WO1993013118A1

WO1993013118A1 - Synthesis of phosphorothioate analogues of oligo- and polynucleotides

Info

Publication number: WO1993013118A1
Application number: PCT/GB1992/002395
Authority: WO
Inventors: Colin Bernard Reese
Original assignee: King's College London
Priority date: 1991-12-23
Filing date: 1992-12-23
Publication date: 1993-07-08
Also published as: GB9127243D0

Abstract

A process for the solid phase production of a phosphorothioate analogue of an oligo- or polynucleotide from phosphoramidite building blocks in the presence of a sulphur-transfer agent characterised in that the sulphur-transfer agent corresponds to the following general formula: X - Sn - X wherein n represents 3 or 4; and X represents RC(O)- or RC(S)- wherein R represents alkyl, cycloalkyl or aryl, (R'O)C(O)- wherein R' represents alkyl, (R2N)C(O)- or (R2N)C(S)- wherein R is as defined above and R2NH may be a cyclic secondary amine, RSO2- wherein R is as defined above, (RO)2P(O)- or (RO)2P(S)- wherein R is as defined above; is disclosed.

Description

SYNTHESIS OF PHOSPHOROTHIOATE ANALOGUES OF OLIGO- AND POLY- NUCLEOTIDES

This invention relates to the synthesis of phosphorothioate analogues of oligo- and poly- nucleotides; more particularly, it relates to the solid phase synthesis of phosphorothioate analogues of, preferably, oligo- and poly- deoxyribonucleotides from phosphora idite building blocks using certain diacyl tri- and tetra-sulphides as sulphur-transfer agents.

(The present invention is equally applicable to the production of "oligo-" and "poly-" nucleotides. As there is no strict demarcation, the former term will generally be used.)

Antisense oligodeoxyribonucleotides, (see, for example, "Oligodeoxynucleotides - Antisense inhibitors of gene expression", Cohen, J.S., Ed., Macmillan, London, 1989), which are complementary to particular RNA targets, may inhibit translation and thereby provide a highly specific method for the inhibition of gene expression. Phosphorothioate analogues of oligonucleotides in which one of the non-bridging oxygen atoms of the internucleotide linkages is replaced by a sulphur atom are much more resistant to nuclease-promoted degradation than corresponding unmodified oligomers, (see, for example, Eckstein, F., Ann. Rev. Biochem. 54_., 367-402, 1985) , and are therefore potentially valuable antisense therapeutic agents.

Although oligonucleotide phosphorothioates have been prepared by the phosphotriester approach in solution, (see, for example, Burgers, P.M.J. , and Eckstein, F. , Biochemistry, 18., 592-596, 1979; and Kemal, 0., et al, J. Chem. Soc, Chem. Co mun., 591-593, 1983) , they are at present much more conveniently prepared by solid phase synthesis using suitably-protected nucleoside H-phosphonate, (see, for example, Froehler, B.C., Tet. Lett., 2_7_, 5575- 5578, 1986) , or phosphoramidite, (see, for example, Stec, W.J., et al, JACS, 106, 6077-6079, 1984; and Stein, C.A., et al. Nucleic Acids Res., JL6_., 3209-3221, 1988) , building blocks. While the H-phosphonate approach has the advantage of requiring only one sulphur-transfer step after the whole sequence has been assembled, the phosphoramidite approach requires a sulphur-transfer step in the course of each synthetic cycle. However, the H-phosphonate approach does not permit the synthesis of oligonucleotides containing both natural phosphodiester and modified (e.g. phosphorothioate diester) internucleotide linkages, and is considered by some workers to be less efficient than the phosphoramidite approach, (see, for example, Iyer, R.P., et a_l, J. Org. Chem., 55., 4693-4699, 1990; and Vu, H. , and Hirschbein, B.L., Tet. Lett., .32_., 3005-3008, 1991).

Both approaches involve the use of a sulphur- transfer agent and, in the case of the phosphoramidite approach, it is of crucial importance that the reagent should be freely soluble in a suitable organic solvent, such as dichloromethane, acetonitrile or tetrahydrofuran, and that it should effect sulphur-transfer as rapidly as possible so as not unduly to extend the duration of each synthetic cycle. Elemental sulphur, which was first used for this purpose, (see, for example, Stec, loc cit; and Stein, loc cit) , does not have particularly satisfactory solubility properties and also does not effect the conversion of phosphite into phosphorothioate triesters very rapidly. For these reasons, there has recently been a search for more suitable sulphur-transfer agents, leading to recommendations that di-(phenylacetyl) disulphide:

O 3H - 1,2 - benzodithiole - 3 - one 1,1 - dioxide:

and tetraethylthiuram disulphide:

should be used in the phosphoramidite approach to the solid phase synthesis of o1 igodeoxyribonuc1eotide phosphorothioates, (see, for example, Kamer, P.C.J., et al, Tet. Lett., 2JD, 6757-6760, 1989; Iyer, R.P., et a_l, JACS, 112 , 1253-1254, 1990, and Iyer, et aJL, supra; and Vu and Hirschbein, loc cit; respectively) . Although all three of these reagents appear to have good solubility properties, only 3H-l,2-benzodithiole-3-one 1,1-dioxide appears to transfer sulphur very rapidly. However, it is not particularly easy to prepare and this could be disadvantageous if oligonucleotide phosphorothioates prove to be useful in antisense chemotherapy and are consequently required in relatively large quantities.

The present invention relates to a process for the solid phase production of phosphorothioate analogues of oligo- or poly- nucleotides from phosphoramidite building blocks in the presence of a sulphur-transfer agent characterised in that the sulphur-transfer agent corresponds to the following general formula:

X - S„ - X

wherein n represents 3 or ;

and

5 X represents RC(O)- or RC(S)- wherein R represents alkyl, cycloalkyl or aryl, (R'O)C(O)- wherein R represents alkyl, (R₂N)C(0)- or (R₂N)C(S)- wherein R is as defined above and R₂NH may be a cyclic 10 secondary amine,

RS0₂- wherein R is as defined above, (R0)₂P(0)- or (RO)₂P(S)- wherein R is as defined above.

15 According to a preferred embodiment of the present invention, it has now been found that dibenzoyl tetrasulphide:

an easily prepared crystalline solid (see, for example, Bloch, I., and Berg ann, M. , Ber. , 5_3_, 961-977, 1920), may 5 be used as an efficient sulphur-transfer agent which appears both to have good solubility properties and to transfer sulphur very rapidly. Furthermore, the presently-preferred reagent may be readily prepared in 81% yield in a one-step reaction from thiobenzoic acid and sulphur monochloride, 0 which are both relatively cheap, commercially-available starting materials; it may be isolated as a stable, pale yellow crystalline solid that is extremely soluble in chloroform, dichloro ethane and tetrahydrofuran, but not in acetonitrile. 5

2 Ph — C -- S_jCI; Et₃N. t≤trarivαroturan / S ^s. ^~^Ph

V-.

SH 7 Ph S S'

'.0 In fact, the presently-preferred dibenzoyl tetrasulphide was prepared by an advantageous modification of the original literature procedure. A solution of triethylamine (16.2 ml, 0.12 mol) in dry tetrahydrofuran (100 ml) was added dropwise over a period of 20 minutes to a stirred solution of redistilled thiobenzoic acid (16.5 g, 0.12 mol) and sulphur monochloride (6.0 ml, 0.075 mol) in dry tetrahydrofuran (150 ml) at 0°C (ice-water bath) . The products were stirred for a further period of 1 hour at room temperature and were then filtered. The filtrate was concentrated under reduced pressure and the residue was extracted with warm ethyl acetate (200 ml) . After filtration, the extract was cooled and concentrated under reduced pressure. The residue was crystallized from ethyl acetate to give dibenzoyl tetrasulphide. [Observed: C, 49.5, H, 2.95; S, 37.5. Calculated for C₁₄H₁₀O₂S₄: C, 49.7; H, 3.0; S, 37.9%], m.p. 84-85°C; yield, 16.4 g (81%) ; <S_C[CDC1₃ ,90.6 MHz] 128.0, 128.9, 134.4, 134.9, 186.8.

Although one preferred embodiment of the present invention concerns the use of, in particular, dibenzoyl tetrasulphide as a sulphur-transfer agent, it is not so- limited. As will be generally appreciated by those skilled in the art, the main requirements for a sulphur-transfer agent in the context of oligonucleotide phosphorothioate synthesis are as follows:

(1) it should be a readily-prepared solid and preferably crystalline;

(2) it should be indefinitely stable in the solid state and relatively stable in solution at or below room temperature; if it does decompose slowly in solution, the decomposition products should also be soluble in the solvent used;

(3) it should be very soluble (say, at least 0.5M) in one or more of the solvents, including acetonitrile, tetrahydrofuran and dichloromethane, that are commonly used in solid phase oligonucleotide synthesis;

(4) in the case of solid phase oligodeoxyribonucleotide synthesis by the phosphite triester approach, (i.e. by using standard nucleoside phosphoramidite 2-cyanoethyl ester building blocks) , it should be possible to transfer sulphur to the phosphite triester groups within ca. 60 seconds in each synthetic cycle to such an extent that the internucleotide linkages of unblocked oligonucleotide products containing up to 20 nucleoside residues are at least 99% phosphorothioates, as determined by ³¹P-NMR spectroscopy.

As indicated above, in addition to the presently- preferred dibenzoyl tetrasulphide and the almost equally effective but somewhat less soluble dibenzoyl trisulphide, (see, for example, Bloch and Bergmann, loc cit) , other diacyl tetra- and tri-sulphides: X-S₄-X and X-S₃-X, respectively, may be considered for use as effective sulphur-transfer agents. The latter compounds may be derived from (i) other carboxylic acids (i.e. X = RC(0)-, wherein R represents an alkyl, cycloalkyl or aryl group) ; (ii) thiocarboxylic acids (i.e. X = RC(S)-, wherein R represents an alkyl, cycloalkyl or aryl group) , (iii) carbonate esters (i.e. X = (RO)C(O)-, wherein R represents alkyl); (iv) carbamic acids and thiocarbamic acids (i.e. X = (R₂N)C(0)- or (R₂N)C(S)-, when R represents an alkyl, cycloalkyl or aryl group, and R₂NH may be a cyclic secondary amine); (v) sulphonic acids (i.e. X = RS0₂-_/ wherein R represents an alkyl, cycloalkyl or aryl group) ; and (vi) phosphate or phosphorothioate diesters (i.e. X = (R0)₂P(0)- or X = (RO)₂P(S)-, wherein R represents an alkyl, cycloalkyl or aryl group) . In the case of tri- and tetra- sulphides derived from simple carboxylic acids (i.e. X = R(C0)-), it is preferred that R represents an aryl group, which may be substituted, for example, with para- or meta- methyl. The preparation of such reagents is within the competence of those skilled in the art.

The following further illustrates the present invention:

The conversion of tri ethyl phosphite, (MeO)₃P, (<S (CDC1₃) 142.0) into trimethyl phosphorothioate, (MeO)₃P=S, (5 (CDCI₃) 73.7) in CDCI₃ solution at 35°C was investigated by ³¹P-NMR spectroscopy. A 0.1M solution of trimethyl phosphite was allowed to react in turn with 2.0 mol. equiv. of (i) dibenzoyl tetrasulphide, BzSSSSBz, (ii) dibenzoyl trisulphide BzSSSBz, and (iii) dibenzoyl disulphide BzSSBz, (see, for example, Organic Syntheses, Coll. Vol. Ill, 116- 118, 1955) . While reactions (i) and (ii) went to completion in less than 1 minute to give the phosphorothioate as the sole product, reaction (iii) required more than 24 minutes

(tι ~ 4.5 minutes) to go to completion. (It cannot be assumed that dibenzoyl tetrasulphide is quantitatively converted by trimethyl phosphite into dibenzoyl trisulphide as dibenzoyl trisulphide also reacts rapidly with trimethyl phosphite. It has not been ascertained which sulphur atom, i.e. one adjacent or one not adjacent to a carbonyl group, is transferred from the tetrasulphide to the phosphite.)

As dibenzoyl tetrasulphide was found to be the most reactive of the three reagents here examined in most detail, it was decided to investigate its use in solid phase oligonucleotide synthesis. The efficacy of dibenzoyl tetrasulphide as a sulphur-transfer agent was compared with that of the recently-recommended tetraethylthiura disulphide in the solid phase synthesis of the nonadecadeoxyribonucleoside octadecaphosphorothioate , d[A(S)T(S)T(S)C(S)C(S)G(S)G(S)A(S)C(S)T(S)C(S)G(S)T(S)C(S)- C(S)A(S) C(S) C(S)A] . Solid phase synthesis was carried out on a 0.2 μmolar scale in an Applied Biosystems 381A DNA Synthesizer. 5'-0-(4,4'-dimethoxytrityl) -2'-deoxy- ribonucleoside 3 '-(2-cyanoethyl) -N,N-diisopropylphosphor- a idites with standard base protecting groups (i.e. benzoyl on adenine and cytosine, and isobutyryl on guanine) were used as building blocks, and lH-tetrazole was used as activating agent. Sulphur-transfer was effected in each of the 18 synthetic cycles either with (a) 0.4M dibenzoyl tetrasulphide in commercially-supplied, base-free 0.025% butylated-hydrσxytoluene(anti-peroxidation stabilizer) - containing tetrahydrofuran for 60 seconds or (b) 0.5M tetraethylthiuram disulphide in acetonitrile for 900 seconds as recommended by Vu and Hirschbein (loc c_it) . In both experiments (a) and (b) , the crude d[DMTrA(S)T(S)T(S)C(S) - C(S)G(S)G(S)A(S)C(S)T(S)C(S)G(S)T(S)C(S)C(S)A(S)C(S)C(S)A] was examined by ³¹P-NMR spectroscopy after ammonolytic release from the solid support and deacylation. In accompanying Figure la (relating to experiment (a) ) the ratio of the integrals of the signals at δ ca. 56.2 (assigned to phosphorothioate phosphorus) and δ ca . -0.3 (assigned to phosphate phosphorus) is estimated to be ca. 99:1, and in accompanying Figure lb (relating to experiment (b) ) , this ratio is estimated to be ca. 98:2. When treatment with 0.5M tetraethylthiuram disulphide was decreased to 450 seconds per synthetic cycle, the latter ratio fell to ca. 92.5:7.5.

It appears from the above results that the presently-preferred dibenzoyl tetrasulphide transfers sulphur at least twenty times as rapidly as tetraethylthiuram disulphide. On a 0.2 μmolar scale, this allowed the duration of the synthetic cycle to be decreased from ca. 23 to ca. 9 minutes. It was apparent from LC analysis (accompanying Figure 2) that the quality of the synthetic nonadecadeoxyribonucleoside octadecaphosphoro- thioate obtained was similar in both cases. The isolated yields obtained were also similar. It is possible that with the optimisation of conditions, solvents and sulphur- transfer agents, for example, cycle times may be further reduced and required concentrations reduced.

Preliminary results obtained in the synthesis of phosphorothioate analogues of oligoribonucleotides suggest that sulphur transfer occurs much more slowly and that dibenzoyl tetrasulphide is also a suitable reagent to use for this purpose. Hence, the present invention is particularly directed to the production of oligodeoxyribo- nucleotides.

Claims

Claims ; -

1. A process for the solid phase production of a phosphorothioate analogue of an oligo- or poly- nucleotide from phosphoramidite building blocks in the presence of a sulphur-transfer agent characterised in that the sulphur- transfer agent corresponds to the following general formula:

X - Sn„ - X

wherein n represents 3 or 4; and

X represents RC(0)- or RC(S)- wherein R represents alkyl, cycloalkyl or aryl,

(R 0)C(0)- wherein R' represents alkyl, (R₂N)C(0)- or (R₂N)C(S)- wherein R is as defined above and R₂NH may be a cyclic secondary amine, RS0₂- wherein R is as defined above,

(R0)₂P(0)- or (R0)₂P(S)- wherein R is as defined above.

2. A process as claimed in claim l wherein R represents aryl, which may be substituted.

3. A process as claimed in claim 1 or claim 2 wherein the sulphur-transfer agent is dibenzoyl tetra- or tri¬ sulphide.

4. A process as claimed in any of claims 1 to 3 wherein the oligo- or poly- nucleotide is an oligo- or poly- deoxyribonucleotide.