5416-5424 Nucleic Acids Research, 1994, Vol. 22, No. 24 © 1994 Oxford University Press

Oligonucleotides with novel, cationic backbonesubstituents: aminoethylphosphonates

Reza Fathi*, Qing Huang, George Coppola, William Delaney, Rebecca Teasdale1,Arthur M.Krieg1 and Alan F.CookPharmaGenics, Inc., 4 Pearl Court, Allendale, NJ 07401 and 1 University of Iowa Hospitals andClinics, Division of Rheumatology, Department of Internal Medicine and Veterans AdministrationMedical Center, Iowa City, IA 52242, USA

Received July 20, 1994; Revised and Accepted October 14, 1994

ABSTRACT

Ollgonucleotide (2-amlnoethyl)phosphonates In whichthe backbone consisted of isomerically pure, alter-nating (2-aminoethyl)-phosphonate andphosphodlester linkages have been prepared andcharacterized. One of these single isomeroligonucleotides (Rp) formed a more stable duplexwith ONA or RNA than its corresponding naturalcounterpart. Hybrid stability was more pH-dependent,but less salt-dependent than a natural duplex. Thespecificity of hybridization was examined by hybridiza-tion of an ollgonucleotide containing one(2-aminoethyl)phosphonate to oligonucleotidespossessing mismatches in the region opposite to theamlnoethyl group. In contrast to oligonucleotides con-taining (amlnomethyl)-phosphonate linkages,ollgonucleotide (2-aminoethyl)phosphonates werecompletely stable to hydrolysis in aqueous solution.These oligonucleotides were resistant to nuclease ac-tivity but did not induce RNase H mediated cleavageof a complementary RNA strand. Incubation In aserum-contalnlng medium resulted In minimaldegradation over 24 hours. Studies of cell uptake byflow cytometry and confocal microscopydemonstrated temperature dependent uptake and In-tracellular localization. (2-Amlnoethyl)phosphonatesrepresent a novel approach to the Introduction ofpositive charges Into the backbone ofoligonucleotides.

INTRODUCTION

For oligonucleotides to be effective as antisense agents, theyneed to penetrate cell membranes and be resistant to degrada-tion by nucleases. These requirements have stimulated effortsto prepare backbone-modified derivatives which might becapable of penetrating membranes more readily whilst retain-ing their resistance to degradation. One widely used approachto improve cell penetration involves removal of the negativecharges to produce neutral backbones such as. for example.

methyl phosphonates [1], phosphoramidates [2] or peptidenucleic acids [3]. Methylphosphonates have been shown to betransported into cells by endocytosis [1] and were more resis-tant to degradation by nucleases [4].

An alternate approach to backbone modification is to replacethe anionic phophodiester groups with cationic groups. The in-troduction of a single cationic group could be advantageous inenhancement of binding to a specific site of a target protein orreceptor, whereas the introduction of cationic groups at variouspositions throughout the oligonucleotide provides a method forproducing molecules with widely differing charge densities. Ca-tionic molecules have previously been used to enhance thedelivery of a variety of conjugates into cells. The iron-transportprotein transferrin, for example, has been conjugated to cationicmolecules such as protamine or poly-lysine in order to transportnucleic acids [5], and cationic lipids have been used in aliposome-mediated transfection protocol for the introduction ofDNA into the nuclei of animal cells. The delivery of proteinsand peptides such as albumin [6] and enkephalin [7] across theblood-brain barrier has been enhanced by conjugation to cationicmolecules such as hexamethylenediamine, and a variety ofdrugs, including methotrexate and daunomycin, have been con-jugated to poly-lysine in attempts to improve their uptake andefficacy [8].

A few cationic molecules have been attached to oligo-nucleotides in attempts to improve their cellular uptake and/oractivity in vivo. The cationic protein avidin enhanced the cellularuptake of a biotin-oligonucleotide conjugate, whereas the anionicprotein streptavidin did not facilitate uptake of the sameoligonucleotide [9]. Conjugation of poly-lysine tooligonucleotides directed against VSV [10] and HTV [11] se-quences has been shown to produce enhanced effects as com-pared with their unmodified counterparts. Other methods whichhave been used for the introduction of positive charges intooligonucleotides include the synthesis of 2'-amino [12] or2'-alkylamino substituents [13] and the attachment of amino-containing groups to the purine or pyrimidine bases [14]. Ca-tionic substituents have been attached to the phosphorus atomvia phosphoramidate linkages [15], although the biological pro-perties of these compounds have not been reported.

*To whom correspondence should be addressed

by guest on May 5, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research, 1994, Vol. 22, No. 24 5417

Cationic groups can be introduced into oligonucleotides viaphosphonate derivatives in which the anionic oxygen is replac-ed by a cationic group via a phosphorus-carbon bond. Aprevious paper in this series [16] described die synthesis ofoligonucleotide (aminomethyl)phosphonates, the simplest ex-ample in this series. During the course of this prior work, itbecame evident that although (aminomethyl)phosphonatespossessed interesting properties, they were shown to behydrolytically labile in aqueous solution. The present paperdescribes the synthesis and properties of a more stable represen-tative of this class of compounds, (2-aminoethyl)-phosphonates.

EXPERIMENTAL PROCEDURESMaterials and methodsSynthesis and purification of oligonucleotides was carried outas previously described [16]. Extinction coefficients of oligo-nucleotides were obtained using a published procedure [17].NMR spectra were recorded using a Varian Unity Plus 500 MHzspectrometer. Unless otherwise stated, NMR spectra were runin DMSO-d6. 31P NMR were broad band decoupled, referenc-ed to H3PO4 as an external standard, and 'H NMR spectrawere referenced to external tetramethylsilane as standard.

Pyridinium [2-(3,4,5,6-tetrabromophthalimido)ethyl]phosphonate(2-Aminoethyl)phosphonic acid (2g, 16 mmol) was dissolvedin water (20 mL), neutralized with IN NaOH to pH 7.0 andpassed through a column of Dowex 50X2-400 (40 mL,pyridinium form) ion exchange resin. The column was washedwith water (3 X50 mL) and the combined eluates were concen-trated in vacuo to give 2.33 g (52%) of die pyridinium salt of(2-aminoethyl)phosphonic acid as a white solid. A portion ofthis material (lg, 3.53 mmol) was suspended in DMF (100 mL)and treated with 3,4,5,6-tetrabromo-phthalic anhydride (9.8 g,21.2 mmol) under reflux overnight. The reaction mixture wascooled to room temperature to precipitate the product. Theprecipitate was filtered, washed with DMF (3 X50 mL) followedby hexane (2 x 50 mL) and dried in vacuo over P2O5 to givethe title compound (2.57 g, 70%) as an off-white solid. 'HNMR: 8 (ppm) 8.5, 7.7, 7.3 (m, 12H, pyridinium), 3.7 (m,2H, CH2), 1.8 (m, 2H, CH2).

3IP NMR: 8 (ppm) 21.6 (s).The tetrachloro derivative was prepared using a similar pro-

cedure in which pyridinium (2-aminoethyl)phosphonate (lg,3.53 mmol) was suspended in DMF (100 mL) and treated with3,4,5,6-tetrachlorophthalic anhydride (6.05 g, 21.2 mmol).After isolation 1.94 g (89%) of pyridinium[2-(3,4,5,6-tetrachloro-phthalimido)-ethyl] phosphonate was ob-tained. 'H NMR: <5 (ppm) 78.5, 7.7, 7.3 (m, 12H,pyridinium), 3.7 (m, 2H, CH2), 1.8 (m, 2H, CHj). 31P NMR:5 (ppm) 21.4 (s).

Synthesis of 5'-0-(Dimethoxytrityl)thymidine-3'-[2-(3,4,5,6-tetrabromophthalimido)ethyl]phosphonate (1)

5'-O-(DimedK)xytrityl)thyniidine (0.822 g , 1.5 mmol) was driedby coevaporation with pyridine (2x5 mL), dissolved in drypyridine (40 mL), and treated with triisopropylbenzenesulfonylchloride (TPS-C1, 1.45g, 4.8 mmol) followed by pyridinium[2-(3,4,5,6-tetrabromo-phthalimido)ethyl]phosphonate (1.0 g,1.814 mmol). The resulting suspension was stirred overnightunder a dry nitrogen atmosphere, diluted with water (150 mL),

SERES* RpBOUERSERIES b: SplSOMER

DMTr

R o

D M T W > - L A J

OCHJCHICN

5: R2 = DlfTr10: R, =. H 6: R = Br

7: R = Cl

Figure 1. Synthesis of dinucleotide (2-aminoethyl)phosphonates.

and extracted with ethyl acetate (2x75 mL). The combinedorganic layers were washed with water (100 mL), dried overanhydrous sodium sulfate, filtered, and the filtrate wasevaporated to dry ness to give crude 1. This material was dissolv-ed in CH2C12/CH3OH (99:1, 10 mL) and filtered through a onemicron glass filter. The filtrate was purified on a Dynamax nor-mal phase silica column (8 /i, 21.4 mmx25 cm) with a gra-dient of 1-20% CH3OH in CH2C12 containing 0.1%triethylamine (TEA). The fractions eluting at 26-34 minuteswere collected and concentrated to yield 0.32 g (19%) of pure1 as an off-white foam. 'H NMR: 5 (ppm)11.33 (s, 1H, NH),7.49 (s, 1H, Hg), 6.8-7.38 (m, 13H, trityl), 6.19 (t, 1H, H r ) ,4.77 (m, 1H, H3.), 4.09 (s, 1H, Hv), 3.69 (s, 6H, OMe), 3.62(m, 2H, CH2), 3.2-3.6 (m, 2H, H3.5.), 2.32 (m, 2H, H2.2.),1.68 (m, 2H, CH2), 1.4 (s, 3H, 5-CH3).

31P NMR: 5 (ppm)17.35 (s).

The corresponding tetrachloro derivative 3 was prepared asan off-white foam by a similar procedure (63% yield) 'HNMR: 5 (ppm) 11.32 (s, 1H, NH), 7.47 (s,lH, He),6.84-7.35 (m, 13 H, trityl), 6.12 (t, 1H, H r ) , 4.75 (m, 1H,H r ) , 4.04 (m, 1H, H4.), 3.68 (s, 6H, OMe), 3.64 (m, 2H,H2), 3.16-3.23 (m, 2H, Hy), 2.31 (m, 2H, H2.), 1.70 (m,2H, H2), 1.38 (s, 3H, CH3). 31P NMR: 8 (ppm)16.4 (s).

O-[5'-O-(Dimethoxytrityl)thymidin-3'-yl]-O-(thy-midin-5'-yD-[2-(3,4,5,6-tetrabromophthalimido)ethyl]phos-phonate (2)Crude 1 (9 g, 7.6 mmol) and thymidine (1.5 g, 6.2 mmol) weredried by coevaporation with pyridine (2x 10 mL), redissolvedin dry pyridine (100 mL), and treated with TPS-C1 (8.06 g, 26.6

by guest on May 5, 2016

http://nar.oxfordjournals.org/D

ownloaded from

5418 Nucleic Acids Research, 1994, Vol. 22, No. 24

mmol) followed by tetrazole (0.53 g, 7.6 mmol). The reactionmixture was stirred under dry nitrogen for 4 h and partitionedbetween 2.5% aqueous NaHCO3 (300 mL) and methylenechloride (2 x300 mL). The organic layer was washed widi water(100 mL), dried over sodium sulfate and evaporated to a foam.This material (10 g) was dissolved in CH2C12 (30 mL), filteredthrough a glass funnel to remove insoluble materials, and injectedin three separate aliquots onto a normal phase silica column whichwas eluted at 40 mL/min with a linear gradient from 0-10%isopropanol in CH2C12 containing 0.1% pyridine. 2a and 2bwere eluted at 32.2 and 31.0 minutes respectively. Theappropriate fractions were pooled and evaporated to give whitesolids. This procedure yielded 470 mg of 2a, 500 mg of 2b, and690 mg of a mixture of isomers.

2a: 'H NMR: 6 (ppm) 11.35 (s, 1H, NH), 11.27 (s, 1H, NH)7.44 (s, 2H, H^, 6.8-7.43 (m, 13H, trityl), 6.15 (t, 1H, H r ) ,6.09 (t, 1H, H r ) , 5.41 (d, 1H, 3'-OH) , 5.12 (m, 1H, H3.),4.05-4.21 (m, 3H, Hy <,• 5>), 3.88 (m, 1H, H4), 3.68 (m, 1H,H4.) 3.6 (s, 6H, OCH3'),'3.22 (m, 2H, CH2), 3.15 (m, 2H,H5.5.), 2.42 (m, 2H, H2.2.), 2.2 (m, 4H, H2. T, CHz), 1.67 (s,3H; CH3), 1.35 (s, 3H,CH3). 31P NMR: S'(ppm) 30.45 (s).

2b: >H NMR: <5 (ppm) 11.36 (s, 1H, NH), 11.26 (s, 1H, NH)7.45 (s, 1H, Hfi), 7.39 (s, 1H, Hg), 6.8-7.34 (m, 13H, trityl),6.16 (t, 1H, H,-), 6.08 (t, 1H, H r ) , 5.4 (d, 1H, 3'-OH), 5.15(m, 1H, H3.), 4.1 (m, 1H, H3.), 3.66-4.0 (m, 3H, H4.5. y),3.6 (s, 6H, OCH3), 3.4 (m, 2H, CH2), 3.18 (m, 2H, H 5 ' v ) ,2.45 (m, 2H, H2. r), 2.18-2.32 (m, 4H, H2.2.,CH2), 1.68 (s,3H, CH3), 1.4 (s, 3H, CH3). 31P NMR: 8 (ppm) 31.03 (s).

The chloro derivatives 4a and 4b were prepared by a similarmethod.

4a: 'H NMR: 8 (ppm) 11.34 (s, 1H, NH), 11.26 (s, 1H, NH)7.45 (s, 2H, Ho), 6.8-7.3 (m, 13H, trityl), 6.16 (t, 1H, H,.),6.10 (t, 1H, H,-), 5.1 (m, 1H, H3-), 4.1-4.26 (m, 3H,H3. J . J , ) , 4.14 (m, 1H, R,.), 3.88 (m, 1H, H ^ 3.7 (s, 6H,OCH3), 3.35 (m, 2H, CH2), 3.21-3.31 (m, 2H,' Hy y), 2.44(m, 2H, Hrx), 2.25 (m, 2H, CHJ, 2.07 (m, 2H, HT'2.), 1.69(s, 3H, CH3), 1.4 (s, 3H, CH3). 31P NMR: 6 (ppm) 29.52 (s).

4b: 'H NMR: 8 (ppm) 7.46 (s, 1H, H^), 7.40 (s, 1H, Hg)6.8-7.3 (m, 13H, trityl), 6.18 (t, 1H, H r ) , 6.09 (t, 1H, H r ) ,5.13 (m, 1H, H30, 4.0-4.12 (m, 3H, H 3 . j 5 . ) , 3.9 (m, 1H,R,.), 3.77 (m, 3H, H,., CHJ, 3.70 (s, 6H,OCH3), 3.01-3.3(m, 2H, H5. 5.), 2.46 (m, 2H, HTX), 2.24 (m, 2H, CHJ, 2.0(m, 2H, H2;2-), 1-70 (s, 3H, CH3), 1.43 (s, 3H, CH3).

31PNMR: 8 (ppm) 29.26 (s).

0-t5'-0-(Dimethoxytrityl)thvmidin-3'-yl]-0-[thvmidln-3'-(cyanoethoxy-N,N-diisopropylaminophosphine)-5'-yl]-[2-(3,4,5,6-tetrabromophthalimido)ethyl]phosphonate (6)A sample of 2 (500 mg, 0.38 mmol) was dried by coevaporationwith pyridine, dissolved in dry acetonitrile (5 mL) and methylenechloride (7 mL) under nitrogen, and treated with stirring with2^anoethoxy-N,N,N\N'-tetraisopropylphosphorodiamidite (1.2mL, 3.8 mmol), diisopropylamine (0.3 mL, 2.1 mmol), andtetrazole (106 mg, 1.52 mmol). After 2 h at room temp themixture was partitioned between 5 % aqueous sodium bicarbonateand memylene chloride (50 mL of each). The aqueous layer wasextracted with methylene chloride (2 X 70 mL) and the combined

organic layers were washed with water (2 X 50 mL) and driedover magnesium sulfate. The solids were removed by filtrationand the filtrate was concentrated to a gum which was purifiedby column chromatography on silica gel (55 g). The column waseluted with a linear gradient of 0-15% ethanol in methylenechloride containing Et3N (0.1%) and the appropriate fractionswere combined and evaporated to yield the purified material asa foam. This material was dissolved in methylene chloride (10mL) and added dropwise with stirring to pentane (450 mL). Theprecipitate was collected and dried to yield 448 mg (78%) ofpure 6. 31P nmr 8 (ppm) 31.239 (s), 150.858 (s).

The chloro derivative 7 was prepared by a similar method.Yield 173 mg (45%), 31P nmr 8 (ppm) 29.09 (s), 148.65 (s).

O- [ 5 ' - 0 - ( D i m e t h o x y t r i t y l ) t h y m i d i n - 3 ' - y l ] - O -(thymidin-5 '-yl)-(2-aminoethyl)phosphonate (5)

A sample of 2b (36 mg, 0.03 mmol, Sp isomer) was dissolvedin anhydrous CH3CN (800 fiL) and treated withethylenediamine (400 fiL). After one hour at room temperaturethe reaction mixture was dried under reduced pressure andcoevaporated with ethanol (5x500 mL) and toluene (2x500mL). The residue was purified on a C4 semipreparative HPLCcolumn (Waters RCM, 25mm x 10 cm) with a gradient of35-80% acetonitrile in 0.1 M TEA A. The deprotected com-pound, eluting at 14.9 min, was collected and evaporated todryness to give 5b, Sp isomer, as a white solid, 9.3 mg (39%).

'H NMR: 8 (ppm) 7.47 (s, 1H, H^, 7.46 (s, 1H, H*,),6.86-7.43 (m, 13H, trityl), 6.20 (t, 1H, H r ) , 6.15 (t, 1H,H r ) , 5.09 (m, 1H, H3), 4.16 (m, 1H, H3), 4.11 (m, 1H, H4),3.94-4.07 (m, 2H, H5. s>), 3.84 (m, 1H, H4.), 3.71 (s, 6H,OCH3), 3.21-3.28 (m, 2H, H v 5.), 2.74 (m, 2H, CHj), 2.43(m, 2H, H2. 2.), 2.06 (m, 2H, H2. 2.), 1.97 (m, 2H, CH2-P,J = 18.0 Hz), 1.74 (s, 3H, CH3j, 1.43 (s, 3H, CH3). 3IPNMR: 8 (ppm) 33.014 (s).

An identical procedure was used to prepare the Rp isomer5a, using (70 mg, 0.05 mmol) of 2a. The deprotected materialwas eluted at 16.19 min and these fractions were collected andevaporated to dryness to give 5a as a white solid, 46.5 mg(82%).

'H NMR: 8 (ppm) 7.49 (s, 1H, t y , 7.46 (s, 1H, HJ,6.87-7.37 (m, 13H, trityl), 6.20 (t, 1H, H r ) , 6.15 (t, 1H,H r ) , 5.11 (m, 1H, H3.), 4.21 (m, 1H, H3.), 4.03-4.19 (m,3H, H3., Hj. 3.), 3.90 (m, 1H, H,.), 3.72 (s, 6H, OCH3),3.20-3.27 (m, 2H, H5. s.), 2.71 (m, 2H, CH2), 2.41 (m, 2H,H2. 2 ) , 2.09 (m, 2H,H 2 . v), 1.92 (m, 2H, J=17.5 Hz,CH2-P), 1.73 (s, 3H, CH3),' 1.41 (s, 3H, CH3).

3 IP NMR: 8(ppm) 32.754 (s).

O-(Thymidin-3'-yl)-O-(thymidin-5'-yl)-(2-aminoethyl)-phosphonate (10)A sample of 5b (9.3 mg, 0.01 mmol, Sp isomer) was dissolvedin 1 % dichloroacetic acid in CH2C12 (1 mL). After one hour atroom temperature the reaction mixture was dried under reducedpressure, redissolved in lmL H2O and extracted with 3 X 1 mLethyl acetate. The aqueous layer was purified on a C4

semipreparative HPLC column (Waters RCM, 25 mm x 10 cm)with a gradient of 5-70% acetonitrile in 0.1 M TEAA. Thedetritylated compound, eluting at 11.8 min, was collected andevaporated to dryness to give 10b, Sp isomer, as a white solid.'H NMR: 8 (ppm) 7.67 ( s, 1H, H&), 7.48 (s, 1H, H^, 6.18(m, 2H, H r) , 5.0 (m, 1H, H3,), 4.2 (m, 1H, H3.), 4.12 (m, 2H,

by guest on May 5, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research, 1994, Vol. 22, No. 24 5419

TpTpTpTpTpTpTpTpTpTpTpTpTI I I I I IX X X X X X

8: X n CH2CH2NH3*

TpTpTpTpTpTpTpTpTpTpTpT

CHJCHJNHJ*

9

TpTpTpTpTpTpTpTpTpTpTpTpTpTI I I I I IX X X X X X

1 1 : X

Ruor-p-TpTpTpTpTpTpTpTpTpTpTpTpTI I I I I IX X X X X X

12b: X = C H J C H J N H , *

SERIES a: RptSOMER, SERIES b: SplSOMER

Figure 2. Oligonucleotide (2-aminoethyl)phosphonates.

H y y ) , 4.01 (m, 1H, a,.), 3.9 (m, 1H, H,-), 3.58 (m, 2H,H5/5.). 2.77 (q, 2H, CHj), 2.11-2.31 (m, 4H, H2.,2'), 198 (m,2H, CHj), 1.79 (s, 3H, CH3), 1.76 (s, 3H, CH3).

31P NMR:29.71 (s).

An identical procedure was used to prepare the Rp isomer10a, using (46.5 mg, 0.053 mmol) of the compound 5a. Thedetritylated material was eluted at 8.48 min. Fractions were col-lected and evaporated to dryness to give 10a as a white solid.'H NMR: 5 (ppm) 7.67 (s, 1H, HJ, 7.50 (s, 1H, H*;), 6.17(t, 2H, H,-), 5.0 (m, 1H, H3) , 4.23 (m, 1H, H3.), 4.09-4.19(m, 2H, H5.5.), 4.0 (m, 1H, H4-), 3.9 (m, 1H, H4.), 3.58 (m,2H, H5.5.), 2.79 (m, 2H, CH2), 2.07-2.33 (m, 4H, H r , H r ) ,2.0 (m,'2H, CH2), 1.78 (s, 3H, 5-CH3), 1.76 (s, 3H, CH3).3IP NMR: 29.71 (s).

Synthesis of oligonucleotidesOligonucleotides were synthesized using a standard 1 /*mol scalephosphoramidite cycle except that a coupling time of 1 min/cyclewas used, with the dimer phosphoramidites (6 or 7, a or b) be-ing delivered at a concentration of 0.15 M in acetonitrile. Coupl-ing yields of >96% were routinely obtained. For 9a, 9b, l l aand l ib, additions of a conventional thymidine phosphoramiditewere employed to introduce phosphodiester linkages at the ap-propriate positions. At the conclusion of the synthesis, the solidsupport containing the tritylated oligonucleotides was dried invacuo overnight, transferred to a screw cap glass vial and treatedwith ethylenediamine (500 /iL) for 30 min at 55°. The super-natant was collected, evaporated to dryness, and the supportwas washed with ethanol (2 X200 fiL) followed by TEAA (0.1M, 2 x200 fiL) and ethanol (2 X200 /tL). The washes were com-

bined, evaporated to dryness, and the residues from the super-natants and washes were dissolved in water (250 /iL each),combined, filtered through a 0.45 n membrane and loaded on-to a preparative reverse phase C4 HPLC column. The columnwas eluted with 0.1 M TEAA/CH3CN according to the follow-ing conditions: 0 - 5 min 5% CH3CN; 5 - 2 0 min 5-30%CH3CN; 20-25 min 30-50% B, flow rate of 10 mL/min.The fractions containing product, which were generally elutedbetween 22-26 min, were combined, lyophilized, and detrityla-tion was performed using 80% acetic acid (1 mL) for 30 minat room temp. The solution was lyophilized, and the residueswere dissolved in TEAA (1 mL) and injected onto a C4 col-umn. The fractions containing pure material, which weregenerally eluted be ween 14-18 min, were combined,evaporated to dryness and the residue was converted into thesodium form and lyophilized to give the requiredoligonucleotide. Yields varied between 7 and 64 OD260 units.

The fluoresceinated oligonucleotide 12b was synthesized asdescribed above except that a fluorescein phosphoramidite (100fiM, Glen Research, Stirling, VA) was used for addition offluorescein to the 5'-terminus. The fluorescein amidite wasdissolved in dry acetonitrile (0.75 mL) and a double couplingprocedure (15 min each) was employed; under these conditionscoupling yield of >95% was obtained, as determined by tritylassay. A trityl-off ending procedure was employed, followedby deprotection and HPLC purification as described above. Thepeak eluting at 16.75 -17.4 min was collected, lyophilized andconverted into the sodium form to give pure 12b (25 OD260

units).

Thermal denaturation experimentsThese experiments were carried out using a Gilford ResponseII temperature controlled spectrophotometer by monitoring thechanges in absorbance at 260 nm versus temperature, with aheating rate of l°/min from 0-70° . Melting curves were ob-tained both in low salt (150 mM NaCl, 10 mM Na2HPO4) andhigh salt (1 M NaCl, lOmM Na2HPO4) conditions at pH 7.0.Transition temperatures were obtained from the first orderderivative plot of absorbance versus temperature.

Hydrolytic stability of 10a and 10bThe dinucleotides 10a and 10b were incubated at pH 7.1 inTEAA buffer at 37° over 50 h and aliquots were removed atintervals and injected onto a C4 reverse phase HPLC column.The rate of degradation was determined by measurement of thearea remaining under the peak corresponding to startingmaterial. The half lives as determined by this method were 51days for 10a and 45 days for 10b.

Stability to SI nucleaseSolutions of the natural oligomer d-T(pT)12 and the aminoethylmodified oligomer 8a or 8b (3 OD2«,) in 150 /tL of buffer (50mM sodium acetate, 250 mM sodium chloride, 1 mM zincchloride, pH 4.6, containing 50 /tG/mL of bovine serumalbumin) were equilibrated at 25°C for 30 min, treated with10 units of SI nuclease (US Biochemicals) and the solutionswere stored at 25°C. Aliquots (10 /iL) of the reaction mixturewere injected onto an analytical reverse phase HPLC column,eluted with a gradient of 5 —30% acetonitrile in 0.1 M TEAAbuffer (pH 7.1) and the area under the peak corresponding tostarting material was plotted versus time to determine t1/2

values. The procedure for incubation of dimers 10a and 10b

by guest on May 5, 2016

http://nar.oxfordjournals.org/D

ownloaded from

5420 Nucleic Acids Research, 1994, Vol. 22, No. 24

Table 1 Hybridization of oligonucleotkk ammoethylphosphonates to complementary DNA and RNA sequences

Oligo-micleotide

d-T(pT)12

8a8b

d-T(pT),2

8a8b

Isomer Target

n/aRpSp

n/aRpSp

M50mM NaCl, lOmM Na2HPO4, pH 7.blM NaCl, 10 mM Na2HPO4, pH 7.

d-A(pA)n

d-A(pA)12

d-A(pA)12

r-A(pA),2r-A(pA)12

r-A(pA)12

Melting Temp (°C)Low Salt*

345112

3035

High Salt1"

475117

3937

was as previously described for the aminomethyl derivatives[16].

Induction of RNAse H activityA single-stranded DNA - RNA hybrid target molecule was syn-thesized consisting of 8-base deoxynucleotide flanks and a cen-tral r-pA13 core. This molecule was end labeled with T4polynucleotide kinase in the presence of 400 /*Ci of[-y^Pl-ATP (ICN, 7000 Ci/mmol). The labeled compoundswere heated (95°, 5 min) in 50% formamide and purified byseparation on a 12% PAGE with 7M urea using Tris-borate-EDTA buffer (TBE). The isolated bands were frozen, crushedand extracted with 2mM EDTA, filtered through celluloseacetate (0.45 /t), extracted with phenol/chloroform andprecipitated with ethanol. The target molecule (~7 frnol, 40,000cpm) was annealed to 10-, 100- or 1000-fold excess of eitherl la , l i b or d-T(pT),2 as a control by heating to 65° for 10min. in 10 /tL GB (20 mM Hepes pH 7.4, 140 mM NaCl, 5mM KOAc, 5 mM MgC^) followed by slow cooling to roomtemp. One unit of RNAse H (Promega) was added and the solu-tions were incubated at 37° for 20 min. Formamide was addedto 50% concentration and the solutions were heated (95°, 5min), and electrophoresed on a 16% PAGE with 7M urea inTBE. The gels were fixed and dried prior to autoradiography.

Serum stabilityTen pmol of d-T(pT),2 or the alternating aminoethyl oligomersl l a or l i b were 5'-end labelled and isolated as described above.For each time point examined, each oligomer (50,000 cpm) wasincubated at 37° in a 5% humidified CO2 atmosphere with 50/xL of Dulbecco's modified Eagle medium (DMEM, Gibco) con-taining 10% heat-inactivated (56°, 45 min) fetal bovine serum(FBS). Control samples were incubated in plain DMEM. Atthe end of 0,1, 4 or 24 hours, the samples were mixed withan equal volume of formamide, and analyzed by 12% urea-PA-GE. The gels were fixed, washed and dried beforeautoradiography (Kodak XAR-5).

Flow cytometryJY cells (a human B lymphoblastoid cell line) in log phase werewashed and cultured in RPMI containing 10% FBS (heat inac-tivated to 65°C for 30 min.), 50 yM 2-mercaptoethanol, 100U/ml penicillin, 100 /tg/ml streptomycin, and 2 mM L-glutamateat 1.5x10* cells/ml for the indicated times and temperatureswith or without oligonucleotide as indicated. Cells were thenwashed twice in HBSS with 1 % BSA. Data on viable cells were

collected on a FACScan (Becton-Dickinson, Mountain View,CA) containing a 15-mW argon-ion laser with Lysis II software.

Confocal microscopyThis was performed following JY cell culture as described aboveusing a Biorad MRC600, with a krypton/argon laser, Nikonoptiphoto optics with a 60X 1.4 NA objective, and images wereprinted with a SONY UP5000 video printer. Cells were wash-ed but not fixed, and were analyzed in BSS within 15 min.

RESULTS AND DISCUSSION

Our initial strategy for the synthesis of nucleotide (2-aminoethyl)-phosphonates employed the phthalimido group for protection ofthe primary amine of the (2-aminoethyl)phosphonate moiety,since this group was previously used for the preparation of(aminomethyl)-phosphonate derivatives. Accordingly, thethymidine dinucleotide-(2-phthalimidoethyl)phosphonate wasprepared, but difficulties were experienced in the removal of thephthalyl group. A variety of deprotection conditions wereevaluated, including treatment with ethylenediamine in acetonitrilefor up to 6.5 h at room temp, but no cleavage of the phthalylgroup was observed. This result is in marked contrast to theresults with the corresponding dinucleotide-(phthalimido-methyl)phosphonate as previously reported [16] which could becompletely deprotected and isolated in excellent yield upontreatment with ethylenediamine under the same conditions.Hydrazine has routinely been used for the removal of phthalylprotecting groups [18] and deprotection of the base protectinggroups of oligonucleotide methylphosphonates [19]' but thisreagent was not satisfactory for removal of the phthalyl groupssince decomposition was observed.

Because of the difficulties experienced with removal of thephthalyl group, other groups were evaluated for protection ofthe amino functionality. Since halogenated phthalimido groupshave been employed for other purposes [20,21], these wereevaluated for protection of the amino group. Reaction oftetrabromophthalic anhydride with (2-aminoethyl)phosphonicacid produced pyridinium [2-(3,4,5,6-tetrabromophthal-imido)ethyl]phosphonate which could be isolated as a solid.Reaction of this compound with 5'-dimethoxytritylthymidine us-ing TPS-C1 as the coupling agent produced the mononucleotide1, although this reaction was inefficient due to the poor solubilityof the starting material in the reaction medium. Coupling of1 with one equivalent of thymidine gave the dimer as a mixture

by guest on May 5, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research, 1994, Vol. 22, No. 24 5421

of isomers 2a and 2b which could be separated by HPLC, withno evidence for 3',3'-coupling. After removal of the acid labiledimethoxytrityl group, die tetrabromophthalyl group couldreadily be removed by treatment with ethylenediamine inacetonitrile. Thus the introduction of the halogen substituentsinto the phthalimido group produced a protecting group whichis expected to be compatible wirn other nucleic acid components.

Although the tetrabromophdialimido group was suitable forprotection of the amino function in the synthesis ofoligonucleotides, the low yields obtained in the synthesis of themononucleotide prompted an evaluation of the correspondingtetrachloro derivatives. Reaction of pyridinium [2-{3,4,5,6-tetra-chlorophthalimido)ethyl] phosphonate with thymidine was muchmore efficient than for the tetrabromo derivative and themononucleotide 3 could be isolated in 63% yield. The dimer4 was prepared as described for 2, and deprotection of dietetrachlorophthalyl group could be easily accomplished usingethylenediamine for 10 min at 55°. Thus the tetrachloro-phdialimido group appears to be the protecting group of choicefor the aminoethyl series.

Stereochemistry at phosphorusA preliminary determination of the stereochemistry atphosphorus was obtained by an adaptation of the 2D-NMRmethod of Loschner and Engels [22]. These authors determin-ed the stereochemistry about phosphorus of dinucleosidemethylphosphonates based upon die level of interactions bet-ween the methylphosphonate and the C3' and C4> sugar pro-tons. In diis work the 2D-ROESY NMR spectra of the partiallyprotected dimers 5a and 5b indicated that the relative interac-tions of the P-CH2 and C4- protons as compared with the C3'protons were enhanced for one isomer, which was tentativelyassigned as the Rp isomer 5a; details of the NMR data will bedescribed elsewhere. This assignment is consistent with dataobtained from the hybridization experiments which showed that

TaWe 2. Hybridization of 9a and 9b to partially complementary sequencespossessing a mismatch opposite to the aminoethyl group

Oligonucleotide

d-A(pA)12

A(pA)jPC(pA)6

A(pA)3pG<pA)6

A(pA)3pT(pA)6

A(pA)6pC(pA)5

A(pA)(PG<s>A)i

A(pA)a>T(pA)5

Melting temperature (°C) of duplexwith matched or mismatched sequence*9a

38242524252525

9b

30171818171717

d-T(pT)12

36222323222323

M50mM NaCl, lOmM Na^HPO^, pH 7.

the oligonucleotide with Rp isomer exhibited a much higher Tm

than the corresponding Sp isomer (see below).

Hybridization experimentsThe alternating sequences 8a and 8b were hybridized to the com-plementary sequence d-A(pA)12 and the results are shown inTable 1. As with the aminomethyl series, one isomer (Rp)hybridized more strongly than its natural counterpart, whereasthe other did not form a stable hybrid. It is of interest that underlow salt conditions the Rp isomer of die aminoediyl oligomerformed a more stable hybrid than the correspondingaminomethyl analog (Tm = 51° versus 45° for the latter). Theeffects of salt concentration on the dissociation temperaturesof hybrids of the Rp isomer were also investigated. For diehybrid widi DNA, no salt dependence was observed, whereasfor the hybrid with RNA, the salt dependence was relativelylow as compared with a natural duplex. This type of reducedsalt dependence has previously been observed for otheroligonucleotides with cationic groups [15]. The relatively highTm of the Rp isomer versus the Sp isomer is consistent withdata previously obtained for stereoisomers of oligonucleotidemethylphosphonates [23].

Since die presence of positive charges might potentially in-crease die likelihood of non-specific hybridization, die stabili-ty of hybrids of oligonucleotides 9a and 9b, each possessingone aminoediylphosphonate moiety, widi a fully complemen-tary sequence and widi a series of partially mismatched com-plements was examined. The mismatched bases were introducedopposite to the aminoediyl group and die results are shown inTable 2. In terms of hybridization to the fully complementarytargets, a substantial difference was observed for die twoisomers, one producing an elevation of 2° versus naturalwhereas die odier caused a marked depression of 6°. In all casesdie introduction of mismatches produced a substantial depres-sion of melting temperature, i.e. 12-13° for die Sp isomer 9b,13-14° for the Rp isomer 9a, versus 13-14° for the naturalsequence. Since die magnitude of depression for each mismat-ched pair was approximately the same in all cases, this sug-gests that the introduction of an aminoediyl group does not resultin a loss of specificity. When similar mismatch experiments werecarried out with the alternating sequence 8a, a drop in Tm of9-11° was observed for die mismatched duplexes as comparedwith the fully complementary sequences. Thus the presence orabsence of paired bases appears to be die primary factor forthe stability of diese duplexes.

Since the ionization of die cationic groups depends on the pHof the medium, a series of hybridization experiments was car-ried out widi 8a in which the pH of die solution was varied anddiese results are summarized in Table 3. For the natural se-quences d-T(pT)12 and d-A(pA),2 die dissociation temperature

TaWe 3. Effect of pH on the

Duplex

d-A(pA)12 + d-T(pT)12

8a + d-A(pA)12

melting temperature

5

3443

6

3451

of duplexes of d-T(pT)l2 or 8a with

Dissociation temperature1

PH7 8 9

3451

3651

3547

d-A(pA)12

10

3043

M50mM NaCl, lOmM

by guest on May 5, 2016

http://nar.oxfordjournals.org/D

ownloaded from

5422 Nucleic Acids Research, 1994, Vol. 22, No. 24

was relatively constant over the pH range of 4 - 9 , with adecrease being observed at pH 10. This latter decrease mightbe due to the destabilization caused by the ionization of the im-ino protons of the thy mine residues. The hybridization of 8aversus d-A(pA)12 showed a much more dramatic variation withchanges of pH, with the Tm's being in all cases substantiallyhigher than the natural duplex. A Tm of 51 ° was observed overthe pH range of 5 - 7 with a substantial decrease at pH 8-10.These data suggest that at pH 5 - 7 the ionized aminoethyl groupsplay an important role in enhancing stability either by interstrandor intrastrand ionic attractions, whereas at higher pH valuesthese attractive forces are absent since the amino groups areunionized, or only partially ionized. These data are consistentwith the pK values reported for (2-aminoethyl)phosphonic acid[24].

The hybridization of the alternating-backbone oligonucleotides8a and 8b to RNA was also studied, using r-A(pA)[2 as thetarget. Under low salt conditions the Rp isomer 8a exhibiteda Tm 5° higher than the natural counterpart, whereas underhigh salt conditions the Tm of 8a was actually lower than thenatural sequence. As with the hybridization to DNA, themodified oligonucleotides show less salt dependence, as wouldbe expected for a hybrid in which the charge-charge repulsionsof the backbone have been reduced. The Sp isomer 8b did notexhibit a detectable transition under any salt conditions.

A

4 5

Resistance to SI nucleaseThe unprotected dinucleotides 10a and 10b were incubated withSI nuclease, which is capable of degrading single-strandedDNA, and the digestion products were examined by HPLC. Theaminoethyl-phosphonate derivatives were found to be completelyresistant after a 30 minute incubation period whereas the naturaldimer was completely degraded after this time. The action ofSI endonuclease on the alternating sequences 8a and 8b wasalso examined. The natural sequence d-T(pT)|2, which was us-ed as a control, was rapidly degraded with a t1/2 of only fourminutes, whereas the Sp isomer 8b was completely undegrad-ed after 24 hours and the Rp isomer 8a showed a small amount(approx. 6%) of degradation. At the end of these experimentsthe activity of the enzyme was confirmed by addition of d-T(pT),2 to the digestion mixture followed by a short incuba-tion period and HPLC analysis, which confirmed that the en-zyme was still capable of degrading the natural oligonucleotide.

u.

150

100

50

a 12b, 37'C• 1 2b, 4'Co PO, 37'C• PO, 4'C

3 4 5 6TIME (HOURS)

B

4 5

B300

200

100

• 12b, 37'C• 12b, 4'Co PO, 37'C• PO, 4'C

I102 103 104

ODN CONCENTRATION (nM)10=

Figure 3. Serum stability of d-T(pT)12 (panel A), and l la (panel B). Lanes1.3,5,7 were incubated in the presence of FBS, and lanes 2.4,6,8 without FBS.Incubation times: lanes 1 and 2, 0 h; lanes 3,4, 1 h; lanes 5,6, 4 h: lanes 7,8,24 h.

Figure 4. A: Time course of oligonucleotide uptake. Fluorescein conjugatedalternating aminoethyl oligonucleotide 12b or phosphodiester (PO)oligonucleotide (2 /iM concentration) was cultured with JY cells at the indicatedtemperature and times before washing and flow cytometry to determine the meanfluorescence intensity (MFI). B: Concentration dependence of oligonucleotideuptake. Cells were cultured for 3 h with the oligos from panel A at the concen-trations shown prior to washing and analysis.

by guest on May 5, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research, 1994, Vol. 22, No. 24 5423

Figure 5. Intracellular oligonucleotide localization. Panels A and B, FTT-C-phosphodiester; panels C and D, FITC-alternating aminoethyl oligonucleoude12b (2 fiM concentration). Panels A and C, phase contrast; panels C and D,confocal microscopy.

Thus the enzyme was apparently incapable of degradingphosphodiester linkages which are flanked by aminoethyl-phosphonates. This protective effect might be steric in nature,or alternatively the aminoethyl groups might exert a protectiveeffect by the formation of internal ion pairs. Similar results onthe enzymatic degradation of oligonucleotides withstereoregular, alternating methylphosphonate/phosphodiesterbackbones have been reported by Miller et al [25].

Induction of RNase H activityIt was of interest to determine if the alternating oligomers whichform stable hybrids would induce RNAse H activity whenhybridized to a target RNA molecule. For this experiment anoligonucleotide with an rA13 core and deoxynucleotide flank-ing regions was synthesized and 32P-end labelled for use as thetarget. Cleavage of the RNA region of the target would thusbe detected by the appearance of a shorter 32P-labelled frag-ment. After incubation of samples of 8a, 8b and d-T(pT)13

with labelled target, the products were examined by native PA-GE. Analysis of the autoradiograms confirmed the lack ofhybridization for the Sp isomer 8b, whereas the expected retar-dations due to hybridization for the control oligomer and theRp isomer 8a were observed (data not shown). Under the con-ditions employed, the d-T(pT)i3 control was effective in induc-ing RNAse H activity with as little as a 10-fold molar excess

over target, whereas the Rp isomer 8b showed no sign of suchactivity even with a 1000-fold excess of oligonucleotide.

Serum stability32P labelled oligonucleotides were needed in order to assess theserum stability of the alternating aminoethylphosphonate-/phosphodiester backbones. Preliminary experiments showed that8a and 8b could not be labelled effectively, presumably due tothe presence of the aminoethyl moiety adjacent to the 5'-terminus.Oligonucleotides l la , l ib , possessing phosphodiester linkagesadjacent to the 5'-terminus, were therefore prepared and 32Plabelled at the 5'-position. The labelled oligonucleotides wereincubated in conditioned medium containing 10% FBS and theproducts were examined by autoradiography. As shown in Figure3, the nuclease stability of the alternating aminoethyl linkageswas superior to the d-T(pT)12 oligonucleotide control. Thenatural oligomer showed significant degradation after only onehour, and full-length oligomer was virtually undetectable after24 hours. In contrast, the degadation of l l a and l i b was barelyvisible after 24 hours. The slight retardation observed on PAGEafter longer time periods up to 24 hours might be an indicationof cleavage of the phosphodiester at the 3'-position to producea 12mer with one less negative charge. Although a 12mer wouldnormally be expected to migrate faster for a phosphodiester, theremoval of a negative charge in the alternating oligomer couldhave a greater retarding effect than the loss of one base.

Cell uptake

Flow cytometry was used to determine whether anoligonucleotide with alternating positive and negative chargeson the backbone might be taken up by cells. This technique hasbeen demonstrated to be useful for studying and comparing theuptake and intracellular localization of different oligonucleotidesconjugated to fluorescent molecules such as fluorescein [26].Oligonucleotide uptake in cell lines has previously been shownto be saturable, sequence independent, and temperature andenergy dependent [27]. JY cells were cultured with the5'-fluoresceinated oligonucleotide 12b (Sp isomer), or with a5'-fluoresceinated, twenty base, mixed sequence oligonucleotideas control, at 2 mM concentration at 4°C for varying times andcompared to those cultured at 37 °C for the same time beforewashing and flow cytometry to determine the mean fluorescenceintensity (MFI). As shown in Figure 4, (panel A), uptake ofboth oligonucleotides was highly temperature-dependent, com-patible with uptake through an endocytic pathway. Cell surfacebinding (defined as uptake in cells cultured at 4°C) of botholigonucleotides was very similar. To evaluate the concentra-tion dependence, oligonucleotides were cultured for 3 hours atthe indicated temperature and concentration before washing andflow cytometry as above. In contrast, the level of cellular up-take of the oligonucleotides at the three hour time point studiedwas highly dependent on the oligonucleotide concentration(Figure 4, panel B). Both experiments were performed threetimes with similar results. It is of interest that at lower concen-trations (less than 1 /xM) the level of phosphodiester uptakegenerally exceeded that of the alternating aminoethyl, while athigher concentrations, the converse was true. Thus theoligonucleotide with alternating positive and negative chargeson the backbone appears to be taken up as well as, or undercertain conditions better than, the control (phosphodiester) se-

by guest on May 5, 2016

http://nar.oxfordjournals.org/D

ownloaded from

5424 Nucleic Acids Research, 1994, Vol. 22, No. 24

quence. Further studies are needed to determine if uptake isdependent upon the stereochemistry of the aminoethyl group.

The temperature dependence of cellular oligonucleotideassociation seen in the experiments above suggested that cultureat 37°C resulted in oligonucleotide internalization. To test thishypothesis, confocal microscopy was used to confirm that boththe phosphodiester and alternating aminoethyl oligonucleotideswere present intracellularly (Figure 5). The top panels showthe phase and confocal images for same field of view of tworepresentative cells that had been incubated with the fluores-ceinated phosphodiester oligonucleotide for 3 hours and the bot-tom panels show the phase and confocal images for a cellincubated with 12b for the same time. The pattern of thisfluorescence was typically speckled and cytoplasmic with verylittle nuclear fluorescence. This result suggests that theoligonucleotides were sequestered in endosomal compartments.

CONCLUSIONS

(2-Aminoethyl)phosphonates represent an interesting new typeof oligonucleotide backbone modification which introducespositive charges into the molecule. The stereoisomers of thy minedinucleotides are readily resolved by procedures that can be ap-plied to the synthesis of isomers of dimers of other nucleotides.The studies described above using oligonucleotides with alter-nating (2-aminoethyl)phosphonate/phosphodiester backbones in-dicate that the Rp isomers are capable of forming more stableduplexes than their natural counterparts. The Sp isomers do notform stable hybrids, presumably due to unfavorable steric in-teractions between the aminoethyl groups and other portionsof the duplex. The increase in duplex stability is more markedwith DNA than with RNA as the complementary strand, andless dependent upon salt concentration. More stable hybrids areformed at pH's which favor protonation of the amino groups.Surprisingly, oligonucleotide (2-aminoethyl)-phosphonates formmore stable hybrids than the corresponding (aminomethyl)phos-phonate derivatives even though the former possess more stericbulk. Backbones of this type are not degraded by SI nucleaseto any significant extent. The most striking difference between(aminomethyl)- and (2-aminoethyl)phosphonates is their dif-ference in stability in aqueous solution. The (aminomethyl)-phosphonate linkage is rapidly hydrolyzed at pH 7 toaminomethyl mononucleotides, whereas the corresponding(2-aminoethyl)-phosphonate is completely stable under the sameconditions. This difference in stability may be due to the reduc-tion in the electropositivity of the phosphorus atom due to theadditional intervening methylene group. A preliminary studyon cell uptake indicated that under certain conditions a netneutral oligonucleotide was taken up as well as, or somewhatbetter than a phosphodiester control sequence. Since thesestudies demonstrate that oligonucleotides with aminoethyl substi-tuents hybridize well to nucleic acid targets and are taken upby cells, their potential value as antisense therapeutic agentsdeserves further investigation.

REFERENCES

1. Shoji, Y., Akhtar, S., Periasamy, A., Herman, B., and Juliano, R , L.(1991) Nucleic Acids Res., 19, 5543-5550.

2. Froehler, B , Ng, P., and Matteucci, M. (1988) Nucleic Acids Res., 16,4831-4839.

3. Egholm, M. E., Buchardt, O., Nielsen, P. E., and Berg, R. H. (1992)J. Am. Chan. Soc. 114, 1895-1897.

4. Quartin, R. S., Brakel, C. L. and Wetmur, J. G. (1989) Nucleic AcidsRes., 18, 7253-7262.

5. Wagner, E., Zenke, M., Cotten, M.,Beug, H., and Birnstiel.M. L. (1990)Proc. Natl. Acad. Sci. USA, 87, 3410-3414.

6. Kumagai, A. K., Eisenberg, J. B., and Pardridge, W. M. (1987) J Biol.Chem., 262, 15214-15219.

7. Pardridge, W. M., Kumagai, A. K., and Eisenberg, J. B. (1987) Biochem.Biopkys. Res Common., 146, 307-313.

8. Arnold, L. J. (1985) Methods in Enzymology, 112, 270-2859. Pardridge, W. M. and Boado, R. J. (1991) FEBSLens., 288, 30-32.

10. Lemahre, M., Bayard, B., and LeBleu, B. (1987) Proc. Natl. Acad. Sci.USA, 84, 648-652.

11. Stevenson, M. and Iverson P. L. (1989). J. Gen. Virol, 70, 2673-268212. Pieken, W. A., Olsen, D. B., Benseler, F., Aurup, H., and Eckstein, F.

(1991) Science, 253, 314-317.13. Manoharan, M., Guinosso, C. J. and Cook, P. D. (1991) Ten Lens. 32,

7171.14. Hashimoto, H., Nelson, M. G. and Switzer, C. (1993) J. Am. Chem. Soc ,

115, 7128-7134.15. Letsinger, R. L.,Singman, C N., Histand, G., and Salunkhe, M. (1988)

J. Amer. Chem. Soc., 110, 4470-4471.16. Fathi, R., Huang, Q., Delaney, W. and Cook, A. F. (1994) Bioconjugate

Chem., 5, 47-57.17. Rychlik, W., and Rhoads, R. E. (1989) Nucleic Acids Res., 17,8543-8551.18 Seyferth, D., Marmor, R. S. and Hilbert, P. (1971) J. Org. Chem., 36,

1379-1386.19. Miller, P., Cushman, C. D. and Levis, J. T. (1991) in Eckstein, F. (ed),

Oligonucleotides and Analogs-A Practical Approach, IRL Press, Oxford,pp. 137-154.

20. Motawia, M. S., Jacobsen, J. P. and Pedersen, E. B. (1989) Chemica Scrip-ta, 29, 51-55 .

21. Motawia, M. S., Nawwar, G A. M., Andreassen, E. S., Jacobsen, JP. and Pedersen, E. B. (1987) Uebigs Ann. Chem., 1111-1114.

22. Loschner, T. and Engels, J. W. (1990) Nucleic Acids Res., 18, 5083-508823. Lesnikowski, Z. J., Jaworska, M. and Stec, W. J. (1990) Nucleic Acids

Res., 18, 2109-2115.24. Petrov, K. A., Chauzov, V. A. and Erokhina, T. S. (1974) Russian Chem

Reviews, 43, 984-1006.25. Miller, P. S., Dreon, N , Pulford, S. M. and McParland, K. B (1980)

J. Biol. Chem., 255, 9659-9665.26. Zhao, Q., Matson, S., Herrera, C. J., Fisher, E., Yu, H. and Krieg, A

M. (1993) Antisense Res. Development, 3, 53-66.27. Jaroszewski, J. W. and Cohen, J. S. (1991) Adv. Drug Delivery Revs.,

6, 235-250.

ACKNOWLEDGEMENTS

A.M.K. is a Pfizer Scholar and was supported by grants fromthe Veterans Administration, Carver Foundation and NIH R29AR 42556-01.

by guest on May 5, 2016

http://nar.oxfordjournals.org/D

ownloaded from