A novel concept for ligand attachment to oligonucleotides via a 2'-suc
http://www.100md.com
《核酸研究医学期刊》
Institute for Pharmaceutical Chemistry, Universit?t Wien, Pharmaziezentrum, Althanstra?e 14, 1090 Wien, Austria 1 Department of Clinical Pharmacology, Section of Experimental Oncology and 2 Department of Dermatology, Universit?t Wien, W?hringer Gürtel 18–20, 1090 Wien, Austria
*To whom correspondence should be addressed. Tel: +431 4277 55103; Fax: +431 4277 9551; Email: christian.noe@univie.ac.at
ABSTRACT
Conjugation of ligands to antisense oligonucleotides is a promising approach for enhancing their effects. In this report, a new method for synthesizing oligonucleotide conjugates is described. 2'-Amino-2'-deoxy-5'-dimethoxytrityl-uridine was select ively acylated with a succinic acid linker at the 2' position. This compound was incorporated at the 3' end of an oligonucleotide corresponding to the sequence of Oblimersen. The carboxyl group was protected for oligonucleotide synthesis as a benzyl ester, which could be selectively cleaved at the solid phase by a catalytic phase transfer reaction using palladium nanoparticles as catalyst. An oligonucleotide–fluorescein conjugate was prepared by condensation of aminofluorescein. Circular dichroism spectroscopic experiments showed a B-DNA type structure. The melting temperature of the duplex was only slightly lower than that of Oblimersen. Biological activity measured by western blotting resulted in a Bcl-2 target downregulation nearly identical to that of control Oblimersen on human melanoma cells, proving that this method is attractive for the binding of ligands located in the minor groove.
INTRODUCTION
Several antisense oligonucleotides, designed for sequence-specific binding to a target mRNA, are now in various phases of clinical trials, many of them as anticancer agents (1). Initial attempts at increasing the efficacy, nuclease resistance and pharmacokinetic properties of antisense oligonucleotides have been focused on the chemical structure, leading to a plethora of modifications (2). Lately, an approach used quite often to specifically enhance the characteristics of antisense oligonucleotides is the attachment of various ligands (3).
Such attached functional moieties may be effectors, increasing the antisense effect; modulators, which can change their biophysical or pharmacodynamic properties; or detectors, which can be employed as diagnostic tools. For example, attachment of lipophilic molecules such as cholesterol can improve the cellular uptake of antisense oligonucleotides (4). Reporter groups such as biotin or fluorescent dyes such as fluorescein are used extensively in DNA-based diagnostics as well as for following cellular trafficking of antisense oligonucleotides (5). Much effort has recently been put into the field of oligonucleotide–peptide conjugates (6).
Due to the multivalent structure of oligonucleotides, there are many different synthetic possibilities for the attachment of ligands (Fig. 1). The majority of these conjugated compounds have been derivatized via a suitable linker at a terminal hydroxyl group of the oligonucleotide (7–9). Since the synthesis of oligonucleotides usually proceeds from 3' to 5', attachment to the 5'-hydroxy group can be achieved in a quite straightforward manner. Conjugation to the 3'-hydroxy function can be done by using an appropriate, orthogonally protected linker. Alternatively, incorporation of ligands can be achieved at nucleobases with a suitable linker attached to an amino group of adenosine, guanosine or cytosine, or the carbon 5 of pyrimidine bases. Attachment to the phosphorous backbone can be achieved by reaction of phosphate with an amino group of the peptide.
Figure 1. Possibilities for attachment of ligands to oligonucleotides.
In contrast, attachment of ligands to the carbohydrate moiety has been investigated less extensively. The increased binding affinity of short 2'-O-alkyl and 2'-O-ethylene glycol-modified oligonucleotides shows that this position is attractive for modifications (10–12) and for achieving structural diversity in the minor groove. Generally, conjugations at the 2' position of the carbohydrate moiety interfere less with base pairing than conjugations at backbone or nucleoside sites. By covalent binding of an aminoalkyl or a thioalkyl chain at the 2' position, post-synthetic derivatization has been made possible (13). Usually, a building block containing a suitably protected linker, mostly with an amino group, is incorporated before or during oligonucleotide synthesis and the amino group is either selectively unmasked or deprotected at the same time as the amino groups of the nucleobases. Linkers which have been used include aminoethyl (14), aminobutyl (15) and aminohexyl (14). To perform the conjugation reaction while the oligonucleotide is still resin bound and fully protected holds obvious advantages regarding selectivity. To date, few possibilities of achieving this have been published. For example, photolabile protecting groups can be used (15). After unmasking, peptides can be coupled to amines by reaction of the activated carboxyl groups. Other ligands can be attached to an amine via an isothiocyanate linkage (16).
Another approach consists of coupling ligands bearing a short alkyl linker directly to 2'-amino-2'-deoxyuridine (17). Due to the steric hindrance at the amino position of 2'-amino-2'-deoxyuridine, coupling yields were relatively low using this procedure (15).
Herein we describe a new strategy, in which the linker is bound directly to the starting nucleoside of solid phase synthesis. This approach has the advantage that possible negative properties of a ligand regarding hybridization affinity as well as problems during solid phase synthesis related to the steric hindrance are minimized. Compared with the dominantly employed 2'-O-aminoalkyl tether, the usage of 2'-amino-2'-deoxyuridine as the modified nucleoside building block greatly facilitates linker attachment. A high selectivity of the amine group in acylation can be expected, whereas the formation of 3'-O-alkyl derivatives can cause considerable problems in the preparation and purification of 2'-O-aminoalkyl nucleosides (2,10). In contrast to the reaction of carboxylic acids with 2'-amino-2'-deoxyuridine after solid phase oligonucleotide synthesis (17), the use of a succinic linker promises better conjugation yields due to the lower steric hindrance. Additionally, the use of a 4–6 carbon tether has been shown to be beneficial with regard to the properties of the resulting conjugates (13).
To test this strategy, we decided to prepare conjugates of Oblimersen, an 18mer all-phosphorothioate targeted against the first six codons of the open reading frame of the apoptosis inhibiting protein Bcl-2 (18,19). It is one of the most progressed and prolific new antisense drugs. Primary data from its phase III clinical trial against malignant melanoma have recently been published.
MATERIALS AND METHODS
Reagents were used in standard quality for synthesis and were purchased from Fluka or Aldrich. Melting points were measured on a Kofler melting point apparatus and are uncorrected. Anhydrous solvents were obtained as follows: tetrahydrofuran and toluene were refluxed on sodium and then distilled; methanol was heated over magnesium methoxide and then distilled; pyridine was refluxed on potassium hydroxide and distilled; and dimethyl formamide (DMF) was stored over a molecular sieve (4 ?). Purified water was obtained from a Milli-Q-apparatus. NMR spectra were recorded on a Bruker Avance 200 MHz or a Varian Unity 300 MHz machine. Shifts are reported relative to the solvent peak ; coupling constants are in Hz. Thin-layer chromatography (TLC) was performed using silica gel 60-F254 pre-coated aluminium plates from Merck. Column chromatography was performed with Merck silica gel 60. Elemental analyses were done by J. Theiner (Mikroanalytisches Laboratorium, University of Vienna).
Succinic acid monobenzyl ester (3)
Succinic acid anhydride (2 g, 20.0 mmol) was dissolved in 10 ml of dried pyridine, and benzyl alcohol (2.2 g, 20.3 mmol) and 4-dimethylaminopyridine (10 mg, 0.08 mmol) were added. The mixture was heated to 50°C and stirred for 24 h. Pyridine was removed in vacuo; the residue was dissolved in dichloromethane and extracted with sodium bicarbonate solution. The aqueous phase was washed with dichloromethane and then acidified with 2 M HCl until a pH of 4 was reached, setting free a white precipitation. Extraction with ethyl acetate and evaporation of the solvent afforded succinic acid mono benzyl ester (3; 3.42 g, 82%).
Fp: 60°C. 1H-NMR (CDCl3, 200 MHz): = 9.56 (bs, 1H, COOH), 7.26 (bs, 5H, Ar-H), 5.15 (s, 2H, PhCH2), 2.70 (s, 4H, CH2-CH2). 13C-NMR (CDCl3, 50 MHz): = 178.20 (COOH), 171.99 (COOR), 135.60 (ArC-1), 128.52 (ArC-3,5), 128.24 (ArC-4), 128.14 (ArC-2,6), 66.60 (PhCH2), 28.87 and 28.82 (CH2–CH2). C11H12O4 calculated: C 63.45 H 5.81, found: C 63.21 H 5.85.
2'-(4-Benzylsuccinyl)amido-2'-deoxy-5'-(4,4'-dimethoxytriphenyl)methyluridine (4)
2'-Amino-2'-deoxy-5'-(4,4'-dimethoxytriphenyl)methyluridine (1) (20) (493 mg, 0.90 mmol) was dissolved under an argon atmosphere in 10 ml of dried tetrahydrofuran. 1-Hydroxy-1H-benzotriazol (138 mg, 1.02 mmol), N,N'-diisopropylcarbodiimide (128 mg, 1.01 mmol), succinic acid mono benzyl ester (3; 190 mg, 0.91 mmol) and 4-dimethylaminopyridine (10 mg, 0.08 mmol) were added subsequently and the reaction mixture was stirred at room temperature for 18 h. The solvent was evaporated in vacuo and the residue was chromatographed (dichloromethane:methanol 100:2). Fractions containing the product were dried, affording N-acylated product 4 (644 mg, 97%) as a white powder.
Fp: 149–150°C, 1H-NMR (CDCl3, 200 MHz): = 7.64 (d, J56 = 8.2, 1H, H-6), 7.42 – 7.25 (m, 14 H, Ar-H), 6.83 (d, J = 8.8, 4H, Ar-H), 6.20 (d, J2'1' = 8.6, 1H, H-1'), 5.40 (d, J65 = 8.1, 1H, H-5), 5.07 (s, 2H, PhCH2), 4.66 (dd, J = 8.3, H-2'), 4.49 (d, J = 5.1, 1H, H-3'), 4.21 (bs, 1H, H-4'), 3.75 (s, 3H, OCH3), 3.39 (bs, 2H, H-5'a, H-5'b), 2.90–2.45 (m, 4H, CH2–CH2). 13C-NMR (CDCl3, 50 MHz): = 173.68 and 172.93 (COOR, CONH), 162.93 (C-4), 158.66 (ArC-4, methoxyphenyl), 151.24 (C-2), 144.01 (ArC-1, phenyl), 139.93 (C-6), 135.39, 135.21 and 135.00 (ArC-1, methoxyphenyl and benzyl), 130.14 (ArC-2,6, methoxyphenyl), 128.58 (ArC-3,5, benzyl), 128.35 (Ar-C), 128.22 (ArC-2,6, benzyl), 128.04 (ArC), 127.12 (ArC-4, phenyl), 113.24 (ArC-3,5, methoxyphenyl), 103.10 (C-5), 87.10 (CPh3), 85.61 and 85.32 (C-1', C-4'), 71.45 (C-3'), 66.93 (PhCH2), 63.78 (C-5'), 57.80 (C-2'), 55.23 (OCH3), 30.69 and 29.65 (CH2–CH2). C41H41N3O10 calculated for 4% H2O: C 64.25 H 5.82 N 5.48, found: C 64.27 H 6.04 N 5.41.
2'-(Benzylsuccinyl)amido-2'-deoxy-5'-(4,4'-dimethoxytriphenyl)methyl-3'-O-succinyluridine (5)
N-Benzylsuccinyl derivative 4 (250 mg, 0.34 mmol) was dissolved in 7 ml of dried pyridine and one drop of N-methylmorpholine was added. After addition of succinyl anhydride (70 mg, 0.70 mmol), the mixture was stirred under exclusion of moisture at 30°C for 5 days. The solvent was removed in vacuo and the residue was co-evaporated three times with dichloromethane. A 10 ml aliquot of dichloromethane was added and the precipitated succinic acid was filtered off. The filtrate was purified by column chromatography (dichlormethane:methanol 100:2) affording hemisuccinate 5 (200 mg, 70%) as a white solid.
1H-NMR (CDCl3, 200 MHz): = 10.46 (bs, 1H, COOH), 7.65 (d, J56 = 8.2, 1H, H-6), 7.48–7.22 (m, 14 H, Ar-H), 6.83 (d, J = 8.7, Ar-H), 6.09 (d, J2'1' = 9.0, 1H, H-1'), 5.46 (d, J2'3' = 5.2, 1H, H-3'), 5.39 (d, J65 = 8.1, 1H, H-5), 5.10 (m, 1H, H-2'), 5.04 (s, 2H, PhCH2), 4.18 (m, 1H, H-4'), 3.75 (s, 6H, OCH3), 3.48 (dd, J5'b5'a = 9.2, 1H, H-5'a), 3.39 (dd, J5'a5'b = 9.1, 1H, H-5'b), 2.64–2.55 (m, 8H, CH2–CH2). 13C-NMR (CDCl3, 50 MHz): = 176.48 (COOH), 173.24 and 172.63 , 170.65 , 163.80 (C-4), 158.72 (ArC-4, methoxyphenyl), 151.30 (C-2), 143.90 (ArC-1, phenyl), 140.16 (C-6), 135.72, 135.04 and 134.89 (ArC-1, methoxyphenyl and phenyl), 130.11 (ArC-2,6, methoxyphenyl), 128.48 (ArC-3,5, benzyl), 128.15, 128.11, 128.03 and 127.75 (ArC), 127.18 (ArC-4, phenyl), 113.38 (ArC-3,5, methoxyphenyl), 102.97 (C-5), 87.45 (CPh3), 85.11 (C-1'), 82.77 (C-4'), 74.14 (C-3'), 66.53 (PhCH2), 63.57 (C-5'), 55.22 (OCH3), 54.08 (C-2'), 30.41, 29.66 and 29.26 (CH2–CH2). C45H45N3O13 calculated: C 64.66, H 5.43, N 5.03; found: C 64.37, H 5.67, N 4.78.
Loading of solid support
TentaGel-NH2 (230 mg, 60 μmol free NH2 groups), purchased from Fluka, was put into a 50 ml flask and suspended in 3 ml of dried DMF for swelling. Succinyl derivative 5 (100 mg, 120 μmol) was separately dissolved in 1 ml of DMF, and 1-hydroxy-1H-benzotriazol (17 mg, 126 μmol) was added. This solution was added to the flask with TentaGel, as were N,N'-diisopropylcarbodiimide (16 mg, 127 μmol) and 4,4'-dimethylaminopyridine (1 mg, 8 μmol). After 16 h of stirring at room temperature, acetic anhydride (12 mg, 118 μmol) in 3 ml of pyridine was added for capping of free amino groups, and stirring was continued for 1 h. TentaGel was filtered off and washed with DMF, methanol and dichloromethane three times each. The resin was dried in vacuo and the loading was quantified by letting a small sample react with 3% trichloroacetic acid and determining the trityl cation concentration by UV spectroscopy (498 nm, = 71 700).
Oligonucleotide synthesis
Standard phosphoramidites and tetrazole solution were obtained from Carl Roth GmbH+Co (Karlsruhe, Germany). Oligonucleotide synthesis was carried out on an ABI 392 Applied Biosystems DNA/RNA-Synthesizer in standard 1 μmol scale in DMT-on mode. Coupling of the first nucleotide to solid-supported modified nucleotide was prolonged to 2 min. Acetonitrile was heated over potassium carbonate and distilled, followed by heating over calcium hydride and distillation. Dichloromethane was dried over phosphorus pentoxide and distilled. Other reagents were obtained in analytical quality and purified before use. Sulfurization was done with tetraethyldiuram disulfide, which was purchased from Sigma and recrystallized (21).
Palladium nanoparticles
Preparation of palladium nanoparticles was done following a procedure by Hirai et al. (22), but with an increased amount of poly-N-vinyl-pyrrolidon (PVP). Palladium(II)chloride (6 mg, 0.034 mmol) was dissolved in 25 ml of methanol, and PVP (mol. wt 25 000 Da, 187 mg, 1.681 mmol of monomer) was separately dissolved in 20 ml of methanol. The solution of PVP was added to the palladium and the mixture was heated to 50°C under exclusion of moisture. After 30 min, a solution of sodium hydroxide (2.65 mg, 0.067 mmol) in 5 ml of methanol was added dropwise. Heating was continued for 5 min and, after cooling, the dark brown suspension was transferred to a dark bottle, where it was stored under an argon atmosphere.
Preparation of oligonucleotide conjugates (7,9)
The synthesis column was removed from the synthesizer and opened. The resin-bound oligonucleotide was transferred to a 5 ml reaction tube of a Quest 210 manual synthesizer. Cleavage of the benzyl protecting group was achieved by phase transfer hydrogenation with PVP-stabilized palladium nanoparticles (0.7 ml) and 1,4-cyclohexadiene (0.05 ml) as hydrogen donor. Cyclohexadiene was added to the PVP-stabilized suspension of palladium, and the resulting solution was microfiltrated (0.22 μm) to ensure no precipitated palladium was applied to the resin. The reaction mixture was stirred at room temperature for a total time of 16 h.
The resin was washed three times with methanol and three times with DMF. Coupling was done by adding a solution of 1-hydroxy-1H-benzotriazol (0.4 mg, 3.0 μmol), N,N'-diisopropylcarbodiimide (0.4 mg, 3.2 μmol) and 5-aminofluorescein (1.1 mg, 3.2 μmol) or putrescine (0.3 mg, 3.4 μmol) in 1 ml of DMF. The mixture was stirred for 18 h and the solution was again filtered off. The resin was washed three times with DMF and three times with water. For cleavage and deprotection of oligonucleotides, 2 ml of concentrated ammonia was added and the mixture was heated to 55°C for 18 h. The solution was filtered and the solvent was evaporated in vacuo.
Purification and analyses of antisense oligonucleotides
Oligonucleotide concentrations were determined by measuring the OD260 in a Hitachi U3000 spectrophotometer. Molar extinction coefficients were calculated as the sum of nucleotides (A, 15 400; G, 11 700; C, 7300, T, 8800; and U, 9950).
Initial purification of oligonucleotides was done with Poly-Pak? columns obtained from Carl Roth GmbH according to the manufacturer. The typical yield after purification was 250 nmol. Analytical HPLC was performed on a Nucleosil? CC 250/4 100-5 C18 column and a LiChroSphere? 100 RP 18 column with the following gradient system: A, 0.1 M triethylammonium acetate in water; B, 0.1 M triethylammonium acetate in 80% acetonitrile, linear gradient 10–40% B in 0–30 min, flow rate 1 ml/min. If Poly-Pak? purification was not satisfactory, preparative HPLC was performed on a LiChrosphere? 100 RP-18 using the same gradient.
Mass spectrometric analysis was done on a Kratos seq MALDI mass spectrometer. Sample preparation was done following the method described by Pieles et al. (23). A 1 μl aliquot of sample (100 pmol/μl in water) was briefly vortexed with 10 μl of a 0.5 M solution of 2,4,6-trihydroxyacetophenone in ethanol and 5 μl of a 0.1 M solution of di-ammonium hydrogen citrate. A 1 μl aliquot of this mixture was spotted on the target and allowed to air dry. The mass spectrometer was run in the negative ion and reflectron mode, and spectra were obtained by summing 50–100 single laser pulses.
Circular dichroism (CD) spectrometry was performed on a Jasco J-810 spectropolarimeter equipped with a Neslab RTE 7 thermostatic unit. Oligonucleotides were diluted to a 9 μM solution in 0.15 M NaCl and 0.01 M Tris–HCl (pH 7.0) buffer. CD spectra were collected from 320 to 210 nm using a quartz cuvette (Hellma 100-QS) with a pathlength of 1 mm. Duplexes were measured after heating the equimolar mixture of complementary strands to 50°C for 10 min and slow cooling to room temperature.
Melting temperatures (Tms) of the duplexes were determined in the Jasco J-810 by slowly heating from 30 to 90°C (50°C/h) and recording CD at 248 nm or OD at 260 nm as a function of temperature. Melting curves showed cooperative form, and Tms were obtained from the maxima of the first derivative plots. Temperatures given are those of the cuvette holder.
Treatment of cells with oligonucleotides and conjugate
Human melanoma 607B cells were grown in 6-well plates. After 24 and 48 h, cells were incubated for 4 h at 37°C with oligonucleotides or oligonucleotide–fluorescein conjugate complexed with lipofectin at a concentration of 100 nM in Opti-MEM medium. After a total of 72 h, cells were harvested.
Western blot analysis
After incubation with oligonucleotides and oligonucleotide conjugate, respectively, whole-cell extracts of cultured cells were prepared as described previously (24). Briefly, melanoma cells were washed with phosphate-buffered saline (PBS) and then scraped into 80 μl of lysis buffer . Cell lysates were freeze–thawed in liquid nitrogen three times and centrifuged at 15 000 r.p.m. for 20 min at 4°C to remove cell debris. Protein concentrations were determined by Bradford protein assay.
Aliquots of cell extracts containing 15 μg of protein were resolved by 12% SDS–PAGE and blotted to a PVDF membrane (ImmobilonTM P). The membrane was blocked for 1 h in 0.2% I-block (Tropix, Bedford, MA) in PBS and cut in half. The top part was incubated overnight at 4°C with actin antibody (2.0 μg/ml in I-block solution, Sigma, St Louis, MO) and the bottom part under the same conditions with monoclonal antibody recognizing Bcl-2 (2.5 μg/ml in I-block solution, Zymed Laboratories, Inc., San Francisco, CA). Secondary antibody incubation was carried out using Bcl-2 goat anti-mouse and actin goat anti-rabbit conjugated to alkaline phosphatase (2 μl in 10 ml of I-block solution, Tropix). The membranes were subsequently washed with I-block and assay buffer. For visualization, chemiluminescence substrate CSPD (0.25 M, 5 min, Tropix) was used. Expression levels were quantified by densitometry of autoradiograms with a Herolab EASY RH densitometer (Herolab, Wiesloch, Germany) and the EASY Win32 software. Signal strengths of Bcl-2 signals were normalized to actin, and the ratios between Bcl-2 protein in antisense oligonucleotide-treated extracts and tumour control extracts were calculated.
RESULTS AND DISCUSSION
2'-Amino-2'-deoxy-5'-dimethoxytrityluridine (1) was synthesized in good yield following the four-step procedure published by McGee et al. (20). We decided to use a succinyl linker for post-synthetic attachment of various ligands by an amide bond. This 4-carbon linker has a suitable length for minimizing the steric hindrance that can be seen for direct acylation of the 2'-amino group (15,25) and is advantageous for introduction of ligands in the minor groove. For solid phase synthesis, the carboxyl group had to be protected. A suitable protecting group had to be stable at all stages of the oligonucleotide synthesis cycle, but also labile enough to be cleaved selectively after oligonucleotide chain elongation without loss of base protecting groups. The benzyl ester proved to be a suitable masking group, since it is stable under acid and slightly basic conditions, but is readily cleaved by hydrogenolysis (26). Whereas using hydrogen gas leads to significant reduction of double bonds of pyrimidine bases, the use of cyclohexadiene as a hydrogen donor in a phase transfer catalysis shows no such side reactions (27). Bajwa reported the selctive cleavage of benzyl esters and N-benzyl groups under these conditions (28). 2'-Amino-2'-deoxy-5'-dimethoxytrityluridine (1) reacted in a highly specific way with succinic acid anhydride to give only N-acylated product 2. As expected, no 3'-acylated product was formed. As benzylation of the free carboxyl group proceeded only in poor yield, we prepared the monobenzyl ester of succinic acid (3) and coupled it to the 2'-amino group using N,N'-diisopropylcarbodiimide, 1-hydroxy-1H-benzotriazol and 4-dimethylaminopyridine in DMF. In accordance with our expectations, again only N-acylated product was isolated in excellent yield (Scheme 1). The structures and conformations of these products were verified by 1H–1H and 1H–13C correlation NMR experiments.
Scheme 1. Synthesis of nucleoside derivative for attachment to solid phase. (a) Succinic acid anhydride, pyr, r.t., 95%. (b) Succinic acid monobenzyl ester 3, N,N'-diisopropylcarbodiimide, 1-hydroxy-1H-benzotriazol, 4-dimethylaminopyridine, tetrahydrofuran, 97%. (c) Succinic acid anhydride, pyr, 30°C, 70%.
NMR spectral data suggest the adoptation of a C2'-endo configuration of these N-acyl derivatives of uridine. The chemical shifts and especially the coupling constants of carbons of the ribose moiety are comparable with those found by other authors for 2'-amido-2'-deoxyadenosine derivatives (29) and 2'-acetamido-2'-deoxyuridine and guanosine, respectively (30) (Table 1). As has been shown by modeling experiments (29), coupling constants for H1'–H2' (8.0 for 2 and 8.6 for 4) and H2'–H3' (5.6 and 5.1, respectively) are a clear indication of the C2'-endo configuration. In contrast to 2'-O-substituted derivatives, 2'-N-acyl nucleosides, the substituents having a low electronegativity, tend towards a B-DNA configuration, which is typical for DNA strands, whereas the ribose of RNA is fixed in the C3'-endo pucker and, consequently, an A-type DNA helix. Although Hendrix et al. (30) concluded that the preferred B-DNA conformation was responsible for the lower melting temperature for oligonucleotides with incorporated 2'-acetamido-2'-deoxyuridine (2.7–7.2°C), the locked A-DNA configuration of 2'-O-alkyl-modified oligonucleotides was found to be a rather disadvantageous property for antisense agents despite the high stability against enzymatic degradation, since such derivatives show no activation of RNase H, nowadays believed to be one of the most important aspects of antisense activity.
Table 1. Comparison of 1H-NMR data of some 2'-amido-2'-deoxynucleosides (21,22)
Based on these data, 4 seems to be a suitable building block for the addition of ligands to oligonucleotide. Its preparation is straightforward and the conformation of the ribose moiety seems to be advantageous for antisense agents.
For the synthesis of oligonucleotide conjugates, the modified nucleoside should be incorporated at the 3' end of the oligonucleotide. Thus, it had to be attached directly to the solid phase. Controlled pore glass (CPG) is the ideal resin for oligonucleotide synthesis, but has disadvantages regarding reactions in polar solvents such as DMF, used in peptide chemistry (31). Therefore, we decided to use TentaGel, which is suitable for both oligonucleotide (32) and peptide chemistry (33). The addition of 4 to TentaGel with an attached succinyl linker gave only very poor loading yields (<10 μmol/g), obviously due to the steric bulk of the 2'-substituent. Thus, we derivatized 4 by reaction with succinic anhydride to give hemisuccinate 5 (Scheme 1). As TLC controls revealed, the reaction with an excess of succinic acid anhydride progressed only slowly, further confirmation of the great steric hindrance of the 3' position. After reaction times of several days, good yields were achieved. Coupling of hemisuccinate 5 to TentaGel-NH2 resulted in a much improved loading of 190 μmol/g.
Solid phase-coupled derivative 5 could be employed in standard oligonucleotide synthesis cycles. The coupling time of the first phosphoramidite was prolonged to improve yields. ASWYs (average stepwise yields) as determined by trityl cation assay were routinely >95%.
The 18mer oligonucleotide bearing the sequence of Oblimersen (5'-TCTCCCAGCGTGCGCCAU-3'), a complete phosphorothioate backbone and with incorporation of the modified uridine nucleoside at the 3' end, was synthesized. Direct cleavage and deprotection by reaction of ammonia gave the product with a free carboxyl group (6) as determined by MALDI mass spectrometry. Treatment of the modified monomer with 10% palladium on carbon and cyclohexadiene in a phase transfer catalysis cleaved the benzyl ester in liquid phase. Reaction of resin-bound oligonucleotide with palladium on charcoal proved to be difficult. The resin turned black after a short time and, even after repeated washing, a small fraction of activated carbon remained on the solid support. After treatment with concentrated ammonia, virtually no product was found. The oligonucleotide was presumably absorbed at the charcoal still present at the resin surface. Therefore, other possibilities for catalytic hydrogenolysis avoiding the use of activated carbon were researched. Different palladium(0)complexes have been used in solid phase chemistry (33,34). Recently, palladium nanoparticles, stabilized by PVP (21) to prevent aggregation, were used for the hydrogenolysis with hydrogen gas of benzyl ether groups of solid-supported sugar derivatives (35).
The phase transfer reaction of resin-bound oligonucleotide with palladium nanoparticles and 1,4-cyclohexadiene as hydrogen donor selectively cleaved the benzyl ester. After 16 h, the liquid was filtered off and a UV spectrum was recorded. Overlay with the spectrum of unreacted solution of palladium nanoparticles with cyclohexadiene showed a new band at 280 nm, which indicated the formation of toluene, the product of hydrogenolysis. In order to determine the optimum reaction time, UV spectra of the reaction solution filtrated at 1, 2 and 4 h were recorded. The obtained data indicated that debenzylation was not yet quantitative at that time. Therefore, a reaction time of 16 h was used in further syntheses. Although inhibition of reaction due to poisoning of the catalyst by sulfur was feared, no differences were observed in reactivity of phosphorodiesters and phosphorothioates. Literature reports on sulfur poisoning of catalytic phase transfer hydrogenation are conflicting (27,36), but it has been established that palladium catalysts can, in principle, work despite the presence of sulfur (37).
To verify that no alterations of phosphorothioate backbone or reduction of pyrimidine bases occurs, oligonucleotide 6 was treated with 1,4-cyclohexadiene and palladium nanoparticles for 5 days. Repeated HPLC analyses during this time proved that no reaction took place as the chromatograms remained unchanged.
For the first coupling of a ligand, fluorescein was chosen. It is a widely used reporter group and successful attachment can very easily be verified because of its fluorescent activity. Well documented peptide coupling reagents were employed for the reaction (38,39). A solution of N,N'-diisopropylcarbodiimide, 1-hydroxy-1H-benzotriazol and 5-aminofluorescein in DMF was added to the resin and stirred for 18 h to ensure completion of the reaction. The solution was filtered off and, on visual inspection, it was clearly brighter than unreacted solution, indicating successful attachment of fluorescein to the oligonucleotide. Cleavage and deprotection afforded a dark yellow solution that showed fluorescence. A control experiment where the carboxyl deprotection step was omitted did not afford a coloured or fluorescent product. Thus it was shown that no unspecific coupling of fluorescein occurred, proving the structure of conjugate 7 shown in Figure 2.
Figure 2. Structures of oligonucleotide conjugates 7 and 9.
Purification was achieved by Poly-Pak? columns or preparative HPLC. Retention times can be seen in Table 2. Due to the lipophilic character, the retention time of fluorescein conjugate 7 was significantly longer than that of modified oligonucleotide 6. No peak of the unreacted oligonucleotide 6 was found in the chromatogram (Fig. 3), indicating a conjugation yield of >95%.
Table 2. HPLC retention times, MALDI mass and melting temperature (Tm)
Figure 3. HPL-chromatogram of conjugate 7 (right) after Poly-Pak? purification; injected sample 10 μg.
The successful conjugation of fluorescein confirms the selective benzyl ester cleavage in nearly quantitative yield on the solid phase and the favorable steric properties of modified nucleoside building block 4. The reaction of amines and carboxyl groups is a well researched procedure and has been used extensively in oligonucleotide chemistry (6,9). Hence, our method can be used for conjugation of various ligands bearing amino functionalites to oligonucleotides.
To further prove the generality of the method, we chose the polyamine putrescine as ligand. Synthesis following the described protocol using unprotected putrescine as ligand gave the polyamine–oligonucleotide conjugate 9 (Fig. 2) in good yield and excellent purity.
CD spectra of unmodified phosphodiester (8), unmodified phosphorothioate (Oblimersen), unconjugated oligonucleotide 6 and the conjugate with fluorescein (7) were recorded. All experiments were performed in a buffer containing 0.1 M Tris–HCl (pH 7.0) and 0.15 M NaCl. The phosphodiester exhibited clear characteristics of a B-DNA type helix. Modified oligonucleotides showed essentially the same spectra, but with reduced band intensities, especially of the minimum at 248 nm (Fig. 4).
Figure 4. Circular dichroism of 5'-TCTCCCAGCGTGCGCCA(T or U)-3'. Phosphodiester 8 (squares), phosphorothiaote Oblimersen (rhombi), modified phosphorothioate 6 (triangles) and conjugate 7 (circles).
Duplexes of phosphorothioate backbone oligonucleotides with complementary DNA showed nearly identical CD spectra. They are shown in Figure 5. Only small differences were observed regarding oligonucleotide–fluorescein conjugate 7, which gave slightly increased intensities of the positive bands at 278 and 218 nm and lower intensity of the negative band at 248 nm. In comparison with the duplex of 8, the thioates (6, 7 and Oblimersen) all had lower intensities in all bands, an observation that is in accordance with published data about other oligonucleotide sequences (40).
Figure 5. Circular dichroism of duplexes of 8 (squares), Oblimersen (rhombi), 6 (triangles) and 7 (circles) with phosphodiester sense strand.
Experiments determining the melting temperatures of duplex oligonucleotides all gave cooperative curves (Fig. 6). Melting temperatures shown in Table 2 are the average of four different determinations for each substance. The transition temperature of Oblimersen was 9.2°C (0.54°C per modification) lower than that of phosphorodiester 8, consistent with known data (14,41). The decrease in melting temperature of 6 was 4°C compared with Oblimersen, but that of the conjugate 7 was only 1°C. Polyamine conjugate 9 gave a reduction in transition temperature of 2°C. A substitution of 2'-deoxyuridine for thymidine accounts for an average reduction of melting temperature of 0.5°C, as was shown using a considerable number of different sequences (11). Presumably, the additional negative charge of the carboxyl group, located in the minor groove, considerably decreases duplex stability by means of electrostatic repulsion, whereas a conjugation removes this negative charge and significantly increases hybridization affinity in comparison with the derivative with a free carboxyl group despite its greater steric hindrance caused by the ligands. The aromatic fluorescein structure may lead to a stabilization of duplex structure by contributing to the base stacking of nucleobases. This presumption is backed by the slightly increased intensity of the CD band at 278 nm, reflecting an increase in base stacking.
Figure 6. Melting profiles of 6 (squares) and 7 (triangles).
Cell culture tests were performed in human melanoma cell line 607B with an oligonucleotide concentration of 100 nM in the presence of lipofectin as uptake enhancing transfection reagent (23). Cell lysates were analysed by western blot. Downregulation of Bcl-2, normalized against changes in the actin band, was similar for conjugate 7 and control Oblimersen, while modified oligonucleotide 6 exhibited a weaker effect. Treatment of cells with fluorescein–oligonucleotide conjugate 7 gave a reduction of Bcl-2 of 79% in comparison with melanoma cells treated with lipofectin only, while Oblimersen reduced the expression of target protein by 85%. Modified oligonucleotide 6 showed the weakest effect of the three tested substances, with a reduction of Bcl-2 level of 69%. Control reverse and mismatched oligonucleotides showed an inhibition of target protein by only up to 17%, establishing the true antisense effect of the tested products (23). These results confirm the findings of the melting temperature experiments. The conjugation of ligands in the minor groove by a succinic linker tethered to the 2' position of 2'-aminouridine leads only to a minor depletion in hybridization affinity and thus in antisense efficacy.
In conclusion, the fluorescent conjugate 7 has excellent base-pairing properties and in vitro efficacy. The convenient location of the ligand in the minor groove at the 3' end of the oligonucleotide results in very little disturbance of base pairing and consequently supression of gene expression. Our method for conjugation of ligands to oligonucleotides via a succinyl linker at the 2' position of 2'-amino-2'-deoxyuridine is selective, gives excellent yield and can easily be employed. It has the considerable advantage over previously published procedures that the building block used for attaching ligands at the 2' position can be prepared much more easily. A method for selectively cleaving benzyl esters in oligonucleotide solid phase chemistry has been developed by the use of palladium nanoparticles in combination with cyclohexadiene as hydrogen donor. It has been shown that this reaction succeeds as well for phosphorothioates. Studies on the cellular uptake of fluorescent conjugate 7 as well as the preparation and characterization of further oligonucleotide conjugates will be published at a later date.
REFERENCES
Tamm,I., D?rken,B. and Hartmann,G. (2001) Antisense therapy in oncology: new hope for an old idea? Lancet, 358, 489–497.
Urban,E. and Noe,C.R. (2003) Structural modifications of antisense oligonucleotides. Farmaco, 58, 243–258.
Noe,C.R. and Kaufhold,L. (2000) Chemistry of antisense oligonucleotides. In Gualteri,F. (ed.), New Trends in Synthetic Medicinal Chemistry. Wiley-VCH, Weinheim, pp. 261–346.
Manoharan,M., Tivel,K.L. and Cook,P.D. (1995) Lipidic nucleic acids. Tetrahedron Lett., 36, 3651–3654.
Manoharan,M. (1993) Designer antisense oligonucleotides: conjugation chemistry and functionality placement. In Crooke,S.T. and Lebleu,B. (eds), Antisense Research and Applications. CRC Press, Boca Raton, FL, pp. 303–349.
Tung,C.-H. and Stein,S. (2000) Preparation and applications of peptide–oligonucleotide conjugates. Bioconjug. Chem., 11, 605–618.
Beaucage,S.L. and Iyer,R.P. (1993) The functionalization of oligonucleotides via phosphoramitide derivatives. Tetrahedron, 49, 1925–1963.
Manoharan,M. (2001) Oligonucleotide conjugates in antisense technology. In Crooke,S.T. (ed.), Antisense Drug Technology. Marcel Dekker, Inc., New York, NY, pp. 391–471.
Virta,P., Katajisto,J., Niittym?ki,T. and L?nnberg,H. (2003) Solid-supported synthesis of oligomeric bioconjugates. Tetrahedron, 59, 5137–5174.
Werner,D., Brunar,H. and Noe,C.R. (1998) Investigations on the influence of 2'-O-alkyl modifications on the base pairing properties of oligonucleotides. Pharm. Acta Helv., 73, 3–10.
Cotten,M., Oberhauser,B., Brunar,H., Holzner,A., Issakides,G., Noe,C.R., Schaffner,G., Wagner,E. and Birnstiel,M.L. (1991) 2'-O-methyl, 2'-O-ethyl oligoribonucleotides and phosphorothioate oligodeoxyribonucleotides as inhibitors of the in vitro U7 snRNP-dependent mRNA processing event. Nucleic Acids Res., 19, 2629–2635.
Freier,S.M. and Altmann,K.-H. (1997) The ups and downs of nucleic acid duplex stability: structure–stability studies on chemically-modified DNA:RNA duplexes. Nucleic Acids Res., 25, 4429–4443.
Manoharan,M. (1999) 2'-Carbohydrate modifications in antisense oligonucleotide therapy: importance of conformation, configuration and conjugation. Biochim. Biophys Acta, 1489, 117–130.
Manoharan,M., Prakash,T.P., Barber-Peoch,I., Bhat,B., Vasquez,G., Ross,B.S. and Cook,P.D. (1999) N-(2-Cyanoethoxycarbonyloxy) succinimide: a new reagent for protection of amino groups in oligonucleotides. J. Org. Chem., 64, 6468–6472.
Hwang,J.-T. and Greenberg,M.M. (2001) Synthesis of 2'-modified oligodeoxynucleotides via on-column conjugation. J. Org. Chem., 66, 363–369.
Zubin,E.M., Romanova,E.A. and Oretskaya,T.S. (2002) Modern methods for the synthesis of peptide–oligonucleotide conjugates. Russ. Chem. Rev., 71, 239–264.
Hwang,J.-T. and Greenberg,M.M. (1999) Synthesis of modified oligodeoxyribonucleotides on a solid-phase support via derivatization of a selectively revealed 2'-amino-2'-deoxyuridine. Org. Lett., 1, 2021–2024.
Dias,N. and Stein,C.A. (2002) Potential roles of antisense oligonucleotides in cancer therapy. The example of Bcl-2 antisense oligonucleotides. Eur. J. Pharm. Biopharm., 54, 263–269.
Jansen,B., Wacheck,V., Heere-Ress,E., Schlagbauer-Wadl,H., Hoeller,C., Lucas,T., Hoermann,M., Hollenstein,U., Wolff,K. and Pehamberger,H. (2000) Chemosensitisation of malignant melanoma by BCL2 antisense therapy. Lancet, 356, 1728–1733.
McGee,D.P.C., Vaughn-Settle,A., Vargeese,C. and Zhai,Y. (1996) 2'-Amino-2'-deoxyuridine via an intramolecular cyclization of a trichloroacetimidate. J. Org. Chem., 61, 781–785.
Vu,H. and Hirschbein,B.L. (1991) Internucleotide phosphite sulfurization with tetraethylthiuram disulfide. Phosphorothioate oligonucleotide synthesis via phosphoramitide chemistry. Tetrahedron Lett., 32, 3005–3008.
Hirai,H., Yakura,N., Seta,Y. and Hodoshima,S. (1998) Characterization of palladium nanoparticles protected with polymer as hydrogenation catalyst. React. Funct. Polymers, 37, 121–131.
Pieles,U., Zürcher,W., Sch?r,M. and Moser,H.E. (1993) Matrix-assisted laser desorption ionization time-of-flight mass spectrometry: a powerful tool for the mass and sequence analysis of natural and modified oligonucleotides. Nucleic Acids Res., 21, 3191–3196.
Jansen,B., Schlagbauer-Wadl,H., Brown,B.D., Bryan,R.N., van Elsas,A., Müller,M., Wolff,K., Eichler,H.-G. and Pehamberger,H. (1998) bcl-2 antisense therapy chemosensitizes human melanoma in SCID mice. Nature Med., 4, 232–234.
Sharma,R.A., Bobek,M. and Bloch,A. (1975) Preparation and biological activity of some aminoacyl and peptidyl derivatives of 2'-amino-2'-deoxyuridine. J. Med. Chem., 18, 955–957.
Greene,T.W. and Wuts,P.G.M. (1991) Protection for the carboxyl group. In Greene,T.W. and Wuts,P.G.M. (eds), Protective Groups in Organic Synthesis. John Wiley & Sons, New York, NY, pp. 224–276.
Watkins,B.E., Kiely,J.S. and Rapoport,H. (1982) Synthesis of oligodeoxyribonucleotides using n-benzyloxycarbonyl-blocked nucleosides. J. Am. Chem. Soc., 104, 5702–5708.
Bajwa,J.S. (1992) Chemoselective deprotection of benzyl esters in the presence of benzyl ethers, benzyloxymethyl ethers and N-benzyl groups by catalytic transfer hydrogenation. Tetrahedron Lett., 33, 2299–2302.
Van Calenbergh,S., Van Den Eeckhout,E., Herdewijn,P., De Bruyn,A., Verlinde,C., Hol,W., Callens,M., Van Aerschot,A. and Rozenski,J. (1994) Synthesis and conformational analysis of 2'-deoxy-2'-(3-methoxybenzamido)adenosine, a rational-designed inhibitor of trypanosomal glyceraldehyde phosphate dehydrogenase (GAPDH). Helv. Chim. Acta, 77, 631–644.
Hendrix,C., Devreese,B., Rozenski,J., Van Aerschot,A., De Bruyn,A., Van Beeumen,J. and Herdewijn,P. (1995) Incorporation of 2'-amido-nucleosides in oligodeoxynucleotides and oligoribonucleotides as a model for 2'-linked conjugates. Nucleic Acids Res., 23, 51–57.
Beaucage,S.L. and Iyer,R.P. (1992) Advances in the synthesis of oligonucleotides by the phosphoramidite approach. Tetrahedron, 48, 2223–2311.
Wright,P., Lloyd,D., Rapp,W. and Andrus,A. (1993) Large scale synthesis of oligonucleotides via phosphoramidite nucleosides and a high-loaded polystyrene support. Tetrahedron Lett., 34, 3373–3376.
de la Torre,B., Avino,A., Tarrason,G., Piulats,J., Albericio,F. and Eritja,R. (1994) Stepwise solid-phase synthesis of oligonucleotide–peptide hybrids. Tetrahedron Lett., 35, 2733–2736.
Ma,S., Duan,D. and Wang,Y. (2002) Palladium(0)-catalyzed coupling-cyclization reaction of polymer-supported aryl iodides with 1,2-allenyl carboxylic acids. Solid phase parallel synthesis of butenolides. J. Comb. Chem., 4, 239–247.
Kanie,O., Grotenberg,G. and Wong,C.-H. (2000) Removal of benzyl protecting groups from solid-supported compounds by hydrogenolysis using palladium nanoparticles. Angew. Chem. Int. Ed., 39, 4545–4547.
Anantharamaiah,G.M. and Sivanandaiah,K.M. (1977) Transfer hydrogenation; a convenient method for removal of some commonly used protecting groups in peptide synthesis. J. Chem. Soc. Perkin Trans. I, 5, 490–491.
Castro,F.J., Meyer,G. and Zampieri,G. (2002) Effects of sulfur poisoning on hydrogen desorption from palladium. J. Alloys Comp., 330–332, 612–616.
Lloyd-Williams,P., Albericio,F. and Giralt,E. (1993) Convergent solid phase peptde synthesis. Tetrahedron, 49, 11065–11133.
Henkel,B., Zhang,L. and Bayer,E. (1997) Investigations on solid-phase peptide synthesis in N-to-C direction (inverse synthesis). Liebigs Ann./Receuil, 2161–2168.
Clark,C.L., Cecil,P.K., Singh,D. and Gray,D.M. (1997) CD, absorption and thermodynamic analysis of repeating dinucleotide DNA, RNA and hybrid duplexes 12·12 and the influence of phosphorothioate substitution. Nucleic Acids Res., 25, 4098–4105.
Mier,W., Eritja,R., Mohammed,A., Haberkorn,U. and Eisenhut,M. (2000) Preparation and evaluation of tumor-targeting peptide–oligonucleotide conjugates. Bioconjug. Chem., 11, 855–860.(Johannes Winkler, Ernst Urban, Doris Los)
*To whom correspondence should be addressed. Tel: +431 4277 55103; Fax: +431 4277 9551; Email: christian.noe@univie.ac.at
ABSTRACT
Conjugation of ligands to antisense oligonucleotides is a promising approach for enhancing their effects. In this report, a new method for synthesizing oligonucleotide conjugates is described. 2'-Amino-2'-deoxy-5'-dimethoxytrityl-uridine was select ively acylated with a succinic acid linker at the 2' position. This compound was incorporated at the 3' end of an oligonucleotide corresponding to the sequence of Oblimersen. The carboxyl group was protected for oligonucleotide synthesis as a benzyl ester, which could be selectively cleaved at the solid phase by a catalytic phase transfer reaction using palladium nanoparticles as catalyst. An oligonucleotide–fluorescein conjugate was prepared by condensation of aminofluorescein. Circular dichroism spectroscopic experiments showed a B-DNA type structure. The melting temperature of the duplex was only slightly lower than that of Oblimersen. Biological activity measured by western blotting resulted in a Bcl-2 target downregulation nearly identical to that of control Oblimersen on human melanoma cells, proving that this method is attractive for the binding of ligands located in the minor groove.
INTRODUCTION
Several antisense oligonucleotides, designed for sequence-specific binding to a target mRNA, are now in various phases of clinical trials, many of them as anticancer agents (1). Initial attempts at increasing the efficacy, nuclease resistance and pharmacokinetic properties of antisense oligonucleotides have been focused on the chemical structure, leading to a plethora of modifications (2). Lately, an approach used quite often to specifically enhance the characteristics of antisense oligonucleotides is the attachment of various ligands (3).
Such attached functional moieties may be effectors, increasing the antisense effect; modulators, which can change their biophysical or pharmacodynamic properties; or detectors, which can be employed as diagnostic tools. For example, attachment of lipophilic molecules such as cholesterol can improve the cellular uptake of antisense oligonucleotides (4). Reporter groups such as biotin or fluorescent dyes such as fluorescein are used extensively in DNA-based diagnostics as well as for following cellular trafficking of antisense oligonucleotides (5). Much effort has recently been put into the field of oligonucleotide–peptide conjugates (6).
Due to the multivalent structure of oligonucleotides, there are many different synthetic possibilities for the attachment of ligands (Fig. 1). The majority of these conjugated compounds have been derivatized via a suitable linker at a terminal hydroxyl group of the oligonucleotide (7–9). Since the synthesis of oligonucleotides usually proceeds from 3' to 5', attachment to the 5'-hydroxy group can be achieved in a quite straightforward manner. Conjugation to the 3'-hydroxy function can be done by using an appropriate, orthogonally protected linker. Alternatively, incorporation of ligands can be achieved at nucleobases with a suitable linker attached to an amino group of adenosine, guanosine or cytosine, or the carbon 5 of pyrimidine bases. Attachment to the phosphorous backbone can be achieved by reaction of phosphate with an amino group of the peptide.
Figure 1. Possibilities for attachment of ligands to oligonucleotides.
In contrast, attachment of ligands to the carbohydrate moiety has been investigated less extensively. The increased binding affinity of short 2'-O-alkyl and 2'-O-ethylene glycol-modified oligonucleotides shows that this position is attractive for modifications (10–12) and for achieving structural diversity in the minor groove. Generally, conjugations at the 2' position of the carbohydrate moiety interfere less with base pairing than conjugations at backbone or nucleoside sites. By covalent binding of an aminoalkyl or a thioalkyl chain at the 2' position, post-synthetic derivatization has been made possible (13). Usually, a building block containing a suitably protected linker, mostly with an amino group, is incorporated before or during oligonucleotide synthesis and the amino group is either selectively unmasked or deprotected at the same time as the amino groups of the nucleobases. Linkers which have been used include aminoethyl (14), aminobutyl (15) and aminohexyl (14). To perform the conjugation reaction while the oligonucleotide is still resin bound and fully protected holds obvious advantages regarding selectivity. To date, few possibilities of achieving this have been published. For example, photolabile protecting groups can be used (15). After unmasking, peptides can be coupled to amines by reaction of the activated carboxyl groups. Other ligands can be attached to an amine via an isothiocyanate linkage (16).
Another approach consists of coupling ligands bearing a short alkyl linker directly to 2'-amino-2'-deoxyuridine (17). Due to the steric hindrance at the amino position of 2'-amino-2'-deoxyuridine, coupling yields were relatively low using this procedure (15).
Herein we describe a new strategy, in which the linker is bound directly to the starting nucleoside of solid phase synthesis. This approach has the advantage that possible negative properties of a ligand regarding hybridization affinity as well as problems during solid phase synthesis related to the steric hindrance are minimized. Compared with the dominantly employed 2'-O-aminoalkyl tether, the usage of 2'-amino-2'-deoxyuridine as the modified nucleoside building block greatly facilitates linker attachment. A high selectivity of the amine group in acylation can be expected, whereas the formation of 3'-O-alkyl derivatives can cause considerable problems in the preparation and purification of 2'-O-aminoalkyl nucleosides (2,10). In contrast to the reaction of carboxylic acids with 2'-amino-2'-deoxyuridine after solid phase oligonucleotide synthesis (17), the use of a succinic linker promises better conjugation yields due to the lower steric hindrance. Additionally, the use of a 4–6 carbon tether has been shown to be beneficial with regard to the properties of the resulting conjugates (13).
To test this strategy, we decided to prepare conjugates of Oblimersen, an 18mer all-phosphorothioate targeted against the first six codons of the open reading frame of the apoptosis inhibiting protein Bcl-2 (18,19). It is one of the most progressed and prolific new antisense drugs. Primary data from its phase III clinical trial against malignant melanoma have recently been published.
MATERIALS AND METHODS
Reagents were used in standard quality for synthesis and were purchased from Fluka or Aldrich. Melting points were measured on a Kofler melting point apparatus and are uncorrected. Anhydrous solvents were obtained as follows: tetrahydrofuran and toluene were refluxed on sodium and then distilled; methanol was heated over magnesium methoxide and then distilled; pyridine was refluxed on potassium hydroxide and distilled; and dimethyl formamide (DMF) was stored over a molecular sieve (4 ?). Purified water was obtained from a Milli-Q-apparatus. NMR spectra were recorded on a Bruker Avance 200 MHz or a Varian Unity 300 MHz machine. Shifts are reported relative to the solvent peak ; coupling constants are in Hz. Thin-layer chromatography (TLC) was performed using silica gel 60-F254 pre-coated aluminium plates from Merck. Column chromatography was performed with Merck silica gel 60. Elemental analyses were done by J. Theiner (Mikroanalytisches Laboratorium, University of Vienna).
Succinic acid monobenzyl ester (3)
Succinic acid anhydride (2 g, 20.0 mmol) was dissolved in 10 ml of dried pyridine, and benzyl alcohol (2.2 g, 20.3 mmol) and 4-dimethylaminopyridine (10 mg, 0.08 mmol) were added. The mixture was heated to 50°C and stirred for 24 h. Pyridine was removed in vacuo; the residue was dissolved in dichloromethane and extracted with sodium bicarbonate solution. The aqueous phase was washed with dichloromethane and then acidified with 2 M HCl until a pH of 4 was reached, setting free a white precipitation. Extraction with ethyl acetate and evaporation of the solvent afforded succinic acid mono benzyl ester (3; 3.42 g, 82%).
Fp: 60°C. 1H-NMR (CDCl3, 200 MHz): = 9.56 (bs, 1H, COOH), 7.26 (bs, 5H, Ar-H), 5.15 (s, 2H, PhCH2), 2.70 (s, 4H, CH2-CH2). 13C-NMR (CDCl3, 50 MHz): = 178.20 (COOH), 171.99 (COOR), 135.60 (ArC-1), 128.52 (ArC-3,5), 128.24 (ArC-4), 128.14 (ArC-2,6), 66.60 (PhCH2), 28.87 and 28.82 (CH2–CH2). C11H12O4 calculated: C 63.45 H 5.81, found: C 63.21 H 5.85.
2'-(4-Benzylsuccinyl)amido-2'-deoxy-5'-(4,4'-dimethoxytriphenyl)methyluridine (4)
2'-Amino-2'-deoxy-5'-(4,4'-dimethoxytriphenyl)methyluridine (1) (20) (493 mg, 0.90 mmol) was dissolved under an argon atmosphere in 10 ml of dried tetrahydrofuran. 1-Hydroxy-1H-benzotriazol (138 mg, 1.02 mmol), N,N'-diisopropylcarbodiimide (128 mg, 1.01 mmol), succinic acid mono benzyl ester (3; 190 mg, 0.91 mmol) and 4-dimethylaminopyridine (10 mg, 0.08 mmol) were added subsequently and the reaction mixture was stirred at room temperature for 18 h. The solvent was evaporated in vacuo and the residue was chromatographed (dichloromethane:methanol 100:2). Fractions containing the product were dried, affording N-acylated product 4 (644 mg, 97%) as a white powder.
Fp: 149–150°C, 1H-NMR (CDCl3, 200 MHz): = 7.64 (d, J56 = 8.2, 1H, H-6), 7.42 – 7.25 (m, 14 H, Ar-H), 6.83 (d, J = 8.8, 4H, Ar-H), 6.20 (d, J2'1' = 8.6, 1H, H-1'), 5.40 (d, J65 = 8.1, 1H, H-5), 5.07 (s, 2H, PhCH2), 4.66 (dd, J = 8.3, H-2'), 4.49 (d, J = 5.1, 1H, H-3'), 4.21 (bs, 1H, H-4'), 3.75 (s, 3H, OCH3), 3.39 (bs, 2H, H-5'a, H-5'b), 2.90–2.45 (m, 4H, CH2–CH2). 13C-NMR (CDCl3, 50 MHz): = 173.68 and 172.93 (COOR, CONH), 162.93 (C-4), 158.66 (ArC-4, methoxyphenyl), 151.24 (C-2), 144.01 (ArC-1, phenyl), 139.93 (C-6), 135.39, 135.21 and 135.00 (ArC-1, methoxyphenyl and benzyl), 130.14 (ArC-2,6, methoxyphenyl), 128.58 (ArC-3,5, benzyl), 128.35 (Ar-C), 128.22 (ArC-2,6, benzyl), 128.04 (ArC), 127.12 (ArC-4, phenyl), 113.24 (ArC-3,5, methoxyphenyl), 103.10 (C-5), 87.10 (CPh3), 85.61 and 85.32 (C-1', C-4'), 71.45 (C-3'), 66.93 (PhCH2), 63.78 (C-5'), 57.80 (C-2'), 55.23 (OCH3), 30.69 and 29.65 (CH2–CH2). C41H41N3O10 calculated for 4% H2O: C 64.25 H 5.82 N 5.48, found: C 64.27 H 6.04 N 5.41.
2'-(Benzylsuccinyl)amido-2'-deoxy-5'-(4,4'-dimethoxytriphenyl)methyl-3'-O-succinyluridine (5)
N-Benzylsuccinyl derivative 4 (250 mg, 0.34 mmol) was dissolved in 7 ml of dried pyridine and one drop of N-methylmorpholine was added. After addition of succinyl anhydride (70 mg, 0.70 mmol), the mixture was stirred under exclusion of moisture at 30°C for 5 days. The solvent was removed in vacuo and the residue was co-evaporated three times with dichloromethane. A 10 ml aliquot of dichloromethane was added and the precipitated succinic acid was filtered off. The filtrate was purified by column chromatography (dichlormethane:methanol 100:2) affording hemisuccinate 5 (200 mg, 70%) as a white solid.
1H-NMR (CDCl3, 200 MHz): = 10.46 (bs, 1H, COOH), 7.65 (d, J56 = 8.2, 1H, H-6), 7.48–7.22 (m, 14 H, Ar-H), 6.83 (d, J = 8.7, Ar-H), 6.09 (d, J2'1' = 9.0, 1H, H-1'), 5.46 (d, J2'3' = 5.2, 1H, H-3'), 5.39 (d, J65 = 8.1, 1H, H-5), 5.10 (m, 1H, H-2'), 5.04 (s, 2H, PhCH2), 4.18 (m, 1H, H-4'), 3.75 (s, 6H, OCH3), 3.48 (dd, J5'b5'a = 9.2, 1H, H-5'a), 3.39 (dd, J5'a5'b = 9.1, 1H, H-5'b), 2.64–2.55 (m, 8H, CH2–CH2). 13C-NMR (CDCl3, 50 MHz): = 176.48 (COOH), 173.24 and 172.63 , 170.65 , 163.80 (C-4), 158.72 (ArC-4, methoxyphenyl), 151.30 (C-2), 143.90 (ArC-1, phenyl), 140.16 (C-6), 135.72, 135.04 and 134.89 (ArC-1, methoxyphenyl and phenyl), 130.11 (ArC-2,6, methoxyphenyl), 128.48 (ArC-3,5, benzyl), 128.15, 128.11, 128.03 and 127.75 (ArC), 127.18 (ArC-4, phenyl), 113.38 (ArC-3,5, methoxyphenyl), 102.97 (C-5), 87.45 (CPh3), 85.11 (C-1'), 82.77 (C-4'), 74.14 (C-3'), 66.53 (PhCH2), 63.57 (C-5'), 55.22 (OCH3), 54.08 (C-2'), 30.41, 29.66 and 29.26 (CH2–CH2). C45H45N3O13 calculated: C 64.66, H 5.43, N 5.03; found: C 64.37, H 5.67, N 4.78.
Loading of solid support
TentaGel-NH2 (230 mg, 60 μmol free NH2 groups), purchased from Fluka, was put into a 50 ml flask and suspended in 3 ml of dried DMF for swelling. Succinyl derivative 5 (100 mg, 120 μmol) was separately dissolved in 1 ml of DMF, and 1-hydroxy-1H-benzotriazol (17 mg, 126 μmol) was added. This solution was added to the flask with TentaGel, as were N,N'-diisopropylcarbodiimide (16 mg, 127 μmol) and 4,4'-dimethylaminopyridine (1 mg, 8 μmol). After 16 h of stirring at room temperature, acetic anhydride (12 mg, 118 μmol) in 3 ml of pyridine was added for capping of free amino groups, and stirring was continued for 1 h. TentaGel was filtered off and washed with DMF, methanol and dichloromethane three times each. The resin was dried in vacuo and the loading was quantified by letting a small sample react with 3% trichloroacetic acid and determining the trityl cation concentration by UV spectroscopy (498 nm, = 71 700).
Oligonucleotide synthesis
Standard phosphoramidites and tetrazole solution were obtained from Carl Roth GmbH+Co (Karlsruhe, Germany). Oligonucleotide synthesis was carried out on an ABI 392 Applied Biosystems DNA/RNA-Synthesizer in standard 1 μmol scale in DMT-on mode. Coupling of the first nucleotide to solid-supported modified nucleotide was prolonged to 2 min. Acetonitrile was heated over potassium carbonate and distilled, followed by heating over calcium hydride and distillation. Dichloromethane was dried over phosphorus pentoxide and distilled. Other reagents were obtained in analytical quality and purified before use. Sulfurization was done with tetraethyldiuram disulfide, which was purchased from Sigma and recrystallized (21).
Palladium nanoparticles
Preparation of palladium nanoparticles was done following a procedure by Hirai et al. (22), but with an increased amount of poly-N-vinyl-pyrrolidon (PVP). Palladium(II)chloride (6 mg, 0.034 mmol) was dissolved in 25 ml of methanol, and PVP (mol. wt 25 000 Da, 187 mg, 1.681 mmol of monomer) was separately dissolved in 20 ml of methanol. The solution of PVP was added to the palladium and the mixture was heated to 50°C under exclusion of moisture. After 30 min, a solution of sodium hydroxide (2.65 mg, 0.067 mmol) in 5 ml of methanol was added dropwise. Heating was continued for 5 min and, after cooling, the dark brown suspension was transferred to a dark bottle, where it was stored under an argon atmosphere.
Preparation of oligonucleotide conjugates (7,9)
The synthesis column was removed from the synthesizer and opened. The resin-bound oligonucleotide was transferred to a 5 ml reaction tube of a Quest 210 manual synthesizer. Cleavage of the benzyl protecting group was achieved by phase transfer hydrogenation with PVP-stabilized palladium nanoparticles (0.7 ml) and 1,4-cyclohexadiene (0.05 ml) as hydrogen donor. Cyclohexadiene was added to the PVP-stabilized suspension of palladium, and the resulting solution was microfiltrated (0.22 μm) to ensure no precipitated palladium was applied to the resin. The reaction mixture was stirred at room temperature for a total time of 16 h.
The resin was washed three times with methanol and three times with DMF. Coupling was done by adding a solution of 1-hydroxy-1H-benzotriazol (0.4 mg, 3.0 μmol), N,N'-diisopropylcarbodiimide (0.4 mg, 3.2 μmol) and 5-aminofluorescein (1.1 mg, 3.2 μmol) or putrescine (0.3 mg, 3.4 μmol) in 1 ml of DMF. The mixture was stirred for 18 h and the solution was again filtered off. The resin was washed three times with DMF and three times with water. For cleavage and deprotection of oligonucleotides, 2 ml of concentrated ammonia was added and the mixture was heated to 55°C for 18 h. The solution was filtered and the solvent was evaporated in vacuo.
Purification and analyses of antisense oligonucleotides
Oligonucleotide concentrations were determined by measuring the OD260 in a Hitachi U3000 spectrophotometer. Molar extinction coefficients were calculated as the sum of nucleotides (A, 15 400; G, 11 700; C, 7300, T, 8800; and U, 9950).
Initial purification of oligonucleotides was done with Poly-Pak? columns obtained from Carl Roth GmbH according to the manufacturer. The typical yield after purification was 250 nmol. Analytical HPLC was performed on a Nucleosil? CC 250/4 100-5 C18 column and a LiChroSphere? 100 RP 18 column with the following gradient system: A, 0.1 M triethylammonium acetate in water; B, 0.1 M triethylammonium acetate in 80% acetonitrile, linear gradient 10–40% B in 0–30 min, flow rate 1 ml/min. If Poly-Pak? purification was not satisfactory, preparative HPLC was performed on a LiChrosphere? 100 RP-18 using the same gradient.
Mass spectrometric analysis was done on a Kratos seq MALDI mass spectrometer. Sample preparation was done following the method described by Pieles et al. (23). A 1 μl aliquot of sample (100 pmol/μl in water) was briefly vortexed with 10 μl of a 0.5 M solution of 2,4,6-trihydroxyacetophenone in ethanol and 5 μl of a 0.1 M solution of di-ammonium hydrogen citrate. A 1 μl aliquot of this mixture was spotted on the target and allowed to air dry. The mass spectrometer was run in the negative ion and reflectron mode, and spectra were obtained by summing 50–100 single laser pulses.
Circular dichroism (CD) spectrometry was performed on a Jasco J-810 spectropolarimeter equipped with a Neslab RTE 7 thermostatic unit. Oligonucleotides were diluted to a 9 μM solution in 0.15 M NaCl and 0.01 M Tris–HCl (pH 7.0) buffer. CD spectra were collected from 320 to 210 nm using a quartz cuvette (Hellma 100-QS) with a pathlength of 1 mm. Duplexes were measured after heating the equimolar mixture of complementary strands to 50°C for 10 min and slow cooling to room temperature.
Melting temperatures (Tms) of the duplexes were determined in the Jasco J-810 by slowly heating from 30 to 90°C (50°C/h) and recording CD at 248 nm or OD at 260 nm as a function of temperature. Melting curves showed cooperative form, and Tms were obtained from the maxima of the first derivative plots. Temperatures given are those of the cuvette holder.
Treatment of cells with oligonucleotides and conjugate
Human melanoma 607B cells were grown in 6-well plates. After 24 and 48 h, cells were incubated for 4 h at 37°C with oligonucleotides or oligonucleotide–fluorescein conjugate complexed with lipofectin at a concentration of 100 nM in Opti-MEM medium. After a total of 72 h, cells were harvested.
Western blot analysis
After incubation with oligonucleotides and oligonucleotide conjugate, respectively, whole-cell extracts of cultured cells were prepared as described previously (24). Briefly, melanoma cells were washed with phosphate-buffered saline (PBS) and then scraped into 80 μl of lysis buffer . Cell lysates were freeze–thawed in liquid nitrogen three times and centrifuged at 15 000 r.p.m. for 20 min at 4°C to remove cell debris. Protein concentrations were determined by Bradford protein assay.
Aliquots of cell extracts containing 15 μg of protein were resolved by 12% SDS–PAGE and blotted to a PVDF membrane (ImmobilonTM P). The membrane was blocked for 1 h in 0.2% I-block (Tropix, Bedford, MA) in PBS and cut in half. The top part was incubated overnight at 4°C with actin antibody (2.0 μg/ml in I-block solution, Sigma, St Louis, MO) and the bottom part under the same conditions with monoclonal antibody recognizing Bcl-2 (2.5 μg/ml in I-block solution, Zymed Laboratories, Inc., San Francisco, CA). Secondary antibody incubation was carried out using Bcl-2 goat anti-mouse and actin goat anti-rabbit conjugated to alkaline phosphatase (2 μl in 10 ml of I-block solution, Tropix). The membranes were subsequently washed with I-block and assay buffer. For visualization, chemiluminescence substrate CSPD (0.25 M, 5 min, Tropix) was used. Expression levels were quantified by densitometry of autoradiograms with a Herolab EASY RH densitometer (Herolab, Wiesloch, Germany) and the EASY Win32 software. Signal strengths of Bcl-2 signals were normalized to actin, and the ratios between Bcl-2 protein in antisense oligonucleotide-treated extracts and tumour control extracts were calculated.
RESULTS AND DISCUSSION
2'-Amino-2'-deoxy-5'-dimethoxytrityluridine (1) was synthesized in good yield following the four-step procedure published by McGee et al. (20). We decided to use a succinyl linker for post-synthetic attachment of various ligands by an amide bond. This 4-carbon linker has a suitable length for minimizing the steric hindrance that can be seen for direct acylation of the 2'-amino group (15,25) and is advantageous for introduction of ligands in the minor groove. For solid phase synthesis, the carboxyl group had to be protected. A suitable protecting group had to be stable at all stages of the oligonucleotide synthesis cycle, but also labile enough to be cleaved selectively after oligonucleotide chain elongation without loss of base protecting groups. The benzyl ester proved to be a suitable masking group, since it is stable under acid and slightly basic conditions, but is readily cleaved by hydrogenolysis (26). Whereas using hydrogen gas leads to significant reduction of double bonds of pyrimidine bases, the use of cyclohexadiene as a hydrogen donor in a phase transfer catalysis shows no such side reactions (27). Bajwa reported the selctive cleavage of benzyl esters and N-benzyl groups under these conditions (28). 2'-Amino-2'-deoxy-5'-dimethoxytrityluridine (1) reacted in a highly specific way with succinic acid anhydride to give only N-acylated product 2. As expected, no 3'-acylated product was formed. As benzylation of the free carboxyl group proceeded only in poor yield, we prepared the monobenzyl ester of succinic acid (3) and coupled it to the 2'-amino group using N,N'-diisopropylcarbodiimide, 1-hydroxy-1H-benzotriazol and 4-dimethylaminopyridine in DMF. In accordance with our expectations, again only N-acylated product was isolated in excellent yield (Scheme 1). The structures and conformations of these products were verified by 1H–1H and 1H–13C correlation NMR experiments.
Scheme 1. Synthesis of nucleoside derivative for attachment to solid phase. (a) Succinic acid anhydride, pyr, r.t., 95%. (b) Succinic acid monobenzyl ester 3, N,N'-diisopropylcarbodiimide, 1-hydroxy-1H-benzotriazol, 4-dimethylaminopyridine, tetrahydrofuran, 97%. (c) Succinic acid anhydride, pyr, 30°C, 70%.
NMR spectral data suggest the adoptation of a C2'-endo configuration of these N-acyl derivatives of uridine. The chemical shifts and especially the coupling constants of carbons of the ribose moiety are comparable with those found by other authors for 2'-amido-2'-deoxyadenosine derivatives (29) and 2'-acetamido-2'-deoxyuridine and guanosine, respectively (30) (Table 1). As has been shown by modeling experiments (29), coupling constants for H1'–H2' (8.0 for 2 and 8.6 for 4) and H2'–H3' (5.6 and 5.1, respectively) are a clear indication of the C2'-endo configuration. In contrast to 2'-O-substituted derivatives, 2'-N-acyl nucleosides, the substituents having a low electronegativity, tend towards a B-DNA configuration, which is typical for DNA strands, whereas the ribose of RNA is fixed in the C3'-endo pucker and, consequently, an A-type DNA helix. Although Hendrix et al. (30) concluded that the preferred B-DNA conformation was responsible for the lower melting temperature for oligonucleotides with incorporated 2'-acetamido-2'-deoxyuridine (2.7–7.2°C), the locked A-DNA configuration of 2'-O-alkyl-modified oligonucleotides was found to be a rather disadvantageous property for antisense agents despite the high stability against enzymatic degradation, since such derivatives show no activation of RNase H, nowadays believed to be one of the most important aspects of antisense activity.
Table 1. Comparison of 1H-NMR data of some 2'-amido-2'-deoxynucleosides (21,22)
Based on these data, 4 seems to be a suitable building block for the addition of ligands to oligonucleotide. Its preparation is straightforward and the conformation of the ribose moiety seems to be advantageous for antisense agents.
For the synthesis of oligonucleotide conjugates, the modified nucleoside should be incorporated at the 3' end of the oligonucleotide. Thus, it had to be attached directly to the solid phase. Controlled pore glass (CPG) is the ideal resin for oligonucleotide synthesis, but has disadvantages regarding reactions in polar solvents such as DMF, used in peptide chemistry (31). Therefore, we decided to use TentaGel, which is suitable for both oligonucleotide (32) and peptide chemistry (33). The addition of 4 to TentaGel with an attached succinyl linker gave only very poor loading yields (<10 μmol/g), obviously due to the steric bulk of the 2'-substituent. Thus, we derivatized 4 by reaction with succinic anhydride to give hemisuccinate 5 (Scheme 1). As TLC controls revealed, the reaction with an excess of succinic acid anhydride progressed only slowly, further confirmation of the great steric hindrance of the 3' position. After reaction times of several days, good yields were achieved. Coupling of hemisuccinate 5 to TentaGel-NH2 resulted in a much improved loading of 190 μmol/g.
Solid phase-coupled derivative 5 could be employed in standard oligonucleotide synthesis cycles. The coupling time of the first phosphoramidite was prolonged to improve yields. ASWYs (average stepwise yields) as determined by trityl cation assay were routinely >95%.
The 18mer oligonucleotide bearing the sequence of Oblimersen (5'-TCTCCCAGCGTGCGCCAU-3'), a complete phosphorothioate backbone and with incorporation of the modified uridine nucleoside at the 3' end, was synthesized. Direct cleavage and deprotection by reaction of ammonia gave the product with a free carboxyl group (6) as determined by MALDI mass spectrometry. Treatment of the modified monomer with 10% palladium on carbon and cyclohexadiene in a phase transfer catalysis cleaved the benzyl ester in liquid phase. Reaction of resin-bound oligonucleotide with palladium on charcoal proved to be difficult. The resin turned black after a short time and, even after repeated washing, a small fraction of activated carbon remained on the solid support. After treatment with concentrated ammonia, virtually no product was found. The oligonucleotide was presumably absorbed at the charcoal still present at the resin surface. Therefore, other possibilities for catalytic hydrogenolysis avoiding the use of activated carbon were researched. Different palladium(0)complexes have been used in solid phase chemistry (33,34). Recently, palladium nanoparticles, stabilized by PVP (21) to prevent aggregation, were used for the hydrogenolysis with hydrogen gas of benzyl ether groups of solid-supported sugar derivatives (35).
The phase transfer reaction of resin-bound oligonucleotide with palladium nanoparticles and 1,4-cyclohexadiene as hydrogen donor selectively cleaved the benzyl ester. After 16 h, the liquid was filtered off and a UV spectrum was recorded. Overlay with the spectrum of unreacted solution of palladium nanoparticles with cyclohexadiene showed a new band at 280 nm, which indicated the formation of toluene, the product of hydrogenolysis. In order to determine the optimum reaction time, UV spectra of the reaction solution filtrated at 1, 2 and 4 h were recorded. The obtained data indicated that debenzylation was not yet quantitative at that time. Therefore, a reaction time of 16 h was used in further syntheses. Although inhibition of reaction due to poisoning of the catalyst by sulfur was feared, no differences were observed in reactivity of phosphorodiesters and phosphorothioates. Literature reports on sulfur poisoning of catalytic phase transfer hydrogenation are conflicting (27,36), but it has been established that palladium catalysts can, in principle, work despite the presence of sulfur (37).
To verify that no alterations of phosphorothioate backbone or reduction of pyrimidine bases occurs, oligonucleotide 6 was treated with 1,4-cyclohexadiene and palladium nanoparticles for 5 days. Repeated HPLC analyses during this time proved that no reaction took place as the chromatograms remained unchanged.
For the first coupling of a ligand, fluorescein was chosen. It is a widely used reporter group and successful attachment can very easily be verified because of its fluorescent activity. Well documented peptide coupling reagents were employed for the reaction (38,39). A solution of N,N'-diisopropylcarbodiimide, 1-hydroxy-1H-benzotriazol and 5-aminofluorescein in DMF was added to the resin and stirred for 18 h to ensure completion of the reaction. The solution was filtered off and, on visual inspection, it was clearly brighter than unreacted solution, indicating successful attachment of fluorescein to the oligonucleotide. Cleavage and deprotection afforded a dark yellow solution that showed fluorescence. A control experiment where the carboxyl deprotection step was omitted did not afford a coloured or fluorescent product. Thus it was shown that no unspecific coupling of fluorescein occurred, proving the structure of conjugate 7 shown in Figure 2.
Figure 2. Structures of oligonucleotide conjugates 7 and 9.
Purification was achieved by Poly-Pak? columns or preparative HPLC. Retention times can be seen in Table 2. Due to the lipophilic character, the retention time of fluorescein conjugate 7 was significantly longer than that of modified oligonucleotide 6. No peak of the unreacted oligonucleotide 6 was found in the chromatogram (Fig. 3), indicating a conjugation yield of >95%.
Table 2. HPLC retention times, MALDI mass and melting temperature (Tm)
Figure 3. HPL-chromatogram of conjugate 7 (right) after Poly-Pak? purification; injected sample 10 μg.
The successful conjugation of fluorescein confirms the selective benzyl ester cleavage in nearly quantitative yield on the solid phase and the favorable steric properties of modified nucleoside building block 4. The reaction of amines and carboxyl groups is a well researched procedure and has been used extensively in oligonucleotide chemistry (6,9). Hence, our method can be used for conjugation of various ligands bearing amino functionalites to oligonucleotides.
To further prove the generality of the method, we chose the polyamine putrescine as ligand. Synthesis following the described protocol using unprotected putrescine as ligand gave the polyamine–oligonucleotide conjugate 9 (Fig. 2) in good yield and excellent purity.
CD spectra of unmodified phosphodiester (8), unmodified phosphorothioate (Oblimersen), unconjugated oligonucleotide 6 and the conjugate with fluorescein (7) were recorded. All experiments were performed in a buffer containing 0.1 M Tris–HCl (pH 7.0) and 0.15 M NaCl. The phosphodiester exhibited clear characteristics of a B-DNA type helix. Modified oligonucleotides showed essentially the same spectra, but with reduced band intensities, especially of the minimum at 248 nm (Fig. 4).
Figure 4. Circular dichroism of 5'-TCTCCCAGCGTGCGCCA(T or U)-3'. Phosphodiester 8 (squares), phosphorothiaote Oblimersen (rhombi), modified phosphorothioate 6 (triangles) and conjugate 7 (circles).
Duplexes of phosphorothioate backbone oligonucleotides with complementary DNA showed nearly identical CD spectra. They are shown in Figure 5. Only small differences were observed regarding oligonucleotide–fluorescein conjugate 7, which gave slightly increased intensities of the positive bands at 278 and 218 nm and lower intensity of the negative band at 248 nm. In comparison with the duplex of 8, the thioates (6, 7 and Oblimersen) all had lower intensities in all bands, an observation that is in accordance with published data about other oligonucleotide sequences (40).
Figure 5. Circular dichroism of duplexes of 8 (squares), Oblimersen (rhombi), 6 (triangles) and 7 (circles) with phosphodiester sense strand.
Experiments determining the melting temperatures of duplex oligonucleotides all gave cooperative curves (Fig. 6). Melting temperatures shown in Table 2 are the average of four different determinations for each substance. The transition temperature of Oblimersen was 9.2°C (0.54°C per modification) lower than that of phosphorodiester 8, consistent with known data (14,41). The decrease in melting temperature of 6 was 4°C compared with Oblimersen, but that of the conjugate 7 was only 1°C. Polyamine conjugate 9 gave a reduction in transition temperature of 2°C. A substitution of 2'-deoxyuridine for thymidine accounts for an average reduction of melting temperature of 0.5°C, as was shown using a considerable number of different sequences (11). Presumably, the additional negative charge of the carboxyl group, located in the minor groove, considerably decreases duplex stability by means of electrostatic repulsion, whereas a conjugation removes this negative charge and significantly increases hybridization affinity in comparison with the derivative with a free carboxyl group despite its greater steric hindrance caused by the ligands. The aromatic fluorescein structure may lead to a stabilization of duplex structure by contributing to the base stacking of nucleobases. This presumption is backed by the slightly increased intensity of the CD band at 278 nm, reflecting an increase in base stacking.
Figure 6. Melting profiles of 6 (squares) and 7 (triangles).
Cell culture tests were performed in human melanoma cell line 607B with an oligonucleotide concentration of 100 nM in the presence of lipofectin as uptake enhancing transfection reagent (23). Cell lysates were analysed by western blot. Downregulation of Bcl-2, normalized against changes in the actin band, was similar for conjugate 7 and control Oblimersen, while modified oligonucleotide 6 exhibited a weaker effect. Treatment of cells with fluorescein–oligonucleotide conjugate 7 gave a reduction of Bcl-2 of 79% in comparison with melanoma cells treated with lipofectin only, while Oblimersen reduced the expression of target protein by 85%. Modified oligonucleotide 6 showed the weakest effect of the three tested substances, with a reduction of Bcl-2 level of 69%. Control reverse and mismatched oligonucleotides showed an inhibition of target protein by only up to 17%, establishing the true antisense effect of the tested products (23). These results confirm the findings of the melting temperature experiments. The conjugation of ligands in the minor groove by a succinic linker tethered to the 2' position of 2'-aminouridine leads only to a minor depletion in hybridization affinity and thus in antisense efficacy.
In conclusion, the fluorescent conjugate 7 has excellent base-pairing properties and in vitro efficacy. The convenient location of the ligand in the minor groove at the 3' end of the oligonucleotide results in very little disturbance of base pairing and consequently supression of gene expression. Our method for conjugation of ligands to oligonucleotides via a succinyl linker at the 2' position of 2'-amino-2'-deoxyuridine is selective, gives excellent yield and can easily be employed. It has the considerable advantage over previously published procedures that the building block used for attaching ligands at the 2' position can be prepared much more easily. A method for selectively cleaving benzyl esters in oligonucleotide solid phase chemistry has been developed by the use of palladium nanoparticles in combination with cyclohexadiene as hydrogen donor. It has been shown that this reaction succeeds as well for phosphorothioates. Studies on the cellular uptake of fluorescent conjugate 7 as well as the preparation and characterization of further oligonucleotide conjugates will be published at a later date.
REFERENCES
Tamm,I., D?rken,B. and Hartmann,G. (2001) Antisense therapy in oncology: new hope for an old idea? Lancet, 358, 489–497.
Urban,E. and Noe,C.R. (2003) Structural modifications of antisense oligonucleotides. Farmaco, 58, 243–258.
Noe,C.R. and Kaufhold,L. (2000) Chemistry of antisense oligonucleotides. In Gualteri,F. (ed.), New Trends in Synthetic Medicinal Chemistry. Wiley-VCH, Weinheim, pp. 261–346.
Manoharan,M., Tivel,K.L. and Cook,P.D. (1995) Lipidic nucleic acids. Tetrahedron Lett., 36, 3651–3654.
Manoharan,M. (1993) Designer antisense oligonucleotides: conjugation chemistry and functionality placement. In Crooke,S.T. and Lebleu,B. (eds), Antisense Research and Applications. CRC Press, Boca Raton, FL, pp. 303–349.
Tung,C.-H. and Stein,S. (2000) Preparation and applications of peptide–oligonucleotide conjugates. Bioconjug. Chem., 11, 605–618.
Beaucage,S.L. and Iyer,R.P. (1993) The functionalization of oligonucleotides via phosphoramitide derivatives. Tetrahedron, 49, 1925–1963.
Manoharan,M. (2001) Oligonucleotide conjugates in antisense technology. In Crooke,S.T. (ed.), Antisense Drug Technology. Marcel Dekker, Inc., New York, NY, pp. 391–471.
Virta,P., Katajisto,J., Niittym?ki,T. and L?nnberg,H. (2003) Solid-supported synthesis of oligomeric bioconjugates. Tetrahedron, 59, 5137–5174.
Werner,D., Brunar,H. and Noe,C.R. (1998) Investigations on the influence of 2'-O-alkyl modifications on the base pairing properties of oligonucleotides. Pharm. Acta Helv., 73, 3–10.
Cotten,M., Oberhauser,B., Brunar,H., Holzner,A., Issakides,G., Noe,C.R., Schaffner,G., Wagner,E. and Birnstiel,M.L. (1991) 2'-O-methyl, 2'-O-ethyl oligoribonucleotides and phosphorothioate oligodeoxyribonucleotides as inhibitors of the in vitro U7 snRNP-dependent mRNA processing event. Nucleic Acids Res., 19, 2629–2635.
Freier,S.M. and Altmann,K.-H. (1997) The ups and downs of nucleic acid duplex stability: structure–stability studies on chemically-modified DNA:RNA duplexes. Nucleic Acids Res., 25, 4429–4443.
Manoharan,M. (1999) 2'-Carbohydrate modifications in antisense oligonucleotide therapy: importance of conformation, configuration and conjugation. Biochim. Biophys Acta, 1489, 117–130.
Manoharan,M., Prakash,T.P., Barber-Peoch,I., Bhat,B., Vasquez,G., Ross,B.S. and Cook,P.D. (1999) N-(2-Cyanoethoxycarbonyloxy) succinimide: a new reagent for protection of amino groups in oligonucleotides. J. Org. Chem., 64, 6468–6472.
Hwang,J.-T. and Greenberg,M.M. (2001) Synthesis of 2'-modified oligodeoxynucleotides via on-column conjugation. J. Org. Chem., 66, 363–369.
Zubin,E.M., Romanova,E.A. and Oretskaya,T.S. (2002) Modern methods for the synthesis of peptide–oligonucleotide conjugates. Russ. Chem. Rev., 71, 239–264.
Hwang,J.-T. and Greenberg,M.M. (1999) Synthesis of modified oligodeoxyribonucleotides on a solid-phase support via derivatization of a selectively revealed 2'-amino-2'-deoxyuridine. Org. Lett., 1, 2021–2024.
Dias,N. and Stein,C.A. (2002) Potential roles of antisense oligonucleotides in cancer therapy. The example of Bcl-2 antisense oligonucleotides. Eur. J. Pharm. Biopharm., 54, 263–269.
Jansen,B., Wacheck,V., Heere-Ress,E., Schlagbauer-Wadl,H., Hoeller,C., Lucas,T., Hoermann,M., Hollenstein,U., Wolff,K. and Pehamberger,H. (2000) Chemosensitisation of malignant melanoma by BCL2 antisense therapy. Lancet, 356, 1728–1733.
McGee,D.P.C., Vaughn-Settle,A., Vargeese,C. and Zhai,Y. (1996) 2'-Amino-2'-deoxyuridine via an intramolecular cyclization of a trichloroacetimidate. J. Org. Chem., 61, 781–785.
Vu,H. and Hirschbein,B.L. (1991) Internucleotide phosphite sulfurization with tetraethylthiuram disulfide. Phosphorothioate oligonucleotide synthesis via phosphoramitide chemistry. Tetrahedron Lett., 32, 3005–3008.
Hirai,H., Yakura,N., Seta,Y. and Hodoshima,S. (1998) Characterization of palladium nanoparticles protected with polymer as hydrogenation catalyst. React. Funct. Polymers, 37, 121–131.
Pieles,U., Zürcher,W., Sch?r,M. and Moser,H.E. (1993) Matrix-assisted laser desorption ionization time-of-flight mass spectrometry: a powerful tool for the mass and sequence analysis of natural and modified oligonucleotides. Nucleic Acids Res., 21, 3191–3196.
Jansen,B., Schlagbauer-Wadl,H., Brown,B.D., Bryan,R.N., van Elsas,A., Müller,M., Wolff,K., Eichler,H.-G. and Pehamberger,H. (1998) bcl-2 antisense therapy chemosensitizes human melanoma in SCID mice. Nature Med., 4, 232–234.
Sharma,R.A., Bobek,M. and Bloch,A. (1975) Preparation and biological activity of some aminoacyl and peptidyl derivatives of 2'-amino-2'-deoxyuridine. J. Med. Chem., 18, 955–957.
Greene,T.W. and Wuts,P.G.M. (1991) Protection for the carboxyl group. In Greene,T.W. and Wuts,P.G.M. (eds), Protective Groups in Organic Synthesis. John Wiley & Sons, New York, NY, pp. 224–276.
Watkins,B.E., Kiely,J.S. and Rapoport,H. (1982) Synthesis of oligodeoxyribonucleotides using n-benzyloxycarbonyl-blocked nucleosides. J. Am. Chem. Soc., 104, 5702–5708.
Bajwa,J.S. (1992) Chemoselective deprotection of benzyl esters in the presence of benzyl ethers, benzyloxymethyl ethers and N-benzyl groups by catalytic transfer hydrogenation. Tetrahedron Lett., 33, 2299–2302.
Van Calenbergh,S., Van Den Eeckhout,E., Herdewijn,P., De Bruyn,A., Verlinde,C., Hol,W., Callens,M., Van Aerschot,A. and Rozenski,J. (1994) Synthesis and conformational analysis of 2'-deoxy-2'-(3-methoxybenzamido)adenosine, a rational-designed inhibitor of trypanosomal glyceraldehyde phosphate dehydrogenase (GAPDH). Helv. Chim. Acta, 77, 631–644.
Hendrix,C., Devreese,B., Rozenski,J., Van Aerschot,A., De Bruyn,A., Van Beeumen,J. and Herdewijn,P. (1995) Incorporation of 2'-amido-nucleosides in oligodeoxynucleotides and oligoribonucleotides as a model for 2'-linked conjugates. Nucleic Acids Res., 23, 51–57.
Beaucage,S.L. and Iyer,R.P. (1992) Advances in the synthesis of oligonucleotides by the phosphoramidite approach. Tetrahedron, 48, 2223–2311.
Wright,P., Lloyd,D., Rapp,W. and Andrus,A. (1993) Large scale synthesis of oligonucleotides via phosphoramidite nucleosides and a high-loaded polystyrene support. Tetrahedron Lett., 34, 3373–3376.
de la Torre,B., Avino,A., Tarrason,G., Piulats,J., Albericio,F. and Eritja,R. (1994) Stepwise solid-phase synthesis of oligonucleotide–peptide hybrids. Tetrahedron Lett., 35, 2733–2736.
Ma,S., Duan,D. and Wang,Y. (2002) Palladium(0)-catalyzed coupling-cyclization reaction of polymer-supported aryl iodides with 1,2-allenyl carboxylic acids. Solid phase parallel synthesis of butenolides. J. Comb. Chem., 4, 239–247.
Kanie,O., Grotenberg,G. and Wong,C.-H. (2000) Removal of benzyl protecting groups from solid-supported compounds by hydrogenolysis using palladium nanoparticles. Angew. Chem. Int. Ed., 39, 4545–4547.
Anantharamaiah,G.M. and Sivanandaiah,K.M. (1977) Transfer hydrogenation; a convenient method for removal of some commonly used protecting groups in peptide synthesis. J. Chem. Soc. Perkin Trans. I, 5, 490–491.
Castro,F.J., Meyer,G. and Zampieri,G. (2002) Effects of sulfur poisoning on hydrogen desorption from palladium. J. Alloys Comp., 330–332, 612–616.
Lloyd-Williams,P., Albericio,F. and Giralt,E. (1993) Convergent solid phase peptde synthesis. Tetrahedron, 49, 11065–11133.
Henkel,B., Zhang,L. and Bayer,E. (1997) Investigations on solid-phase peptide synthesis in N-to-C direction (inverse synthesis). Liebigs Ann./Receuil, 2161–2168.
Clark,C.L., Cecil,P.K., Singh,D. and Gray,D.M. (1997) CD, absorption and thermodynamic analysis of repeating dinucleotide DNA, RNA and hybrid duplexes 12·12 and the influence of phosphorothioate substitution. Nucleic Acids Res., 25, 4098–4105.
Mier,W., Eritja,R., Mohammed,A., Haberkorn,U. and Eisenhut,M. (2000) Preparation and evaluation of tumor-targeting peptide–oligonucleotide conjugates. Bioconjug. Chem., 11, 855–860.(Johannes Winkler, Ernst Urban, Doris Los)