Thermolytic Properties of 3-(2-Pyridyl)-1-propyl and 2-[N-Methyl-N-(2-pyridyl)]aminoethyl Phosphate/Thiophosphate Protecting Groups in Solid-Phase Synthesis of Oligodeoxyribonucleotides

Article abstract of DOI:10.1021/jo0354490

Thermolytic groups may serve as alternatives to the conventional 2-cyanoethyl group for phosphate/ thiophosphate protection in solid-phase oligonucleotide synthesis to prevent DNA alkylation by acrylonitrile generated under the basic conditions used for oligonucleotide deprotection. Additionally, thermolytic groups are attractive in the context of engineering a "heat-driven" process for the synthesis of oligonucleotides on diagnostic microarrays. In these regards, the potential application of pyridine derivatives as thermolytic phosphate/thiophosphate protecting groups has been investigated. Specifically, 2-pyridinepropanol and 2-[N-methyl-N-(2-pyridyl)]aminoethanol were incorporated into deoxyribonucleoside phosphoramidites 7a-d and 9, which were found as efficient as 2-cyanoethyl deoxyribonucleoside phosphoramidites in solid-phase oligonucleotide synthesis. Whereas the removal of 3-(2-pyridyl)-1-propyl phosphate/thiophosphate protecting groups from oligonucleotides is effected within 30 min upon heating at 55 °C in concentrated NH₄OH or in an aqueous buffer at pH 7.0, cleavage of 2-[N-methyl-N-(2-pyridyl)]aminoethyl groups occurs spontaneously when their phosphate or phosphorothioate esters are formed during oligonucleotide synthesis. The deprotection of these groups follows a cyclodeesterification process generating the bicyclic salts 13 and 14 as side products. These salts do not alkylate or otherwise modify any DNA nucleobases and do not desulfurize a phosphorothioate diester model under conditions mimicking large-scale oligonucleotide deprotection.

Full text of DOI:10.1021/jo0354490

Th er m olytic P r op er ties of 3-(2-P yr id yl)-1-p r op yl a n d
2-[N-Meth yl-N-(2-p yr id yl)]a m in oeth yl P h osp h a te/Th iop h osp h a te
P r otectin g Gr ou p s in Solid -P h a se Syn th esis of
Oligod eoxyr ibon u cleotid es
J acek Cies´lak† and Serge L. Beaucage*
Division of Therapeutic Proteins, Center for Biologics Evaluation and Research,
Food and Drug Administration, 8800 Rockville Pike, Bethesda, Maryland 20892
beaucage@cber.fda.gov
Received October 2, 2003
Thermolytic groups may serve as alternatives to the conventional 2-cyanoethyl group for phosphate/
thiophosphate protection in solid-phase oligonucleotide synthesis to prevent DNA alkylation by
acrylonitrile generated under the basic conditions used for oligonucleotide deprotection. Additionally,
thermolytic groups are attractive in the context of engineering a “heat-driven” process for the
synthesis of oligonucleotides on diagnostic microarrays. In these regards, the potential application
of pyridine derivatives as thermolytic phosphate/thiophosphate protecting groups has been
investigated. Specifically, 2-pyridinepropanol and 2-[N-methyl-N-(2-pyridyl)]aminoethanol were
incorporated into deoxyribonucleoside phosphoramidites 7a -d and 9, which were found as efficient
as 2-cyanoethyl deoxyribonucleoside phosphoramidites in solid-phase oligonucleotide synthesis.
Whereas the removal of 3-(2-pyridyl)-1-propyl phosphate/thiophosphate protecting groups from
oligonucleotides is effected within 30 min upon heating at 55 °C in concentrated NH4OH or in an
aqueous buffer at pH 7.0, cleavage of 2-[N-methyl-N-(2-pyridyl)]aminoethyl groups occurs spontane-
ously when their phosphate or phosphorothioate esters are formed during oligonucleotide synthesis.
The deprotection of these groups follows a cyclodeesterification process generating the bicyclic salts
13 and 14 as side products. These salts do not alkylate or otherwise modify any DNA nucleobases
and do not desulfurize a phosphorothioate diester model under conditions mimicking large-scale
oligonucleotide deprotection.
In tr od u ction
2-cyanoethyl phosphate/thiophosphate protecting group,4
which occurs under basic conditions via elimination of
acrylonitrile, a potent carcinogen known to alkylate the
nucleobases of nucleosides and nucleic acids.3a By com-
parison, deprotection of the thermolabile groups inves-
tigated so far for phosphate/thiophosphate protection has
not generated mutagenic side products.2,3 The use of these
phosphate/thiophosphate protecting groups has therefore
been recommended for large-scale syntheses of alkyla-
tion-free therapeutic oligonucleotides.
Although the discovery of thermolytic phosphate/
thiophosphate protecting groups may be valuable in
large-scale preparations of oligonucleotide drugs, it may
also lead to the development of thermolabile 5′-/3′-
hydroxyl protecting groups in the synthesis of oligonucle-
otides on diagnostic microarrays. In this context, given
that ketoalkyl and amidoalkyl phosphate/thiophosphate
protecting groups (shown as 1 and 2) exhibited striking
thermolytic deprotection properties,2 we rationalized that
In recent years we have been investigating cyclic
N-acylphosphoramidites in the solid-phase synthesis of
DNA oligonucleotides and their phosphorothioate ana-
logues.1 This method led to the discovery of several
thermolabile phosphate/thiophosphate protecting groups.1,2
Interestingly, the thermolytic cleavage of these protecting
groups proceeds through an intramolecular cyclodees-
terification mechanism reminiscent to that of the base-
assisted deprotection of 4-[N-(2,2,2-trifluoroacetyl)amino]-
butyl and 4-[N-methyl-N-(2,2,2-trifluoroacetyl)amino]butyl
phosphate protecting groups described earlier by us.3
Such a deprotection mechanism departs from that of the
* To whom correspondence should be addressed. Tel: (301) 827-
5162. Fax: (301) 480-3256.
† On leave from the Institute of Bioorganic Chemistry, Polish
Academy of Sciences, Poznan, Poland.
(1) Wilk, A.; Grajkowski, A.; Phillips, L. R.; Beaucage, S. L. J . Am.
Chem. Soc. 2000, 122, 2149-2156.
(2) (a) Grajkowski, A.; Wilk, A.; Chmielewski, M. K.; Phillips, L. R.;
Beaucage, S. L. Org. Lett. 2001, 3, 1287-1290. (b) Wilk, A.; Chmielews-
ki, M. K.; Grajkowski, A.; Phillips, L. R.; Beaucage, S. L. Tetrahedron
Lett. 2001, 42, 5635-5639. (c) Wilk, A.; Chmielewski, M. K.; Gra-
jkowski, A.; Phillips, L. R.; Beaucage, S. L. J . Org. Chem. 2002, 67,
6430-6438. (d) Wilk, A., Chmielewski, M. K.; Grajkowski, A.; Phillips,
L. R.; Beaucage, S. L. In Current Protocols in Nucleic Acid Chemistry;
Beaucage, S. L., Bergstrom, D. E., Glick, G. D., J ones, R. A., Eds.; J ohn
Wiley & Sons: New York, 2002; pp 3.9.1-3.9.15.
(3) (a) Wilk, A.; Grajkowski, A.; Phillips, L. R.; Beaucage, S. L. J .
Org. Chem. 1999, 64, 7515-7522. (b) Wilk, A.; Grajkowski, A.;
Chmielewski, M. K.; Phillips, L. R.; Beaucage, S. L. In Current
Protocols in Nucleic Acid Chemistry; Beaucage, S. L., Bergstrom, D.
E., Glick, G. D., J ones, R. A. Eds.; J ohn Wiley & Sons: New York,
2001; pp 2.7.1-2.7.12. (c) Wilk, A.; Srinivasachar, K.; Beaucage, S. L.
J . Org. Chem. 1997, 62, 6712-6713.
(4) Tener, G. M. J . Am. Chem. Soc. 1961, 83, 159-168.
10.1021/jo0354490 This article not subject to U.S. Copyright. Published 2003 American Chemical Society
Published on Web 12/03/2003
J . Org. Chem. 2003, 68, 10123-10129 10123

Cies´lak and Beaucage
Resu lts a n d Discu ssion
Synthesis of phosphoramidites 7a -d is achieved by
condensing each of the commercially available deoxyri-
bonucleosides 5a -d with phosphorodiamidite 6 in the
presence of 1H-tetrazole5 in anhydrous dichloromethane
(Figure 1A). The phosphorodiamidite 6 is obtained from
the reaction of 2-pyridinepropanol with bis(N,N-diiso-
propylamino)chlorophosphine generated in situ upon
mixing phosphorus trichloride with an excess N,N-
diisopropylamine in dry benzene.2 The preparation of
phosphoramidite 9 is accomplished through the conden-
sation of 5a with phosphorodiamidite 8 (Figure 1B) in a
manner similar to that described for the synthesis of
phosphoramidites 7a -d . As for 6, phosphorodiamidite 8
is generated from bis(N,N-diisopropylamino)chlorophos-
phine and 2-[N-methyl-N-(2-pyridyl)]aminoethanol, which
is obtained by heating 2-bromopyridine and 2-methyl-
aminoethanol as reported in the literature.6 The crude
deoxyribonucleoside phosphoramidites 7a -d and 9 are
purified by silica gel chromatography and are isolated
in yields ranging from 70% to 80%. Identity of these
phosphoramidites is confirmed by 31P NMR spectroscopy
and high-resolution mass spectrometry. The phosphora-
midites 7a and 9 are then used in the solid-phase
synthesis of dinucleotides to independently assess the
thermolytic deprotection kinetics of both 3-(2-pyridyl)-1-
propyl and 2-[N-methyl-N-(2-pyridyl)]aminoethyl phos-
phate protecting groups through reversed-phase high
performance liquid chromatography (RP-HPLC) analy-
ses.
2-pyridinylalkyl and 2-[N-methyl-N-(2-pyridyl)]aminoalkyl
phosphate/thiophosphate protecting groups (shown as 3
and 4) may significantly accelerate cyclodeesterification
deprotection reactions, considering the inherent nucleo-
philicity of pyridine derivatives.
Th er m olytic P r op er ties of th e 3-(2-P yr id yl)-1-
p r op yl a n d 2-[N-Meth yl-N-(2-p yr id yl)]a m in oeth yl
Gr ou p s a s P h osp h a te/Th iop h osp h a te P r otectin g
Gr ou p s for DNA Oligon u cleotid es. Treatment of the
solid-phase-linked dinucleoside phosphotriester 10 (Fig-
ure 2) with pressurized methylamine gas for 3 min at 25
°C, followed by washing of the support with 0.1 M
triethylammonium acetate (TEAA, pH 7.0)/MeCN (3:2
v/v), afforded a mixture of dinucleotides composed of
phosphotriester 11 (40%) and phosphodiester TpT (12,
60%) on the basis of RP-HPLC analysis of the eluates. A
longer exposure of 10 to pressurized methylamine gas
(30 min, 25 °C) did not change the ratio of 11 and 12.
Complete removal of the 3-(2-pyridyl)-1-propyl group
from 11 is accomplished within 30 min (t1/2 ∼225 s) or 5
min (t1/2 ∼40 s) upon heating the eluates at 55 or 90 °C,7
respectively. By comparison, treatment of 10 with con-
centrated NH4OH for 30 min at 25 °C followed by heating
the ammoniacal solution for an additional 30 min at 55
°C also led to complete cleavage of the 3-(2-pyridyl)-1-
propyl group from 11, as 12 was the only nucleotidic
species detected by RP-HPLC analysis of the deprotection
reaction.
In the absence of NH4OH, the rates at which the 3-(2-
pyridyl)-1-propyl group is removed from 11 in aqueous
buffers are pH-dependent. Specifically, when a citric acid
buffer (pH 4.0) is used instead of 0.1 M TEAA (pH 7.0)/
MeCN (3:2 v/v), removal of the 3-(2-pyridyl)-1-propyl
phosphate protecting group is 99% complete only after
heating for 60 min at 55 °C. The slower deprotection
rates, relative to those determined at pH 7.0 (100%, 30
min, 55 °C), are likely due to the decreased nucleophi-
F IGURE 1. Synthesis of deoxyribonucleoside phosphoramid-
ites 7a -d and 9. Keys: DMTr, 4,4-dimethoxytrityl; Thy,
thymin-1-yl.
We set out to investigate this rationale through the
preparation of deoxyribonucleoside phosphoramidites
7a -d (Figure 1A) and 9 (Figure 1B), and incorporation
of these monomers into oligonucleotides.
10124 J . Org. Chem., Vol. 68, No. 26, 2003

Thermolytic Phosphate/ Thiophosphate Protecting Groups
Syn th esis a n d Ch a r a cter iza tion of Cyclod eester i-
fica tion P r od u cts. On the basis of our previous experi-
ence with cyclodeesterification of thermolytic phosphate/
thiophosphate protecting groups,1,2 formation of the
bicyclic pyridinium salt 13 is expected from the thermal
deprotection of 10 and 11. Indeed, RP-HPLC analysis of
the deprotection reaction revealed, aside from 12 (tR
)
14.7 min), a new fast-eluting compound (tR ) 3.4 min).8
This compound was collected and characterized by high-
resolution mass spectrometry. The accurate mass found
for the fast-eluting material is consistent with that
expected for 13 (see Experimental Section). To further
corroborate the identity of 13, its synthesis is achieved
by heating a solution of 2-pyridinepropanol and trifluo-
roacetic anhydride (TFAA) in MeCN (1:5:5 v/v/v). RP-
HPLC analysis of the reaction product showed a single
peak exhibiting a retention time (3.4 min) identical to
that recorded for 13 under the same chromatographic
conditions. 1H and 13C NMR spectra of the reaction
product are in agreement with the structure proposed
for 13 (see Experimental Section).
When the 3-(2-pyridyl)-1-propyl phosphate protecting
group in 10 is replaced with the 2-[N-methyl-N-(2-
pyridyl)]aminoethyl group, the cyclodeesterification reac-
tion product 14 is immediately formed and washed off
the support. Short exposure of the support-linked di-
nucleotide to pressurized methylamine gas leaves 12 as
the only species detectable by RP-HPLC.
F IGURE 2. Thermolytic cleavage of the 3-(2-pyridyl)-1-propyl
phosphate protecting group from a dinucleotide. Conditions:
(i) MeNH2 gas (2.5 bar), 3 min, 25 °C or concentrated NH4-
OH, 30 min, 25 °C; (ii) 0.1 M TEAA (pH 7.0)/MeCN (3:2 v/v),
30 min, 55 °C or concentrated NH4OH, 30 min, 55 °C. Keys:
Thy, thymin-1-yl; LCAA-CPG, succinyl long chain alkylamine
controlled-pore glass.
licity of the partially protonated pyridyl group at pH 4.0
inhibiting the cyclodeesterification reaction (Figure 2).
Replacement of the 3-(2-pyridyl)-1-propyl group with
the 2-[N-methyl-N-(2-pyridyl)]aminoethyl group for phos-
phate protection in 10 and exposure of the support-bound
dinucleoside phosphotriester to pressurized methylamine
gas for 3 min result in the quantitative formation of 12
according to RP-HPLC analysis of the deprotection
reaction. This result is consistent with the increased
nucleophilicity of the pyridyl group caused by the electron-
donating 2-dialkylamino function, which accelerates the
cyclodeesterification reaction leading to the formation of
12. Because it has been demonstrated that exposure of
10 to pressurized methylamine for either 3 or 30 min has
no significant effect on the production of 12 (vide supra),
it would appear that the cleavage of both 3-(2-pyridyl)-
1-propyl and 2-[N-methyl-N-(2-pyridyl)]aminoethyl phos-
phate protecting groups begins when the phosphite
triester is oxidized to the corresponding phosphate tri-
ester during solid-phase oligonucleotide synthesis. Par-
ticularly noteworthy is the complete removal of the 2-[N-
methyl-N-(2-pyridyl)]aminoethyl phosphate protecting
group during the oxidation reaction (vide infra). Thermal
deprotection of the 3-(2-pyridyl)-1-propyl and 2-[N-meth-
yl-N-(2-pyridyl)]aminoethyl groups from dinucleoside
phosphorothioate triesters under conditions identical to
those used for the parent dinucleoside phosphotriesters
(Figure 2) proceeds, as expected, with similar kinetics
(data not shown). No other thermolytic phosphate/thio-
phosphate protecting groups reported earlier by us
exhibited such a rapid deprotection kinetics.
Thus, a strategy different than that used for isolating
13 had to be devised to capture and identify 14. Specif-
ically, condensation of the phosphoramidite 9 with a
slight excess of 3′-O-acetylthymidine and 1H-tetrazole in
anhydrous MeCN afforded the dinucleoside phosphite
triester product in essentially quantitative yield accord-
ing to 31P NMR analysis of the reaction mixture, which
displayed two characteristic signals at ∼144 ppm. Addi-
tion of solid 3H-1,2-benzodithiol-3-one 1,1-dioxide to the
solution rapidly converted the phosphite triester to the
parent dinucleoside phosphorothioate diester character-
ized by two 31P NMR signals at ∼60 ppm. These signals
confirmed the rapid formation of 14 given the absence of
31P NMR signals corresponding to the dinucleoside phos-
phorothioate triester expected at ∼70 ppm. RP-HPLC
analysis of the solution validated the presence of 14 as a
peak displaying a retention time (4.9 min) comparable
to that of 13 (tR ) 3.4 min) under identical chromato-
graphic conditions. The fast-eluting material was col-
lected and characterized by high-resolution mass spec-
trometry. The results of accurate mass determination are
in accordance with the mass expected for 14 (see Experi-
mental Section). Identity of the bicyclic salt is further
established through its chemical synthesis by heating a
solution of 2-[N-methyl-N-(2-pyridyl)]aminoethanol and
TFAA in MeCN, as described above for the preparation
of 13. The reaction product shows as a single RP-HPLC
peak (tR ) 4.9 min) identical to that of 14 under the same
1
chromatographic conditions. H and 13C NMR spectra of
J . Org. Chem, Vol. 68, No. 26, 2003 10125

Cies´lak and Beaucage
Given the structural similarities of phosphoramidites
9 and 7a -d in regard to phosphorus protection, prepara-
tion of the remaining three 2-[N-methyl-N-(2-pyridyl)]-
aminoethyl deoxyribonucleoside phosphoramidites to as-
sess their suitability for solid-phase oligonucleotide syn-
thesis seems unnecessary. Instead, the site-specific in-
corporation of 9 depicted as T* into the DNA oligonucle-
otide sequence d(AT*T*CGT*AGCT*AAGGT*CAT*GC)
and into that of its phosphorothioated analogue should
be sufficient to demonstrate the comparability of these
oligonucleotides with those synthesized using 7a -d or
2-cyanoethyl deoxyribonucleoside phosphoramidites in
terms of ease of synthesis and deprotection, purity, and
yields. Such a comparability study is validated on the
basis of RP-HPLC and PAGE data (shown as Supporting
Information). In addition, native DNA oligonucleotides
prepared via 7a -d , 9, and 2-cyanoethyl deoxyribonucleo-
side phosphoramidites were exposed to bacterial alkaline
phosphatase and snake venom phosphodiesterase to
assess any nucleobase modifications that might have
occurred through the use of these phosphoramidites. RP-
HPLC analysis of the enzymatic digests shows no detect-
able nucleobase modification and indicates that the use
of 7a -d or 9 (data shown as Supporting Information)
produces oligonucleotides of quality comparable to those
prepared using 2-cyanoethyl deoxyribonucleoside phos-
phoramidites.
We have earlier reported that the concentration of
acrylonitrile that is generated from base-assisted â-
elimination of 2-cyanoethyl phosphate/thiophosphate
groups, under conditions mimicking large-scale oligo-
nucleotide deprotection (>50 mmol), converted thymidine
to N3-(2-cyanoethyl)thymidine to the extent of 11%.3a
Because large-scale preparations of therapeutic oligo-
nucleotides are required for clinical studies, it becomes
imperative to investigate whether oligonucleotides syn-
thesized via 7a -d or 9 would undergo nucleobase alky-
lation under deprotection conditions comparable to those
used for large-scale syntheses of therapeutic oligonucle-
otides.
F IGURE 3. Polyacrylamide gel electrophoresis analysis of
d(ATCCGTAGCTAAGGTCATGC) and its phosphorothioated
analogue under denaturing conditions (7 M urea, 1X TBE
buffer, pH 8.3). Left lane: crude oligomer synthesized from
commercial 2-cyanoethyl deoxyribonucleoside phosphoramid-
ites and deprotected by treatment with concentrated NH4OH
for 10 h at 55 °C. Middle lane: crude oligomer synthesized
from 1-week-old solutions of 7a -d and deprotected under
conditions identical to those used for the 20-mer shown in the
left lane. Right lane: crude phosphorothioated oligomer syn-
thesized from fresh solutions of 7a -d and deprotected under
conditions identical to those used for the 20-mer shown in the
left lane. Unmodified oligonucleotides are visualized as blue
bands and fully phosphorothioated oligonucleotides as purple
bands, upon staining the gel with Stains-all. Bromophenol blue
is used as a marker and shows as a large band in each lane at
the bottom of the gel.
the reaction product agree with the structure proposed
for 14 (see Experimental Section) and support an opera-
tive cyclodeesterification processs for the removal of 2-[N-
methyl-N-(2-pyridyl)]aminoethyl phosphate/thiophos-
phate protecting groups.
Syn th esis a n d Ch a r a cter iza tion of Oligod eoxyr i-
b on u cleot id es. Automated solid-phase synthesis of
d(ATCCGTAGCTAAGGTCATGC) and that of its fully
phosphorothioated analogue9 is performed using fresh
and 1-week-old solutions of 7a -d for the purpose of
comparing coupling efficiency and stability of the phos-
phoramidites in solution with that of commercial 2-cya-
noethyl deoxyribonucleoside phosphoramidites.
The solid-phase-linked oligonucleotides are released
from the support by treatment with concentrated NH4-
OH, and the eluates are heated for 10 h at 55 °C to
complete nucleobase and phosphate/thiophosphate depro-
tection. The crude oligonucleotides are compared with
each other using RP-HPLC and polyacrylamide gel
electrophoresis (PAGE) techniques. Figure 3 shows that
synthesis of DNA oligonucleotides using phosphoramid-
ites 7a -d is as efficient as that achieved with commercial
2-cyanoethyl deoxyribonucleoside phosphoramidites when
evaluating the relative intensity of bands corresponding
to shorter than full-length sequences on the gel. RP-
HPLC chromatograms of the crude oligonucleotides are
consistent with PAGE data (see Supporting Information).
Thus, thymidine, 2′-deoxycytidine, 2′-deoxyadenosine,
2′-deoxyguanosine, N4-benzoyl-2′-deoxycytidine, N6-ben-
zoyl-2′-deoxyadenosine, and N2-isobutyryl-2′-deoxygua-
nosine are mixed with the 2,2,2-trifluoroacetate salt of
13 or that of 14 in NH4OH in concentrations representa-
tive to those existing in large-scale (>50 mmol) oligo-
nucleotide deprotection and heated for 10 h at 55 °C.10
The simulation reaction involving either 13 or 14 is then
analyzed by RP-HPLC to assess the extent of nucleobase
modification. No nucleobase alkylation products or other
nucleoside modifications are detected under these condi-
tions, thus indicating that the use of phosphoramidites
7a -d and/or 9 and its congeners is recommended over
2-cyanoethyl deoxyribonucleoside phosphoramidites for
the large-scale preparation of alkylation-free therapeutic
oligonucleotides.
Because small-scale (0.2 µmol) solid-phase synthesis
of phosphorothioated oligonucleotides via 7a -d or 9
proceeds without significant desulfurization during either
chain assembly or oligonucleotide deprotection, as judged
by 31P NMR analysis of fully deprotected phosphorothio-
ated oligonucleotides (see Supporting Information), it still
(5) (a) Lee, H.-J .; Moon, S.-H. Chem. Lett. 1984, 1229-1232. (b)
Barone, A. D.; Tang, J .-Y.; Caruthers, M. H. Nucleic Acids Res. 1984,
12, 4051-4061.
(6) Weiner, N.; Kaye, I. A. J . Org. Chem. 1949, 14, 9353-9372.
10126 J . Org. Chem., Vol. 68, No. 26, 2003

Thermolytic Phosphate/ Thiophosphate Protecting Groups
remains to demonstrate whether 13 or 14 can induce
desulfurization of phosphorothioated oligonucleotides
when produced in concentrations similar to those pre-
vailing under large-scale oligonucleotide deprotection
conditions. To confirm or reject this possibility, the
potassium salt of O,O-diethyl thiophosphate is used as
a model and is mixed with the 2,2,2-trifluoroacetate salt
of 13 or that of 14 under conditions simulating large-
scale oligonucleotide deprotection (see Experimental Sec-
tion). 31P NMR analysis of the reactions did not reveal
any conversion of O,O-diethyl thiophosphate (δP ∼55
ppm) to O,O-diethyl phosphate (δP ∼2 ppm) caused by
either 13 or 14 (see Supporting Information). It is
therefore likely that 13 or 14 will not induce desulfur-
ization of thioated oligonucleotides produced on a large-
scale process.
decrease exposure of both oligonucleotides and glass
surface to the harsh chemical reagents that are currently
employed in conventional solid-phase oligonucleotide
synthesis. The 3-(2-pyridyl)-1-propyl and 2-[N-methyl-
N-(2-pyridyl)]aminoethyl groups for phosphate/thio-
phospate protection thus add to the repertoire of pro-
tecting groups exhibiting thermolytic properties.
Exp er im en ta l Section
O-[3-(2-P yr id yl)-1-p r op yl] N,N,N′,N′-Tet r a isop r op y-
lp h osp h or od ia m id ite (6). 2-Pyridinepropanol (3.1 mL, 24
mmol) is added by syringe to a stirred solution of bis(N,N-
diisopropylamino)chlorophosphine2c,d (20 mmol) in dry benzene
(100 mL) at 25 °C. 31P NMR analysis of the reaction mixture
indicates that bis(N,N-diisopropylamino)chlorophosphine (δP
135.5 ppm) is converted to the corresponding phosphorodia-
midite (δP 123.3 ppm) within 2 h. The suspension is filtered,
and the filtrate is evaporated under reduced pressure to afford
an oil, which is used without further purification in the
preparation of 7a -d . 1H NMR (300 MHz, CDCl3): δ 8.51 (ddd,
J ) 1.0, 1.8, 4.9 Hz, 1H), 7.56 (ddd, J ) 1.8, 7.6, 7.8 Hz, 1H),
7.15 (m, 1H), 7.07 (ddd, J ) 1.0, 4.9, 7.6 Hz, 1H), 3.60 (dt,
3J PH ) 7.3 Hz, J ) 6.3 Hz, 2H), 3.54 (sept, J ) 6.8 Hz, 2H),
3.50 (sept, J ) 6.8 Hz, 2H), 2.89 (m, 2H), 2.02 (m, 2H), 1.16
(d, J ) 6.8 Hz, 12H), 1.15 (d, J ) 6.8 Hz, 12H). 13C NMR (75
Con clu sion
We have demonstrated that the 3-(2-pyridyl)-1-propyl
and 2-[N-methyl-N-(2-pyridyl)]aminoethyl groups are
appealing phosphate/thiophosphate protecting groups for
solid-phase DNA oligonucleotide synthesis. These pro-
tecting groups are cleaved from DNA oligonucleotides
under the mildest conditions ever applied to thermolytic
phosphate/thiophosphate protecting groups, whether in
the presence or absence of concentrated NH4OH. Like all
thermolytic phosphate/thiophosphate protecting groups
we have studied so far, removal of the 3-(2-pyridyl)-1-
propyl and 2-[N-methyl-N-(2-pyridyl)]aminoethyl groups
from oligonucleotides follows a cyclodeesterification path-
way with the concomitant formation of bicyclic salt side
products, which are innocuous to both DNA nucleobases
and phosphorothioate diester groups. It should be pointed
out that exposure of these functional groups to the
bicyclic salts is minimal during oligonucleotide depro-
tection. Indeed, cleavage of the 3-(2-pyridyl)-1-propyl or
2-[N-methyl-N-(2-pyridyl)]aminoethyl phosphate/thio-
phosphate protecting group is ∼60% or 100% complete,
respectively, after each oxidative conversion of an inter-
nucleosidic phosphite triester linkage to its phosphate/
thiophosphate counterpart throughout oligonucleotide
assembly. Consequently, the bicyclic salts 13 and 14 are
washed off the solid support as they are generated,
thereby minimizing potential nucleobase modification
and desulfurization of phosphorothioate diesters. This
feature would certainly be valuable in the large-scale
preparation of therapeutic DNA oligonucleotides as it
would preserve genetic fidelity and maintain the inherent
ability of thioated oligonucleotides to resist ubiquitous
nucleases. It is also very likely that the 3-(2-pyridyl)-1-
propyl or 2-[N-methyl-N-(2-pyridyl)]aminoethyl phosphate/
thiophosphate protecting group may find applications in
the solid-phase synthesis of oligoribonucleotides, since
these oligomers are currently in high demand considering
the biomedical significance of short interfering double-
stranded RNAs11 in silencing gene expression. It is worth
restating that the use of thermolytic groups for phosphate
and hydroxyl protection in oligonucleotide synthesis on
planar glass surfaces is very attractive because it would
3
MHz, CDCl3): δ 23.7, 23.8, 24.6, 24.7, 31.6 (d, J PC ) 9.5 Hz),
2
35.1, 44.2, 44.3, 63.6 (d, J PC ) 21.2 Hz), 120.9, 122.8, 136.1,
149.3, 162.1. 31P NMR (121 MHz, CDCl3): δ 124.6.
O-[2-[N-Meth yl-N-(2-p yr id yl)]a m in oeth yl] N,N,N′,N′-
Tetr a isop r op ylp h osp h or od ia m id ite (8). This phosphiny-
lating agent is prepared under conditions identical to those
described for the preparation of 6. 1H NMR (300 MHz,
CDCl3): δ 8.12 (ddd, J ) 0.9, 2.0, 4.9 Hz, 1H), 7.39 (ddd, J )
2.0, 7.1, 8.8 Hz, 1H), 6.49 (m, 2H), 3.73 (m, 4H), 3.51 (sept, J
) 6.8 Hz, 2H), 3.48 (sept, J ) 6.8 Hz, 2H), 3.1 (s, 3H), 1.14 (d,
J ) 6.8 Hz, 12H), 1.12 (d, J ) 6.8 Hz, 12H). 13C NMR (75 MHz,
CDCl3): δ 23.77, 23.83, 24.4, 24.5, 37.2, 44.2, 44.4, 51.4 (d,
2
3J PC ) 9.6 Hz), 61.9 (d, J PC ) 20.3 Hz), 105.6, 111.2, 128.2,
136.8, 147.8, 158.5. 31P NMR (121 MHz, CDCl3): δ 123.8.
Gen er a l P r oced u r e for P r ep a r a tion of Deoxyr ibo-
n u cleosid e P h osp h or a m id ites 7a -d . Crude 6 (700 mg, ∼2
mmol) is added by syringe under an inert atmosphere to a
stirred solution of 5a -d (2.2 mmol) in dry CH2Cl2 (10 mL).
Sublimed 1H-tetrazole (112 mg, 1.6 mmol) is then added,
portionwise, under a positive pressure of argon to the solution.
The reaction is monitored by 31P NMR spectroscopy until 6 is
all reacted (<2 h). Triethylamine (1 mL) is added to the
solution, which is then immediately evaporated under reduced
pressure. The material left is purified by silica gel chroma-
tography using benzene/triethylamine (9:1 v/v) as the eluent.
Fractions containing the product are pooled together and
rotoevaporated under reduced pressure to a white foam. A
solution of the foamy material in dry benzene is frozen and
lyophilized under high vacuum to afford pure 7a -d as white
powders in yields ranging from 70% to 80%.
5′-O-(4,4′-Dim eth oxytr ityl)-3′-O-(N,N-diisopr opylam in o)-
[3-(2-p yr id yl)-1-p r op yloxy]p h osp h in yl-2′-d eoxyt h ym i-
d in e (7a ). 31P NMR (121 MHz, CDCl3): δ, 147.1, 147.6. FAB-
HRMS: calcd for C45H55N4O8P (M + Cs)+ 943.2819, found
943.2812.
N4-Ben zoyl-5′-O-(4,4′-d im eth oxytr ityl)-3′-O-(N,N-d iiso-
p r op yla m in o)-[3-(2-p yr id yl)-1-p r op yloxy]p h osp h in yl-2′-
d eoxycytid in e (7b). 31P NMR (121 MHz, CDCl3): δ, 147.4,
147.7. FAB-HRMS: calcd for
C
51H58N5O8P (M + Cs)+
1032.3075, found 1032.3077.
N6-Ben zoyl-5′-O-(4,4′-d im eth oxytr ityl)-3′-O-(N,N-d iiso-
p r op yla m in o)-[3-(2-p yr id yl)-1-p r op yloxy]p h osp h in yl-2′-
d eoxya d en osin e (7c). 31P NMR (121 MHz, CDCl3): δ, 147.3,
147.6. FAB-HRMS: calcd for
1056.3176, found 1056.3190.
(7) Comparatively, thermal deprotection of the 3-(N-tert-butylcar-
boxamido)-1-propyl group from dinucleoside phosphotriesters is sig-
nificantly slower (t1/2 ∼100 s, see ref 2c) under similar conditions (pH
7.0, 90 °C).
C
52H58N7O7P (M + Cs)+
J . Org. Chem, Vol. 68, No. 26, 2003 10127

Cies´lak and Beaucage
N2-Isobu tyr yl-5′-O-(4,4′-d im eth oxytr ityl)-3′-O-(N,N-d i-
isop r op yla m in o)-[3-(2-p yr id yl)-1-p r op yloxy]p h osp h in yl-
2′-d eoxygu a n osin e (7d ). 31P NMR (121 MHz, CDCl3): δ,
147.3, 147.5. FAB-HRMS: calcd for C49H60N7O8P (M + Cs)+
1038.3329, found 1038.3295.
5′-O-(4,4′-Dim eth oxytr ityl)-3′-O-(N,N-diisopr opylam in o)-
[2-[N-m eth yl-N-(2-p yr id yl)]a m in oeth oxy]p h osp h in yl-2′-
d eoxyth ym id in e (9). This compound is prepared from 5a and
8, purified, and isolated in a manner identical to that described
for the preparation of 7a -d . 31P NMR (121 MHz, CDCl3): δ,
152.0, 152.2. FAB-HRMS: calcd for C45H55N5O8P (M - H)+
824.3820, found 824.3788.
P r ep a r a t ion a n d Ch a r a ct er iza t ion of Oligon u cle-
otid es. It is important to dissolve each of the purified
deoxyribonucleoside phosphoramidites 7a -d in dry benzene
and lyophilize each frozen solution to eliminate residual tri-
ethylamine that is carried through the phosphoramidite pur-
ification and postpurification processes. Automated solid-phase
synthesis of d(ATCCGTAGCTAAGGTCATGC) and that of its
phosphorothioated analogue is performed on a 0.2-µmol scale
using phosphoramidites 7a -d as 0.2 M solutions in dry MeCN.
The preparation of d(AT*T*CGT*AGCT*AAGGT*CAT*GC)
and that of its phosphorothioated analogue is also achieved
under similar conditions via 2-cyanoethyl deoxyribonucleoside
phosphoramidites and 9 (T*) as 0.1 M solutions in dry MeCN.
All reagents needed for the preparation of oligonucleotides
were purchased and used as recommended by the instrument’s
manufacturer. The sulfurization reaction that is indicated for
the preparation of phosphorothioated oligodeoxyribonucleo-
sides is performed employing 0.05 M 3H-1,2-benzodithiol-3-
one 1,1-dioxide in MeCN as recommended in the literature.12
The crude oligomers and their phosphorothioated analogues
are released from the CPG support by treatment with con-
centrated NH4OH for 30 min at 25 °C. Each of the oligonucle-
otide solutions is then heated for 10 h at 55 °C to ensure
complete nucleobase and phosphate/thiophosphate deprotec-
tion. The crude oligonucleotides are analyzed “DMTr-ON” by
RP-HPLC to estimate purity and compare synthesis yields
with that of control oligonucleotides synthesized from 2-cya-
noethyl deoxyribonucleoside phosphoramidites. RP-HPLC chro-
matograms are shown in Supporting Information.
might have occurred during oligonucleotide synthesis and de-
protection. Prior to enzymatic digestion, crude d(ATCCGTAGC-
TAAGGTCATGC) (∼25 OD260 units) in ddH2O (250 µL) is ap-
plied to the top of a prepacked PD-10 Sephadex G-25M column
equilibrated in ddH2O to eliminate the bicyclic salt 13. The
crude oligonucleotide is eluted with ddH2O, and 1 mL fractions
are collected. Fractions containing the oligonucleotide are
identified by UV at 260 nm and pooled together. One OD260
unit of the desalted oligonucleotide is evaporated to dryness
under reduced pressure and then subjected to enzymatic diges-
tion. Oligonucleotide d(AT*T*CGT*AGCT*AAGGT*CAT*GC)
does not require desalting prior to enzymatic digestion. Ali-
quots of either digests are analyzed by reversed-phase HPLC
under the conditions reported in ref 8. RP-HPLC profiles of
the digests are shown in Supporting Information.
Isola tion a n d Ch a r a cter iza tion of 2,3-Dih yd r o-1H-
in d olizin iu m Sa lt 13. This compound is isolated from the
thermal deprotection of dinucleotide 11 via RP-HPLC. The
dinucleoside phosphotriester 10 is first prepared manually by
condensing phosphoramidite 7a with thymidine, covalently
attached through a 3′-O-succinyl linker to LCAA-CPG, and a
0.45 M solution of 1H-tetrazole in MeCN. Following standard
iodine oxidation and cleavage of the 5′-dimethoxytrityl group,
10 is treated with concentrated NH4OH for 30 min at 25 °C,
affording 11. Heating the ammoniacal solution for 30 min at
55 °C results in complete removal of the phosphate protecting
group to produce dinucleotide 12 and some bicyclic salt 13.
RP-HPLC analysis of the deprotection reaction shows a fast-
eluting compound (tR ) 3.4 min) under chromatographic
conditions identical to those used for the characterization of
oligonucleotides. The fast-eluting compound is collected, and
the eluate is evaporated to dryness under reduced pressure.
The residue, generated from multiple RP-HPLC runs, is
characterized by high-resolution mass spectrometry. The mass
spectral data agree with the accurate mass expected for 13.
EI-HRMS: calcd for C8H10N1 (M+) 120.0811, found 120.0813.
To further confirm its structure, 13 is synthesized by mixing
a solution of 2-pyridinepropanol (100 µL) in MeCN (0.5 mL)
with TFAA (0.5 mL). The solution is heated in a tightly closed
vial overnight at 55 °C. Evaporation of excess TFAA gives an
oil. 1H NMR (300 MHz, DMSO-d6): δ 9.02 (d, J ) 6.0 Hz, 1H),
8.47 (dt, J ) 1.2, 8.1 Hz, 1H), 8.08 (d, J ) 8.1 Hz, 1H), 7.94
(m, 1H), 4.82 (t, J ) 7.7 Hz, 2H), 3.47 (t, J ) 7.7 Hz, 2H), 2.39
(qt, J ) 7.7 Hz, 2H). 13C NMR (75 MHz, DMSO-d6): δ 21.0,
31.8, 58.6, 116.7 (q, 1J CF ) 299 Hz), 124.4, 125.2, 141.2, 144.8,
157.8 (q, 2J CF ) 32.3 Hz), 158.5. EI-HRMS: calcd for C8H10N1
(M+) 120.0811, found 120.0813. RP-HPLC analysis of the oil
shows a peak having a retention time identical to that obtained
for 13 isolated from the deprotection of 11.
Fully deprotected oligomers (0.25 OD260 unit, each) are
further characterized by electrophoresis on 20% polyacryla-
mide-7 M urea gels (40 cm × 20 cm × 0.75 mm), which were
prepared using electrophoresis purity reagents. Gels are
stained by soaking in a solution of Stains-all as described
elsewhere.3a Photographs of such gels are shown in Figure 3
and in Supporting Information.
Phosphorothioated oligonucleotides prepared through the
use of phosphoramidites 7a -d and 9 were analyzed for
desulfurization by 31P NMR spectroscopy in concentrated NH4-
OH. Data are shown in Supporting Information.
Enzymatic digestion of crude d(ATCCGTAGCTAAGGT-
CATGC) and d(AT*T*CGT*AGCT*AAGGT*CAT*GC) is per-
formed with snake venom phosphodiesterase (Crotalus ada-
manteus) and bacterial alkaline phosphatase according to a
published procedure13 to assess nucleobase modifications that
Isola tion a n d Ch a r a cter iza tion of 1-Meth yl-2,3-d ih y-
d r oim id a zo[1,2-a ]p yr id in iu m sa lt 14. In a 5-mm NMR tube
are added 9 (16.5 mg, 0.020 mmol), 3′-O-acetylthymidine (6.2
mg, 0.022 mmol), 1H-tetrazole (2 mg, 0.028 mmol), and dry
MeCN (0.5 mL). 31P NMR analysis of the reaction indicates
immediate formation of the corresponding dinucleoside phos-
phite triester (δP ∼144 ppm). To the reaction mixture is added
3H-1,2-benzodithiol-3-one 1,1-dioxide (6.0 mg, 0.03 mmol),
which converted the phosphite triester to a dinucleoside
phosphorothioate diester (δP ∼60 ppm) with the concomitant
release of 14. RP-HPLC analysis of the solution reveals the
presence of a fast-eluting compound having a retention time
of 4.9 min under the chromatographic conditions used to
isolate 13. The fast-eluting compound is collected, and the
material accumulated from multiple RP-HPLC runs is char-
acterized by high-resolution mass spectrometry. The mass
spectral data are consistent with the accurate mass expected
for 14. EI-HRMS: calcd for C8H11N2 (M+) 135.0922, found
(8) RP-HPLC analyses were performed using a 5-µm Supelcosil LC-
18S column (25 cm × 4.6 mm) under the following conditions: starting
from 0.1 M triethylammonium acetate pH 7.0, a linear gradient of 1%
MeCN/min is pumped at a flow rate of 1 mL/min for 40 min and then
held isocratic for 20 min.
(9) Solid-phase synthesis of the fully phosphorothioated 20-mer
requires replacing the standard iodine solution with a solution of 0.05
M 3H-1,2-benzodithiol-3-one 1,1-dioxide in MeCN and performing the
capping reaction after the sulfurization step (see ref 12).
(10) Since 13 and 14 are ∼60% and 100% eliminated, respectively,
during each oxidation/sulfurization step of the oligonucleotide as-
sembly, the concentrations of 13 or 14 in the simulation experiments
are that of a worst-case scenario.
(12) Iyer, R. P.; Phillips, L. R.; Egan, W.; Regan, J . B.; Beaucage, S.
L. J . Org. Chem. 1990, 55, 4693-4699. See also: Regan, J . B.; Phillips,
L. R.; Beaucage, S. L. Org. Prep. Proc. Int. 1992, 24, 488-492.
(13) Scremin, C. L.; Zhou, L.; Srinivasachar, K.; Beaucage, S. L. J .
Org. Chem. 1994, 59, 1963-1966.
(11) Novina, C. D.; Murray, M. F.; Dykxhoorn, D. M.; Beresford, P.
J .; Riess, J .; Lee, S.-K.; Collman, R. G.; Lieberman, J .; Shankar, P.;
Sharp, P. A. Nat. Med. (N. Y.) 2002, 8, 681-686.
10128 J . Org. Chem., Vol. 68, No. 26, 2003

Thermolytic Phosphate/ Thiophosphate Protecting Groups
135.0928. To further establish its identity, 14 is synthesized
by mixing a solution of 2-[N-methyl-N-(2-pyridyl)]aminoetha-
nol (100 µL) in acetonitrile (0.5 mL) with TFAA (0.5 mL).
Heating the solution overnight at 55 °C in a tightly closed vial,
followed by evaporation of excess TFAA, produces an oil, which
is characterized by NMR spectroscopy. 1H NMR (300 MHz,
DMDO-d6): δ 8.23 (d, J ) 6.5 Hz, 1H), 7.99 (ddd, J ) 1.7, 7.3,
8.9 Hz, 1H), 7.2 (d, J ) 8.9 Hz, 1H), 6.90 (ddd, J ) 0.8, 6.5,
7.3 Hz, 1H), 4.64 (t, J ) 9.9 Hz, 2H), 3.90 (t, J ) 9.9 Hz, 2H),
3.07 (s, 3H). 13C NMR (75 MHz, CDCl3): δ 32.4, 49.6, 50.0,
107.5, 113.5, 115.6 (q, 1J CF ) 289 Hz), 137.3, 145.0, 154.8, 160.4
(q, 2J CF ) 37.9 Hz). RP-HPLC analysis of the oil shows a peak
having a retention time identical to that of 14 isolated after
sulfurization of the dinucleoside phosphite triester that was
synthesized from 9 and 3′-O-acetylthymidine.
P r oced u r e for Deter m in in g th e DNA-Mod ifyin g P r op -
er ties of 13 a n d 14. To thymidine, 2′-deoxycytidine, 2′-
deoxyadenosine, 2′-deoxyguanosine, N4-benzoyl-2′-deoxycyti-
dine, N6-benzoyl-2′-deoxyadenosine, and N2-isobutyryl-2′-
deoxyguanosine (0.5 mmol each) in a 7-mL screw-capped glass
vial are added 13 or 14 (4 mmol), concentrated NH4OH (1.1
mL), and MeCN (3 mL). The suspension solubilizes within 10
min upon heating at 55 °C. The solution is then kept at 55 °C
for a total time of 10 h. A control reaction is performed under
identical conditions in the absence of 13 or 14 for comparison
purposes. Aliquots of the reaction mixtures are analyzed by
RP-HPLC under the conditions delineated in ref 8. No nucleo-
base modifications are detected as only thymidine, 2′-deoxy-
cytidine, 2′-deoxyadenosine, 2′-deoxyguanosine, benzamide, 13
or 14, and their individual trace amount impurities are present
relative to that observed from the analysis of the control
reaction.
ppm) to O,O-diethyl phosphate (δP ∼2 ppm) (Data shown in
Supporting Information).
Ack n ow led gm en t. This research is supported in
part by an appointment to the Postgraduate Research
Participation Program at the Center for Biologics Evalu-
ation and Research administered by the Oak Ridge
Institute for Science and Education through an inter-
agency agreement between the U.S. Department of
Energy and the U.S. Food and Drug Administration. We
thank Andrzej Wilk for performing a preliminary ex-
periment in this project.
Su p por tin g In for m a tion Ava ila ble: Materials and meth-
ods; 1H, 13C, and 31P NMR spectra of crude 6 and 8; 31P NMR
spectra of 7a -d and 9; 1H and 13C NMR spectra of 13 and 14
as 2,2,2-trifluoroacetate salts; RP-HPLC chromatograms of
crude 5′-DMTr-d(ATCCGTAGCTAAGGTCATGC), 5′-DMTr-
d(AT*T*CGT*AGCT*AAGGT*CAT*GC), and that of their
fully phosphorothioated analogues synthesized using either
7a -d or 9 as indicated; RP-HPLC chromatograms of the snake
venom phosphodiesterase and bacterial alkaline phosphatase
digestion of crude d(ATCCGTAGCTAAGGTCATGC) and
d(AT*T*CGT*AGCT*AAGGT*CAT*GC); 31P NMR spectra of
crude phosphorothioated d(ATCCGTAGCTAAGGTCATGC)
and d(AT*T*CGT*AGCT*AAGGT*CAT*GC); photograph of a
stained 20% polyacrylamide-7 M urea gel showing the elec-
trophoretic profiles of crude d(AT*T*CGT*AGCT*AAGGT*-
CAT*GC) and that of its phosphorothioated analogue, both
synthesized from 9 and standard 2-cyanoethyl deoxyribo-
nucleoside phosphoramidites; 31P NMR spectra of the reaction
of O,O-diethyl thiophosphate with either 13 or 14 under
conditions simulating those of large-scale oligonucleotide
deprotections. This material is available free of charge via the
Internet at http://pubs.acs.org.
P r oced u r e for Deter m in in g th e P h osp h or oth ioa te
Desu lfu r iza tion P r op er ties of 13 or 14. A solution of 0.5
M potassium O,O-diethyl thiophosphate and 0.75 M 13 or 14
in concentrated NH4OH is heated for 10 h at 55 °C. Analysis
of the reaction mixture by 31P NMR spectroscopy does not
reveal any conversion of O,O-diethyl thiophosphate (δP ∼55
J O0354490
J . Org. Chem, Vol. 68, No. 26, 2003 10129