea. Cleavage patterns were analyzed by autoradiography as described below. Assays were done with the optimized enzyme concentration as determined above. Duplex substrate solutions were made of a 3-fold excess of AON relative to the target RNA strand. Assays were conducted in parallel for a given series. Reactions were initiated by addition of 5 µL of a freshly prepared solution of human RNase H1 in sterile water (diluted as explained above) to each test substrate (45 µL final volume: 4 time points + 1 extra ) 5 × 9 µL). 9 µL of the 0 min time point were prepared in a different vial. 10 µL aliquots were taken at five defined times (0, 5, 10, 20, and 30 min) and quenched with an equal volume of loading buffer, followed by heating at 90 °C for 5

Bioconjugate Chem., Vol. 18, No. 2, 2007 425

RNase H Activation by Arabinose−PNA Chimeras

min. For Ha-ras sequences, reactions were initiated by the addition of 25 µL of freshly prepared human RNase H1 in sterile water (diluted as explained above) to each test substrate (30 µL final volume). 9 µL of the control solution (time ) 0 min) were prepared in a different Eppendorf tube. 10 µL aliquots were taken at six defined times (0, 2, 5, 10, 20, and 30 min) and quenched with an equal volume of loading buffer, followed by heating at 90 °C for 5 min. At the completion of the experiment, the samples were heated and then quickly cooled on ice before electrophoretic analysis (16% denaturing PAGE, 1.5-2 h, 2000 V). Gels were prerun for 20 min at 1500 V before loading the samples. Degradation profiles were analyzed by autoradiography and further quantified by densitometric analysis after autoradiography. Quantification of the remaining full-length 32P-RNA signal was obtained by UN-Scan-It digitizing software (Silk Scientific, Inc.).

RESULTS AND DISCUSSION Duplexation Studies. For the homopolymeric sequences shown in Table 2, thermal melting experiments were performed at two wavelengths, 260 and 284 nm, corresponding to the absorbance of duplexes and triplexes, respectively. In the latter case, data obtained at the longer wavelength was primarily undertaken to rule out the existence of triple-helical species in solution (data not shown), which are favored for poly(thymine) PNA oligomers (7). Indeed, no transitions characteristic of triplehelical species were observed. Tm data indicative of duplex formation were thus obtained for the modified oligomers at 260 nm (I-III) and summarized in Table 2. In assessing the thermal profiles of unmodified and PNAcontaining AONs, we observed a significant drop in duplex stability for all modified oligonucleotides investigated. As expected, all oligomers (I, II, and III) showed greater affinity toward dA18 than rA18. Surprisingly, we found that the stability of duplexes formed with rA18 appear sensitive to the site of modification, whereas the complexes formed with dA18 were not. For instance, whether a PNA unit substitutes a nucleotide at position 13 or 9 (numbered from the 5′-end, entries II and III, Table 2), the destabilization is the same (∆Tm ) ∼-10 °C) when associating with a DNA complement. However, when hybridized with rA18 targets, AON III (substituted in the middle at position 9) is 4 °C more stable than AON II in which the point of substitution is closer to the 3′-end (position 13). We next turned our attention toward mixed sequences (A/ T/C) in order to do a comparative study of duplex stability of oligomers containing either a butanediol (sequence X, Table 2) or a PNA linker at position 10 (VI, Table 2). We were also interested to see if there was any beneficial additive effect upon the thermal stability of the duplex following the incorporation of three contiguous PNA residues (VII). In accord with the observation for homopolymeric sequence III, sequence VI showed greater destabilization with its complementary DNA target (∆Tm ) -13 °C) as compared to the hybrid formed with RNA (∆Tm ) -10.5 °C). Similar results have been observed by Bergman et al. (20). who introduced a single PNA building block in a mixed sequence DNA 13-mer. For the same RNA target, the PNA insertion (VI, ∆Tm ) -10.5 °C) was less tolerated than a butyl linker (X, ∆Tm ) -5 °C) (6). However, the destabilizing effect was not additive upon incorporation of three contiguous PNA residues (VII). In fact, the same decrease was observed for one or three insertions (∆Tm ≈ -10 °C). This is not surprising since PNA-DNA chimeras are known to form stable hybrid duplexes with complementary DNA or RNA (21). The stability is dependent upon the overall composition and placement of DNA versus PNA (22), such that higher thermal stability is to be expected upon increasing the number of

Table 3. Melting Temperatures (Tm) of (PNA)T-G Mismatch Experiments target stranda RNA (∆Tm, °C)

DNA (∆Tm, °C)

AON strandb

single mismatch

triple mismatch

single mismatch

VI (single PNA insert) VII (triple PNA insert)

39 (+1)c n.a.

n.a. 27 °C (-11.5 °C)

34 (-0.5) n.a.

a Target sequences correspond to RNA, in uppercase lettering, single mismatch underlined: 5′-GGG AAA GAG AAA AUA UAA-3′. RNA triple mismatch underlined: 5′- GGG AAA GAG GGA AUA UAA-3′. DNA, in lowercase, single mismatch underlined: 5′-ggg aaa gag aaa ata taa-3′. Aqueous solutions of ∼2 × 10-6 M of each oligonucleotide, 140 mM KCl, 1 mM MgCl2, 5 mM Na2HPO4 buffer (pH 7.2); uncertainty of (0.5 °C for Tm; n.a. - not applicable. b See Table 2. c ∆Tm: the difference between the melting temperature of the perfectly matched PNA-containing AON and the hybrids containing one or three mismatches.

consecutive PNA residues. It is also interesting to note that the selectivity for the RNA complement increased with the insertion of a PNA triad. In fact, AON VII binds to its DNA complement very weakly at best (Tm ca. 15 °C, broad transition). To gain further insight into the effect of replacing nucleotides with PNA units and the role of base pairing on the stability of the complexes, we undertook Tm studies with sequences containing one or three base mismatches either opposite or in place of PNA in the duplex (sequences VI-IX, Tables 2 and 3). First, introduction of a single mismatch in the DNA or RNA target strand opposite the PNA insert in sequence VI to yield a duplex with a (PNA)T-G mismatch gave a Tm which was the same as the perfectly matched sequence (Tables 2 and 3). Thus, there is no deleterious effect for a mismatch across from the single PNA insert, and both resulted in the same destabilization of the hybrid relative to the non-PNA-containing strands. This result implies that the nucleobase of the PNA residue does not contribute to binding in this sequence context, or its base pair is worth no more energetically than the wobble pair available in (PNA)T-G mismatched strand. A possible explanation for lack of base pairing may be that the rigidity of the PNA insert does not permit it to adopt a geometry that allows interstrand hydrogen bonding with the opposing nucleobase in the complementary strand. Interestingly, a single C-A mismatch (∆Tm ) -11 °C) in the mixed sequence VIII is less destabilizing than a putative matching PNA insert (VI, ∆Tm ) -13 °C), which supports the observation that the presence of an amide bond at the PNA-DNA junction is destabilizing (19, 23). Sequence VII, which contains three contiguous PNA inserts, demonstrates a large destabilization when hybridized to a complement RNA to give three (PNA)T-G mismatches. The magnitude of the depression of the Tm is similar to that observed for three (DNA)C-A mismatches (∆Tm ) -20 °C, Table 2), when hybridized to an RNA target. This change in Tm is much larger than when VII hybridizes to a matched RNA sequence (∆Tm) -11.5 °C, Table 3), thus indicating that the three PNA nucleobases partially participate to binding with the RNA target. This may be understood on the basis that the insertion of a PNA triad is sufficient to nucleate a structural perturbation that can accommodate nucleobase pairing, whereas a single residue insertion is not. This notion is supported by the similarity in CD spectra of the unmodified DNA/RNA heteroduplex (V and complementary RNA) to the spectra of VI and VII with the same complementary RNA. In addition, the global conformation of either the homopolymeric sequence or 2′-FANA sequence is largely unperturbed by insertion of a single PNA residue, as shown by the CD spectra of these complexes. Only the introduction of a triple PNA(T)-G mismatch showed a strong disruption of the overall morphology evidenced by a dramatic change in the CD spectra.

426 Bioconjugate Chem., Vol. 18, No. 2, 2007

Figure 1. UV thermal melting profiles of 2′-FANA and DNA backbones containing a PNA insert complexed with Ha-ras 18-mer RNA. Solid line, XII; crosses, XIV; dashed line, XI; open circles, XV.

Finally, we analyzed the impact of the presence of one PNA insert in a more rigid 2′-fluoroarabinonucleic acid (2′-FANA) backbone (XV), using the all-DNA sequence for comparison. As expected, 2′-FANA sequence XIV showed a greater Tm than the unmodified DNA XI (Table 2) that is rationalized on the basis of greater strand preorganization that entropically favors double helix formation (17, 24, 25). The greater rigidity of the 2′-FANA scaffold may also explain the smaller destabilization that is caused (∆Tm ) -9.5 °C) by a single PNA insert in the middle of the sequence (XV) as compared to the DNA analog XII (∆Tm ) -17 °C). Furthermore, in contrast with the 2′FANA counterpart, the melting curve of sequence XII presents two clearly distinct melting transitions (Figure 1). The reason for the biphasic nature of this transition could lie in the fact that 2′-deoxynucleotides are more flexible than 2′FANA residues and would be more likely to experience greater intrahelical disorder around the PNA domain. The PNA residue possibly creates distortions that prevent one-half of the strand from binding efficiently to the target. Further experiments are needed to investigate this hypothesis. Interestingly, while flanking deoxynucleotides in the AON tolerate PNA inserts poorly in comparison to a butyl linker (sequences XII and XIII, Table 2, ∆Tm ) -17 °C vs -12 °C), 2′-FANA sequences exhibit the same destabilization with either type of insert (XV and XVI, ∆Tm ) -9.5 and -10 °C). RNase H Assays. The ability of the modified oligonucleotides to elicit human RNase H1 activity when targeted to a structured RNA (e.g., hairpin) or unstructured RNA (random coil) was next evaluated. We first examined the activity of homopolymeric AON II and III in conjunction with DNA control (I) and the sequence containing a butyl insert in position 10 (IV). The presence of a PNA linker within the AON strand increases RNase H1 activity compared to control DNA but to a smaller extent than observed with IV, which may be attributed to the loss of flexibility imparted by the rigid amide bond. The noted order of RNA degradation within this series is IV > III > II > I (Figure 2). Interestingly, sequences II and IIIsalthough relatively less efficient at activating RNase H with respect to central butyl-modified AON IVsnonetheless confer the same accelerating trends to their relevant AON-RNA substrates as those described in previous work comparing central vs 3′terminal butyl insertions on RNase H activity (6). Next, we turned our attention toward mixed sequences with DNA (V) as the control and the butyl-containing AON X for comparison. As PNA inserts induce some flexibility in the AON strand, and given the previous result with the poly(T) sequences, we expected VI to have greater RNase H activity than the DNA

Paˆtureau et al.

Figure 2. Time course of RNase H1-mediated cleavage of hybrids containing polypyrimidine sequences I-IV (see Table 1).

Figure 3. Time course of RNase H1-mediated cleavage of hybrids containing mixed sequences (Table 1).

control. However, hybrids of VI with RNA were found to be much less reactive as substrates for RNase H1 than control DNA/RNA. This effect is emphasized in comparison to the control DNA, because this sequence context is inherently a better substrate for RNase H1. Thus, for this sequence, the activity of the control DNA is enhanced by incorporation of a flexible butyl linker and diminished by incorporation of one or three PNA residues. A possible explanation may be that the PNA insert creates a greater local distortion within this duplex, as compared to the butyl linker, and becomes a less acceptable substrate for the enzyme. This idea is supported by the observation that complexes of VI with complementary RNA have a lower Tm than their butyl counterpart X. However, the thermal destabilizations associated with these two AON sequences likely have different thermodynamic origins, with diminished RNA binding likely owing to the absence of a base pairing partner in the case of butyl inserts, and to an incompatible local conformation in the case of a PNA insert. This effect may also be sequencedependent, since it was not observed in other studies. Oligonucleotide VII is the least reactive of all AONs tested, which is not surprising if we consider that this oligonucleotide tends to adopt the profile of a PNA-DNA chimera, with the latter being less reactive than pure DNA, and first synthesized with the goal of introducing an RNase H-active backbone to the nonreactive PNA domain (22). The RNase H1 efficiency observed for each AON is the following: X > V > VI > VII; Figure 3. Finally, we used an in vitro Ha-ras model system to evaluate the ability of the DNA-PNA chimera to elicit RNase H1 activity. This RNA species and its associated structure is of particular interest, since hyperactivated forms of point mutant

RNase H Activation by Arabinose−PNA Chimeras

Bioconjugate Chem., Vol. 18, No. 2, 2007 427

Figure 4. Secondary structures presumably adopted by (A) 18-mer and (B) 40-mer Ha-ras RNA. AON binding region is in blue.

Figure 5. Quantification of the remaining full-length 32P-RNA signal at two time points (20 and 30 min) for the Ha-ras RNA targets. For convenience, AONs are identified on the basis of the single inserts wherein B ) butyl and P ) peptide nucleic acid.

and nonmutated mammalian ras protooncogenes are responsible for various human cancers (26-29), and hence represent a potential therapeutic target. Furthermore, we wished to compare the rate and extent of reactivity of both structured and unstructured RNA targets. As is common in the antisense approach, we believe that RNA cleavage should also depend on the structure, length, and base composition of the target RNA. For this purpose, two oligomers were investigated: an 18-mer and a 40-mer, both of which are complementary to an internal nucleotide region encompassing the translation initiation site of the ras gene and corresponding, respectively, to residues -8 to +10 and -19 to +21 of the full-length transcript. The apparent secondary structures of both targets have previously been determined (6) using the MFOLD program (30, 31) and supported by UV thermal profiles. No intramolecular association was observed for the 18-mer target, whereas the 40-mer was capable of self-associating into a stable 25 nucleotide hairpin structure (6) (Figure 4). The ability of PNA-containing Ha-ras sequences (XII) to activate human RNase H1 was evaluated in conjunction with butyl-containing Ha-ras-DNA (XIII) and 2′-FANA sequences (XV, XVI), as well as DNA (XI) and 2′-FANA (IX) control

AONs (Figure 5). The linker-containing DNA AON (XIII) did not enhance target degradation, which is consistent with the work of other groups (32, 33). Obviously, target accessibility is an important determinant in the ability of the AONs studied to elicit RNase-H-mediated cleavage of the RNA. For the 18-mer unstructured RNA target, complete disappearance was observed after 20 min when in the presence of complementary DNA or the butyl linker containing 2′-FANA (XVI), whereas the structured 40-mer target was ∼30% intact when treated with the most active oligomers XVI or XV (Figure 6). In fact, all the FANA AONs were less effective at mediating RNase H1 directed destruction of the structured RNA target as compared to the same oligomers tested against the 18-mer unstructured target. The best modified AON for the 18-mer target was the 2′-FANA possessing a butyl linker (XVI), whereas the PNA-containing oligomer was more efficient at eliciting the cleavage of the 40-mer target. For 2′-FANA-based AONs, the order of enzyme efficiency is roughly the same for both the 18- and 40-mer targets: XVI ≈ XV > XIV. This is similar to the trend we observed earlier (IV > III, II > I) and likely correlates with a gradual decrease in the flexibility of the AON strand. In other words, the all-

428 Bioconjugate Chem., Vol. 18, No. 2, 2007

Paˆtureau et al.

Figure 6. Degradation patterns of Ha-ras-RNA (A) 18-mer and (B) 40-mer hybridized with various AONs. See experimental procedures for details. T ) 37 °C; time points: 0, 2, 5, 10, 20, and 30 min. The slowly migrating band observed for XIV, XVI, and XV hybridized to the 18-nt and 40-nt RNA targets corresponds to the duplexed species that does not dissociate despite heating of the sample prior to loading and the denaturing conditions of the gel electrophoresis, due to high Tm.

FANA strand XIV likely does not present enough flexibility to the enzyme, which would therefore require greater energy to adjust the hybrid to adopt the correct conformation for hydrolysis. Polyacrylamide electropherograms can be inspected to qualitatively assess the effect of a non-nucleotide insert into DNA or 2′-FANA on the pattern of RNase-H-directed RNA hydrolysis. Figure 6 shows the gel from which the data presented in Figure 5 was obtained. For both the 18-mer and 40-mer targets, the DNA oligomer XI causes a predominance of short hydrolysis products, likely a result of multiple cleavage events. For the modified oligomers, there is a specific hydrolysis more central in the targeted RNA resulting in a slower moving band. This is clearly seen in the linker-modified DNA oligomers regardless of the identity of the linker, that is, both the butyl- and PNAlinked DNA oligomers produce similar cleavage patterns.

Interestingly, the nature of the linker in a 2′-FANA sequence appears to affect the pattern of cleavage. The 2′-FANA oligomer (XIV) directs a central cleavage, assumed to be a primary hydrolysis event, as does the butyl linker-modified 2′-FANA (XVI). In contrast, the PNA-modified 2′-FANA (XV) presents a different cleavage pattern, with a product resulting from central hydrolysis being absent. We interpret this observation as being a result of the structural perturbation that a single PNA insert causes as well as the relative stiffness of the amide backbone, as compared to a butyl insert, which makes it a less suitable enzyme substrate.

CONCLUSIONS Oligonucleotide analogs capable of recruiting RNase H for the destruction of mRNA have potential use as artificial regulators of gene expression. Our previous work has shown

RNase H Activation by Arabinose−PNA Chimeras

that the ability of 2′-FANA to direct the RNase-H-mediated hydrolysis of RNA is dramatically improved by incorporation of a flexible non-nucleotide butyl linker. This modification has the deleterious effect of reducing the thermal stability of the complexes formed with complementary nucleic acids. The current study has investigated the incorporation of a neutral peptide nucleic acid linker, which we have found diminishes the thermal stability of complexes formed with complementary oligonucleotides. Although this is presumably due to some degree of structural changes at the PNA-DNA or PNA-FANA junctions in the chimeras, this was not found by comparative examination of the complexes by CD spectroscopy (data not shown). In general, PNA-modified oligonucleotides did not lead to a substantial improvement in induction of RNase H1 activity as compared to butyl derivatized oligonucleotides. However, homopolymeric DNA sequences containing a PNA insert were more efficient at eliciting RNase H1 activity relative to the unmodified homopolymeric DNA presumably due to added flexibility. Nevertheless, the greater rigidity imparted to the oligonucleotide strand by the PNA unit as compared to the butyl linker is believed to be accountable, at least in part, for the decrease in enzymatic activity. Another effect that cannot be ruled out from the present study is a partial loss of key enzyme interactions due to the uncharged nature of the amide backbone at the insertion junction. Nonetheless, it is clear that PNAmodified substrates can still support elevated cleavage over unmodified (i.e., all-DNA or all-FANA) strands in certain sequence contexts, especially given the general inability of PNA to activate RNase H. These trends thus corroborate our previous claims on substrate flexibility as a crucial element for high RNase H activity, and also illustrate the versatility with which a non-nucleotide residue, when strategically placed in the antisense oligonucleotide, can still maintain appreciable RNase H activation. Conversely, AON sequences containing three PNA inserts were less efficient at recruiting RNase H activity due to the increasing degree of PNA character to the oligomer and associated topological consequences on overall helix architecture. Together, these and other studies reinforce the hypothesis that local flexible sites within AON/RNA hybrids are essential for high RNase H catalysis, (6, 34-36) in which an enzyme-induced altered trajectory of the bound substrate (37, 38) could facilitate optimal interaction with the catalytic site of RNase H1.

ACKNOWLEDGMENT The authors (M.J.D., R.H.E.H.) gratefully acknowledge the Natural Science and Engineering Research Council of Canada and the Canadian Institutes of Health Research (M.J.D.) for funding. This work was also supported by a scholarship award from the Chemical Biology Strategic Training Initiative of the Canadian Institutes of Health Research to B.P. We thank Dr. M. M. Mangos for several helpful discussions, encouragement, and assistance with the writing of this manuscript.

LITERATURE CITED (1) Uhlmann, E., and Peyman, A. (1990) Antisense oligonucleotides: a new therapeutic principle. Chem. ReV. 90, 543-584. (2) Walder, R. Y., and Walder, J. (1988) Role of RNase H in hybridarrested translation by antisense oligonucleotides. Proc. Natl. Acad. Sci. U.S.A. 85, 5011-5015. (3) Minasov, G., Teplova, M., Nielsen, P., Wengel, J., and Egli, M. (2000) Structural basis of cleavage by RNase H of hybrids of arabinonucleic acids and RNA. Biochemistry 39, 3525-3532. (4) Zamaratski, E., Pradeepkumar, P. I., and Chattopadhyaya, J. (2001) A critical survey of the structure-function of the antisense oligo/

Bioconjugate Chem., Vol. 18, No. 2, 2007 429 RNA heteroduplex as substrate for RNase H. J. Biochem. Biophys. Methods 48, 189-208. (5) Christensen, N. K., Peterssen, M., Nielsen, P., Jacobsen, J. P., Olsen, C. E., and Wengel, J. (1998) A novel class of oligonucleotide analogs containing 2′-O,3′-C-linked [3.2.0]bicycloarabinonucleoside monomers: synthesis, thermal affinity studies, and molecular modeling. J. Am. Chem. Soc. 120, 5458-5463. (6) Mangos, M. M., Min, K.-L., Viazovkina, E., Galarneau, A., Elzagheid, M. I., Parniak, M. A., and Damha, M. J. (2003) Efficient RNase H-directed cleavage of RNA promoted by antisense DNA or 2′F-ANA constructs containing acyclic nucleotide inserts. J. Am. Chem. Soc. 125, 654-661. (7) Nielsen, P. E., Egholm, M., Berg, R. H., and Buchardt, O. (1991) Sequence-selective recognition of DNA by strand displacement with a thymine-substituted polyamide. Science 254, 1497-1500. (8) Rabinowtiz, J. L., and Gurin, S. (1953) Some pyrimidine derivatives. J. Am. Chem. Soc. 75, 5758-5759. (9) Heimer, E. P., Gallo-Torres, H. E., Felix, A. M., Ahmad, M., Lambros, T. J., Scheidl, F., and Meienhofer, J. (1984) Synthesis of analogs and oligomers of N-(2-aminoethyl)glycine and their gastrointestinal absorption in the rate. Int. J. Pept. Res. 23, 203-211. (10) Will, D. W., Breipohl, G., Langner, D., Knolle, J., and Uhlmann, E. (1995) The synthesis of polyamide nucleic acids using a novel monomethoxytrityl protecting-group strategy. Tetrahedron 51, 1206912082. (11) Stetsenko, D. A., Lubyako, E. N., Potapov, V. K., Azhikina, T. L., and Sverdlov, E. D. (1996) New approach to the solid phase synthesis of polyamide nucleic acid analogues (PNA) and PNADNA conjugates. Tetrahedron Lett. 37, 3571-3576. (12) Finn, P. J., Gibson, N. J., Fallon, R., Hamilton, A., and Brown, T. (1996) Synthesis and properties of DNA-PNA chimeric oligomers. Nucleic Acids Res. 24, 3357-3363. (13) Viazovkina, E. V., Mangos, M. M., Elzagheid, M. I., and Damha, M. J. (2002). Synthesis of 2′-fluoroarabino nucleoside phosphoramidites and their use in the synthesis of 2′F-ANA. In Current Protocols in Nucleic Acid Chemistry, Chapter 4.15, W. J. Sons, New York. (14) Yamamoto, I., Sekine, M., and Hata, T. (1980) One-step synthesis of 5′-azido nucleosides. J. Chem. Soc., Perkin Trans. 1 1, 306310. (15) Mag, M., and Engels, J. W. (1989) Synthesis and selective cleavage of oligodeoxyribonucleotides containing non-chiral internucleotide phosphoramidate linkages. Nucleic Acids Res. 17, 59731588. (16) Wolfe, J. L., Kawate, T., Belenky, A., and Stanton, V., Jr. (2002) Synthesis and polymerase incorporation of 5′-amino-2′,5′-dideoxy5′-N-triphosphate nucleotides. Nucleic Acids Res. 30 (17), 37393747. (17) Damha, M. J., Wilds, C. J., Noronha, A., Brukner, I., Borkow, G., Arion, D., and Parniak, M. A. (1998) Hybrids of RNA and arabinonucleic acids (ANA and 2′F-ANA) are substrates of ribonuclease H. J. Am. Chem. Soc. 120, 12976-12977. (18) Wilds, C. J., and Damha, M. J. (2000) 2′-Deoxy-2′-fluoro-β-Darabinonucleosides and oligonucleotides (2′F-ANA): synthesis and physicochemical studies. Nucleic Acids Res. 28, 3625-3635. (19) van der Laan, A. C., Brill, R., Kuimelis, R. G., Kuyl-Yeheskiely, E., van Boom, J. H., Andrus, A., and Vinayak, R. (1997) A convenient automated solid-phase synthesis of PNA-(5′)-DNA-(3′)PNA chimera. Tetrahedron Lett. 38, 2249-2252. (20) Bergmann, F., Bannwarth, W., and Tam, S. (1995) Solid phase synthesis of directly linked PNA-DNA hybrids. Tetrahedron Lett. 36, 6823-6826. (21) Wenninger, D., and Seliger, H. (1997) Synthesis and hybridization properties of modified oligonucleotides with PNA-DNA dimmer blocks. Nucleosides Nucleotides 16, 977-980. (22) Uhlmann, E., Peyman, A., Breipohl, G., and Will, D. W. (1998) PNA: Synthetic polyamide nucleic acids with unusual binding properties. Angew. Chem., Int. Ed. 37, 2796-2823. (23) van der Laan, A. C., Havenaar, P., Oosting, R. S., KuylYeheskiely, E., Uhlmann, E., and van Boom, J. H. (1998) Optimization of the binding properties of PNA-(5′)-DNA chimerae. Bioorg. Med. Chem. Lett. 8, 663-668. (24) Berger, I., Tereshko, V., Ikeda, H., Marquez, V. E., and Egli, M. (1998) Crystal structures of B-DNA with incorporated 2′-deoxy-2′-

430 Bioconjugate Chem., Vol. 18, No. 2, 2007 fluoro-arabino-furanosyl thymines: implications of conformational preorganization for duplex stability. Nucleic Acids Res. 26, 24732480. (25) Denisov, A. Y., Noronha, A. M., Wilds, C. J., Trempe, J. F., Pon, R. T., Gehring, K., and Damha, M. J. (2001) Solution structure of an arabinonucleic acid (ANA)/RNA duplex in a chimeric hairpin: comparison with 2′-fluoro-ANA/RNA and DNA/RNA hybrids. Nucleic Acids Res. 29, 4284-4293. (26) Chakraborty, A. K., Cichutek, K., and Duesberg, P. H. (1991) Transforming function of proto-ras genes depends on heterologous promoters and is enhanced by specific point mutations. Proc. Natl. Acad. Sci. U.S.A. 88, 2217-2221. (27) Kiaris, H., and Spandidos, D. A. (1995) Mutations of Ras genes in human tumors. Int. J. Oncol. 7, 413-421. (28) Kiaris, H., and Spandidos, D. A. (1995) Analysis of H-Ras, K-Ras and N-Ras genes for expression, mutation and amplification in laryngeal tumors. Int. J. Oncol. 7, 75-80. (29) Hua, V. Y., Wang, W. K., and Duesberg, P. H. (1997) Dominant transformation by mutated human ras genes in vitro requires more than 100 times higher expression than is observed in cancers. Proc. Natl. Acad. Sci. U.S.A. 94, 9614-9619. (30) Zuker, M., Mathews, D. H., and Turner, D. H., Eds. (1999) Biochemistry and Biotechnology, Kluwer Academic Publishers, Norwell, MA. (31) Mathews, D. H., Sabina, J., Zuker, M., and Turner, D. H. (1999) Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure. J. Mol. Biol. 288, 911-940. (32) Vorobjev, P. E., Pyshnaya, I. A., Pyshnyi, D. V., Venyaminova, A. G., Ivanova, E. M., Zarytova, V. F., Bonora, G. M., Scalfi-Happ, C., and Seliger, H. (2001) Nuclease resistance and RNase H

Paˆtureau et al. sensitivity of oligonucleotides bridged by oligomethylenediol and oligoethylene glycol linkers. Antisense Nucleic Acid Drug DeV. 11, 77-85. (33) Vorobjev, P. E., Pyshnaya, I. A., Pyshnyi, D. V., Repkova, M. N., Venyaminova, A. G., Zenkova, M. A., Ivanova, E. M., ScalfiHapp, C., Seliger, H., Bonora, G. M., and Zarytova, V. F. (2000) Cleavage of RNA in hybrid duplexes by the E. coli ribonuclease H: II. Substrate properties of oligonucleotides containing nonnucleotide linkers. Bioorg. Khim. 26, 758-764. (34) Mangos, M. M., and Damha, M. J. (2002) Flexible and frozen sugar-modified nucleic acids - modulation of biological activity through furanose ring dynamics in the antisense strand. Curr. Top. Med. Chem. (Sharjah, United Arab Emirates) 2, 1145-1169. (35) Lok, C.-N., Viazovkina, E., Min, K.-L., Nagy, E., Wilds, C. J., Damha, M. J., and Parniak, M. A. (2002) Potent gene-specific inhibitory properties of mixed-backbone antisense oligonucleotides comprised of 2′-deoxy-2′-fluoro-D-arabinose and 2′-deoxy nucleotides. Biochemistry 41, 3457-3467. (36) Damha, M. J., Noronha, A. M., Wilds, C. J., Trempe, J.-F., Denisov, A., and Gehring, K. (2001) Properties of arabinonucleic acids (ANA & 2′F-ANA): Implications for the design of antisense therapeutics that invoke RNase H cleavage of RNA. Nucleosides, Nucleotides Nucleic Acids 20, 429-440. (37) Premraj, B. J., Raja, S., and Yathindra, N. (2002) Structural basis for the unusual properties of 2′,5′ nucleic acids and their complexes with RNA and DNA. Biophys. Chem. 95, 253-272. (38) Nielsen, J. T., Stein, P. C., and Petersen, M. (2003). NMR structure of an R-L-LNA:RNA hybrid: structural implications for RNase H recognition, Nucleic Acids Res. 31, 5858-5867. BC060300R