Table 3. Phosphitylation of various nucleosides with Py‚TFA
generation of 2′-deoxyphosphorothioate antisense olignucle-
otides, several second-generation antisense oligonucleotides
have entered Phase I clinical trials.22 Most of these oligo-
nucleotides contain a 2′-modification that provides increased
binding affinity for target RNA and enhanced stability in
vivo.23 To make larger quantities of these oligonucleotides,
one requires 2′-substituted amidites, such as 7 and 9.
Therefore we elected to extend the Py‚TFA activator
methodology to (i) 2′-O-TBDMS-, (ii) 2′-O-Me-, and (iii)
2′-O-(CH2)2OCH3-substituted nucleosides.
The data shown in Tables 1 and 2 demonstrate that
unmodified 2′-deoxynucleoside phosphoramidites can be
easily synthesized under optimal conditions described in this
report (See Experimental Section). Almost identical condi-
tions were then successfully applied to the transformation
of 2′-O-Me nucleosides 2 f 7 in ∼75-94% yield, and 2′-
O-methoxyethyl nucleosides 4 f 9 in ∼92-96% yield
(Table 3). The phosphitylation conditions using Py‚TFA as
an activator were further extended to the corresponding
ribonucleosides protected with 2′-O-TBDMS group. Thus,
ribonucleosides (3) were converted into amidites 8 in
excellent yield (82-95%, entries 7-10, Table 3). It is
noteworthy that all of the reactions described in Table 3 were
carried out in CH2C12 as a solvent and found to be complete
in under 3 h at room temperature, thus making it efficient
and convenient for further scale-up work.
Scale-up Trials and Interpretation. With the phosphi-
tylation procedures fully optimized, the scale-up trials were
performed on 10 mmol scale with 2′-deoxynucleosides (1).
The conventional method4 of phosphitylation involves con-
siderable skill in work-up after the reaction, isolation, and
purification of amidites. Furthermore there are problems
associated with storage and handling of reactive P(III)
phosphoramidites,24 such as compounds 6-9. Therefore we
have attempted to develop a universal process for phosphi-
tylation which is simpler and works with a variety of
nucleosides. A detailed procedure is described in the
Experimental Section, and the highlights are summarized
below. The simplicity of this procedure lies in eliminating
the traditional aqueous work-up step for two reasons. First,
any exposure of water to phosphoramidite is bound to
hydrolyze a portion of the product to H-phosphonate,
resulting in lower yield. Second, removal of traces of water
is not only difficult but also cost- and labor-intensive. The
excess of the reagents and byproducts are conveniently
separated from the product by short silica gel flash column
chromatography. Since the byproducts are inert and nonre-
active, their coexistence with the product during chroma-
tography is not an issue. In addition, the presence of traces
of “free” pyridine avoids accidental cleavage of 5′-O-DMT
group from amidites (6-9). This nonaqueous work-up
activator
reaction
time (h)
entry
% yielda
St.Mat.fPdtb
1
2
3
4
5
6
7
8
9
10
11
12
13
14
1
2
<1
2
2
2
<1
2
2
2
<1
2
3
2
60
91
75
94
86
88
95
90
82
84
96
94
95
92
1gf6g
1ff6f
2af7a
2bf7b
2cf7c
2ef7e
3af8a
3bf8b
3cf8c
3ef8e
4af9a
4cf9c
4df9d
4ff9
a,b See Table 1 footnotes for details.
experiment confirmed that the use of Py‚TFA as an activator
will not cause depurination of 1a during phosphitylation.
With the successful use of Py‚TFA as an activator for
phosphitylation of 2′-deoxynucleosides, next we turned our
attention to the application of this methodology for the
synthesis of other important nucleosidic phosphoramidites.
Recently, Hayakawa et al. reported the use of the base-
unprotected phosphoramidite of 2′-deoxyadenosine and its
application in automated synthesis of oligonucleotides.16 It
was intriguing for us to see if our methodology was able to
convert unprotected 1g into 6g. Phosphitylation of 1g with
Py‚TFA as an activator was quick and gave 6g in 60%
isolated yield with a small amount of unidentified products
(Table 3, entry 1). Incorporation of base- or sugar-modified
2′-deoxynucleoside residue into oligonucleotide is a well-
established art.17 One of these modifications, 5-methyl-2′-
deoxycytidine (5-Me-dC, 1f) was recently used in an
antisense oligonucleotide that entered Phase I clinical trials.18
The use of 5-MedC in place of unmodified 2′-deoxycytidine
in an antisense oligonucleotide is anticipated to increase the
affinity for target RNA.19 Therefore, it was pertinent for us
to undertake phosphitylation of 5-Me-dC. Reaction of 1f with
bis-reagent in the presence of Py‚TFA in CH2C12 gave 6f
in excellent yield (Table 3, entry 2).
With the initiation of Phase I clinical trials20 with
Angiozyme, the first synthetic ribozyme molecule, and
increasing demand for large-scale synthesis of RNA,21 it was
necessary to have improved methods for the preparation of
ribophosphoramidites (8a-c,e). In addition to the first-
(16) Hayakawa, Y.; Masanori, K. J. Am. Chem. Soc. 1998, 120, 12395.
(17) Sanghvi, Y. S., Cook, P. D., Eds. Carbohydrate Modification in Antisense
Research; ACS Symposium Series 580; American Chemical Society:
Washington, DC, 1994.
(18) Pramik, M. J. Genet. Eng. News 1999, June 15, 9.
(22) Persidis, A. Nature Biotech. 1999, 17, 403.
(19) Sanghvi, Y. S.; Hoke, G. D.; Freier, S. M.; Zounes, M. C.; Gonzalez, C.;
Cummins, L.; Sasmor, H.; Cook, P. D. Nucleic Acids Res. 1993, 21, 3197.
(20) Sandberg, J. A.; Bouhana, K. S.; Gallegos, A. M.; Agrawal, A. B.; Grimm,
S. L.; Wincott, F. E.; Reynolds, M. A.; Pavco, P. A.; Parry, T. J. Antisense
Nucleic Acid Drug DeVelop. 1999, 9, 271.
(21) Wincott, F. E.; Usman, N. Ribozyme Protocols. In Methods in Molecular
Biology; Turner, P. C., Ed.; Humana Press: Totowa, New Jersey; Vol 74,
Chapter 8, p 59.
(23) Cook, D. In Antisense Research and Application; Crooke, S. T. Ed.;
Sprinter: Berlin, Germany, 1998; Chapter 2, p 51.
(24) Phosphoramidites are stored as free-flowing powder under a blanket of inert
atmosphere in glass bottles at room temperature (16-25 °C). These
compounds must be used within 2-4 weeks after dissolution in acetonitrile.
Particularly, the shelf life of dG amidite 1c is shortest (∼2 weeks) among
other 2′-deoxyamidites (1). All amidites are hygroscopic and should be
handled appropriately.
Vol. 4, No. 3, 2000 / Organic Process Research & Development
•
179