Silylation-mediated transesterification of O-phenyl H-phosphonothioates - A new entry to nucleoside H-phosphonothioate monoesters

Article abstract of DOI:10.1002/ejoc.200400550

O-Phenyl H-phosphonothioate undergoes a facile transesterification with suitably protected nucleosides upon in situ silylation with tert- butyldiphenylsilyl chloride in pyridine/toluene to produce the corresponding 3′-H-phosphonothioates in good yields. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2004.

Full text of DOI:10.1002/ejoc.200400550

FULL PAPER
Silylation-Mediated Transesterification of O-Phenyl H-Phosphonothioates ؊
A New Entry to Nucleoside H-Phosphonothioate Monoesters
[a]
[a]
[a,b]
´
Gaston Laven, Johan Nilsson, and Jacek Stawinski*
Keywords: Phosphonates / Phosphonothioates / Silylation / Transesterification
O-Phenyl H-phosphonothioate undergoes a facile trans-
esterification with suitably protected nucleosides upon in situ
silylation with tert-butyldiphenylsilyl chloride in pyridine/
toluene to produce the corresponding 3Ј-H-phosphono-
thioates in good yields.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2004)
Introduction
phonate diesters,[16] or a dedicated H-phosphonothioate
group-transfer reagent, O-(9-fluorenylmethyl) phosphono-
thioate.[5,17] Although all these methods have their own
During the last decade we have introduced and have been merits and usually afford products in good yields, we would
investigating various aspects of H-phosphonothioates (‘‘H’’ like to expand the array of synthetic methods available for
indicates a PϪH bond)[1Ϫ5] as a new type of synthetic inter- the preparation of H-phosphonothioate monoesters by de-
mediates for the preparation of biologically active phos- veloping an approach based on transesterification of tetra-
phorus compounds based on H-phosphonate methodology. coordinate PIII compounds containing an HϪPϭS func-
Due to their higher stability, ease of handling, and ef- tion. This would provide additional insight into the reactiv-
ficiency in solid-phase synthesis,[6Ϫ8] H-phosphonothioate ity of compounds bearing an HϪPϭS functionality and,
building blocks[1,9] seem to be superior to phosphoro- due to the differences in the underlying chemistry and the
thioamidite derivatives[10] in the synthesis of nucleotide ana- reaction conditions used, it could expand the range of hy-
logues with multiple modifications at the phosphorus droxylic compounds that can be subjected to H-thiophos-
centre that are difficult to prepare in other ways, for phonylation.
example from nucleoside phosphorodithioates,[6,7,11,12]
nucleoside phosphoramidothioates,[13] nucleoside phos- H-phosphonothioate monoester and its use in a silylation-
phorofluoridothioates,[14] etc.
mediated transesterification reaction to obtain various
In this paper we describe the preparation of O-phenyl
Despite their usefulness in the synthesis of various types nucleoside H-phosphonate monoesters.
of phosphorus compounds, there are only a handful of syn-
thetic methods available for the preparation of nucleoside
H-phosphonothioate monoesters. These compounds can be
obtained either by non-oxidative thiation of nucleoside H-
Results and Discussion
phosphonate derivatives (e.g. by sulfhydrolysis of tervalent
PIII species[3] or from aryl nucleoside H-phosphonate de-
rivatives[15]) or by using various thiophosphonylation pro-
tocols.[2,5,16,17] This latter group includes methods based on
reactions of suitably protected hydroxylic compounds with
triethylammonium phosphinate in the presence of a con-
densing agent, followed by sulfurization of the phosphinate
monoesters formed in situ,[2] sulfhydrolysis of aryl H-phos-
We were looking for a reagent which is stable, easy to
prepare, and does not require an additional synthetic step
(e.g. a removal of a phosphate protecting group) to convert
intermediate compounds into H-phosphonothioate monoe-
sters. An attractive candidate for this purpose seemed to be
O-phenyl H-phosphonothioate. As we were aware of the
high propensity of aryl H-phosphonate diesters to undergo
hydrolysis and sulfhydrolysis, it seemed likely that O-phenyl
H-phosphonothioate could be obtained upon treatment of
diphenyl H-phosphonate with hydrogen sulfide. The pres-
ence of a negative charge at the phosphorus centre should
make this compound stable and resistant to spontaneous
air oxidation, but at the same time it would effectively pre-
vent transesterification of the phenoxy group by the nucleo-
[a]
Department of Organic Chemistry, Arrhenius Laboratory,
Stockholm University
10691 Stockholm, Sweden
[b]
Institute of Bioorganic Chemistry, Polish Academy of Sciences
Poznan, Poland
E-mail: js@organ.su.se
Eur. J. Org. Chem. 2004, 5111Ϫ5114
DOI: 10.1002/ejoc.200400550
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5111

´
G. Laven, J. Nilsson, J. Stawinski
FULL PAPER
side. As a viable means to overcome the anticipated low agent, and a suitably protected nucleoside 3a to react
reactivity of this reagent, we considered that carrying out a (Scheme 2). The reactions were not particularly clean and
transesterification reaction with nucleoside in the presence 31P NMR spectroscopy revealed the formation of several
of a silylating agent should make the phosphorus center side products. Most notable among these were dinucleoside
more susceptible to a nucleophilic attack towards oxygen H-phosphonothioate diesters and silylated H-phosphono-
nucleophiles by O-silylating the H-phosphonothioate thioates, when TMSCl was used as the silylating agent.[21]
monoester function.[18] This approach would have an ad- Although far from satisfactory, the best results were ob-
ditional advantage over the methods based on sulfhydro- tained with TBDPSCl. This reagent produced the least
lysis of reaction intermediates[3,15,16] in that inconveniences number of side products during the reaction in pyridine
connected with the preparation of a stock solution of hy- and, due to its low reactivity towards hydroxy groups in
drogen sulfide prior to the reaction would be alleviated.
The synthesis of O-phenyl H-phosphonothioate (1) by
sulfhydrolysis of diphenyl H-phosphonate (DPHP) was in-
vestigated first (Scheme 1). The reaction was carried out by
adding a saturated solution of hydrogen sulfide (3.3 molar
excess), trimethylsilyl chloride (TMSCl, 1.7 equiv.), and tri-
ethylamine (TEA, 7 equiv.) to a solution of DPHP in pyri-
dine. TMSCl was used to ensure anhydrous reaction con-
ditions, and TEA to prevent the formation of phenyl H-
phosphonodithioate.[16] The reaction proceeded as expected
and produced O-phenyl H-phosphonothioate (1) as the
major product along with some small amounts of phenyl
H-phosphonate monoester (5Ϫ10%).[19] Since the removal
of this side product by column chromatography on silica
gel was inefficient, and handling of an oily triethylammon-
ium salt of O-phenyl H-phosphonothioate (1) inconvenient,
we tried to convert reagent 1 into a derivative that would
permit crystallographic purification of the crude reaction
mixture. Among various salts investigated, the (p-chloro-
benzyl)thiouronium derivative turned out to be crystalline,
and provided reagent 1 as a stable solid with good solubility
in organic solvents and free of phenyl H-phosphonate
monoester.[20]
nucleosides 3, it was used for further optimisation studies.
Scheme 2
We assumed that the source of most side reactions ob-
served was, most likely, a too high reactivity of the reaction
system containing pyridine, and thus we searched for a sol-
vent composition that would provide the best compromise
between reactivity and selectivity. After extensive experi-
mentation with various organic solvents, we arrived at a
solvent system consisting of toluene (4 parts) and pyridine
(1 part), which allowed a very clean reaction of 1 with nu-
cleoside 3a in the presence of TBDPSCl, in a reasonable
time period (overnight reaction).[22]
These reaction conditions were then applied for prepara-
tive syntheses involving other protected nucleosides 3, and
the desired nucleoside H-phosphonothioates 4 were isolated
in 72Ϫ77% yield after simple aqueous workup followed by
silica-gel column chromatography (Scheme 2). TLC analysis
indicated, in all instances, the complete disappearance of
the starting material 3,[23] and 31P NMR spectroscopy con-
firmed that the only nucleotide-containing material formed
was the corresponding H-phosphonothioate 4.
Scheme 1
We then investigated the reaction conditions most suit-
able for transferring an H-phosphonothioate moiety into
nucleosides 3. Pyridine, or solvent systems containing pyri-
dine, seemed to be the obvious choice as pyridine should
Conclusions
We have developed a new thiophosphonylating reagent —
provide efficient nucleophile catalysis during both reaction O-phenyl H-phosphonothioate (1) — that, in a silylation-
steps (i.e. during the conversion of 1 into a silyl ester, and mediated transesterification reaction with nucleosides 3 un-
during the transesterification reaction of the produced in- der mild conditions, produces the corresponding nucleoside
termediate of type 2 with nucleoside 3). In addition, the 3Ј-H-phosphonothioates 4 in good yield. This reagent is
weakly basic properties of pyridine should ensure the sta- easy to prepare on a large scale from inexpensive, commer-
bility of the acid-sensitive dimethoxytrityl groups in the cially available materials, is crystalline, and can be stored at
protected nucleoside moieties during the course of the reac- room temperature for a prolonged period of time (months)
tion. Preliminary screening of various silylating agents without noticeable decomposition. This approach can prob-
[TMSCl, tert-butyldimethylsilyl (TBDMS), and tert-butyld- ably be extended to other hydroxylic compounds and thus
iphenylsilyl (TBDPS) chlorides] was carried out in neat expands the array of synthetic methods available for the
pyridine, by allowing equimolar amounts of 1, a silylating preparation of H-phosphonothioate monoesters.
5112
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 5111Ϫ5114

Silylation-Mediated Transesterification of O-Phenyl H-Phosphonothioates
FULL PAPER
added pyridine (3 ϫ 20 mL), and dissolved in a pyridine/toluene
mixture (1:4, v/v; 50 mL). tert-Butyldiphenylsilyl chloride (0.4 mL,
1.5 mmol) was added to this mixture, with vigorous stirring, and it
was left overnight. When complete (TLC analysis), the reaction
mixture was quenched by addition of pyridine (5 mL), triethyl-
amine (1 mL), and water (0.8 mL), concentrated to dryness, and
partitioned between dichloromethane (100 mL) and saturated aq.
NaHCO3 (100 mL). The organic layer was washed with saturated
aq. NaHCO3 (2 ϫ 100 mL), dried with anhydrous Na2SO4, and
the solvents were evaporated. The residue was purified by silica-gel
column chromatography using a gradient of methanol (6Ϫ10%)
and triethylamine (0.01Ϫ0.5%) in dichloromethane to furnish nu-
cleoside H-phosphonothioates 4 as white, amorphous solids (purity
Experimental Section
Material and Methods: 1H and 31P NMR spectra were recorded
with a Varian Unity 400 BB VT spectrometer. The 31P NMR spec-
troscopy experiments were carried out at 25 °C in 5-mm tubes using
0.1  concentrations of phosphorus-containing compounds in ap-
propriate solvents (0.6 mL; the spectra were referenced to 2%
H3PO4 in D2O (external standard). TLC analyses were carried out
on Merck silica gel 60 F254 precoated plates using a CH2Cl2/
CH3OH (9:1, v/v) solvent system. To avoid smearing, the plates
were immersed in CH2Cl2/CH3OH (9:1, v/v) containing 1% of tri-
ethylamine, before chromatography, and dried. Pyridine (LabScan
Ltd.), toluene and anhydrous acetonitrile (LabScan Ltd.) were
1
Ͼ98% by H NMR spectroscopy).
˚
stored over molecular sieves (4 A). Diphenyl H-phosphonate, tri-
methylsilyl chloride, and tert-butyldiphenylsilyl chloride were com-
mercial-grade reagents from Aldrich. Triethylamine (Aldrich) was
freshly distilled. The protected nucleosides were obtained according
to a published procedure.[24] The assignment of signals in the 31P
NMR spectra to particular products or intermediates was carried
out on the basis of their chemical shifts, multiplicity of the signals
in 1H-coupled and 1H-decoupled spectra, by spiking the reaction
mixtures with appropriate species, and, if possible, by isolation of
the compound in question from reaction mixtures. The assignment
of the proton and carbon resonances of compounds 4 was carried
out on the basis of known or expected chemical shifts in conjunc-
tion with 1H-1H, 1H-13C, and DEPT correlated NMR spec-
troscopy.
5Ј-O-Dimethoxytritylthymidine
3Ј-H-Phosphonothioate
(4a):
523 mg (72% yield). Rf ϭ 0.34. 31P NMR (CDCl3): δ ϭ 54.65 and
1
3
53.78 (each dd, JP,H ϭ 580.0 and 583.9 Hz; JP,H ϭ 12.6 and
12.0 Hz) ppm. HRMS for C31H32N2O8PS [M Ϫ TEAH]Ϫ: calcd.
623.1622; found 623.1633. The 1H and 13C NMR spectroscopic
data of the compound were identical to those reported previously.[2]
5Ј-O-Dimethoxytrityl-N4-benzoyldeoxycytidine
3Ј-H-Phosphono-
thioate (4b): 627 mg (77% yield). Rf ϭ 0.36. 31P NMR (CDCl3):
δ ϭ 53.03 and 51.72 (each dd, JP,H ϭ 577.2 and 586.5 Hz; 3JP,H ϭ
1
11.4 and 12.0 Hz) ppm. HRMS for C37H35N3O8PS [M Ϫ TEAH]Ϫ:
calcd. 712.1888; found 712.1897. The 1H and 13C NMR spectro-
scopic data of the compound were identical to those reported pre-
viously.[2]
Synthesis of (p-Chlorobenzyl)thiouronium Chloride: p-Chlorobenzyl
chloride (161 g, 1.0 mol) and thiourea (76 g, 1.0 mol) were refluxed
in ethanol (200 mL) for 2 h. The reaction mixture was cooled down
and the product was filtered off, and recrystallized from 36% aq.
HCl/water (1:1, v/v). Yield: 209 g (88%). The 1H NMR spectrum
of the product was identical with a commercial sample of (p-chloro-
benzyl)thiouronium chloride.
5Ј-O-Dimethoxytrityl-N6-benzoyldeoxyadenosine 3Ј-H-Phosphono-
thioate (4c): 609 mg (72% yield). Rf ϭ 0.38. 31P NMR (CDCl3):
1
δ ϭ 54.35 and 54.30 (each dd, JP,H ϭ 581.5 and 582.1 Hz; 3JP,H ϭ
10.6 and 11.7 Hz) ppm. HRMS for C38H35N5O7PS [M Ϫ TEAH]Ϫ:
calcd. 736.2000; found 736.2010. The 1H and 13C NMR spectro-
scopic data of the compound were identical to those reported pre-
viously.[2]
Synthesis of O-Phenyl H-Phosphonothioate (p-Chlorobenzyl)thio-
uronium Salt 1: A mixture of a freshly prepared solution of H2S in
dioxane (1  solution, 200 mmol, 200 mL), trimethylsilyl chloride
(12.7 mL, 100 mmol), and triethylamine (60 mL, 432 mmol) was
added to a stirred solution of diphenyl H-phosphonate (11.48 mL,
60 mmol) in pyridine (100 mL). After 1 h, the reaction solvents
were evaporated to dryness, the residue dissolved in ethanol
(100 mL), and (p-chlorobenzyl)thiouronium chloride (13.3 g,
56.0 mmol) was added. The mixture was gently heated to obtain a
homogeneous solution, concentrated, and the residue washed with
water (2 ϫ 100 mL). After recrystallization from ethanol, white
crystals were obtained (15.75 g, 70% yield). M.p. 134Ϫ135 °C.
C14H16ClN2O2PS2 (374.85): calcd. C 44.86, H 4.30, N 7.47; found
C 45.05, H 4.53, N 7.33. HRMS for C6H6O2PS [M Ϫ cbtu]Ϫ: calcd.
5Ј-O-Dimethoxytrityl-N2-isobutyryldeoxyguanosine 3Ј-H-Phosphono-
thioate (4d): 624 mg (76% yield). Rf ϭ 0.26. 31P NMR (CDCl3):
1
δ ϭ 54.61 and 53.95 (each dd, JP,H ϭ 585.4 and 579.0 Hz; 3JP,H ϭ
13.3 and 11.4 Hz) ppm. HRMS for C35H37N5O8PS [M Ϫ TEAH]Ϫ:
calcd. 718.2106; found 718.2120. The 1H and 13C NMR spectro-
scopic data of the compound were identical to those reported pre-
viously.[2]
Acknowledgments
Financial support from the Swedish Research Council is grate-
fully acknowledged.
1
172.9826; found 172.9833. H NMR ([D6]DMSO): δ ϭ 9.32 (s, 4
1
[1]
H, 2 ϫ NH2), 9.01 (d, JP,H ϭ 568.9 Hz, PH), 7.47Ϫ7.38 (m, 4 H,
J. Stawinski, M. Thelin, R. Zain, Tetrahedron Lett. 1989, 30,
2157Ϫ2160.
H2, H3), 7.25 (3 m, 2 H, H9), 7.09 (2 m, 2 H, H8), 7.01 (2 m 1 H,
H10), 4.49 (s, 2 H, H5) ppm. 13C NMR ([D6]DMSO): δ ϭ 169.44
(C6), 153.44, 153.32 (C7, JC,P ϭ 8.8 Hz), 135.09 (C4), 133.32 (C1),
131.54 (C3), 129.84 (C9), 129.42 (C2), 123.58 (C10), 121.61, 121.54
(C8, JC,P ϭ 5.4 Hz), 34.03 (C5) ppm. 31P NMR ([D6]DMSO): δ ϭ
49.80 (1JP,H ϭ 568.9 Hz) ppm.
[2]
J. Stawinski, M. Thelin, E. Westman, R. Zain, J. Org. Chem.
1990, 55, 3503Ϫ3506.
[3]
R. Zain, R. Strömberg, J. Stawinski, J. Org. Chem. 1995, 60,
8241Ϫ8244.
[4]
R. Zain, J. Stawinski, J. Chem. Soc., Perkin Trans. 2 1996,
795Ϫ799.
J. Jankowska, J. Cieslak, A. Kraszewski, J. Stawinski, Tetra-
[5]
General Procedure for the Preparation of Protected Nucleoside 3Ј-
H-Phosphonothioates 4 (TEAH؉ Salts): A suitably protected nu-
cleoside 3 (1 mmol; 0.55 g 3a, 0.64 g 3b, 0.66 g 3c, 0.63 g 3d) and
O-phenyl H-phosphonothioate (p-chlorobenzyl)thiouronium salt 1
(0.55 g, 1.5 mmol) were rendered anhydrous by co-evaporation with
hedron Lett. 1997, 38, 2007Ϫ2010.
R. Zain, M. Bollmark, J. Stawinski, Nucleosides Nucleotides
[6]
1997, 16, 1661Ϫ1662.
P. H. Seeberger, M. H. Caruthers, D. Bankaitis-Davis, G. Be-
[7]
aton, Tetrahedron 1999, 55, 5759Ϫ5772.
Eur. J. Org. Chem. 2004, 5111Ϫ5114
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5113

´
G. Laven, J. Nilsson, J. Stawinski
FULL PAPER
[8]
[19]
[20]
K. Kamaike, K. Hirose, Y. Kayama, E. Kawashima, Tetra-
Formation of phenyl H-phosphonate (31P NMR spectroscopy)
is probably due to the presence of adventitious water or to
partial disproportionation of DPHP under the reaction con-
ditions. See: A. Kers, I. Kers, J. Stawinski, M. Sobkowski, A.
Kraszewski, Tetrahedron 1996, 52, 9931Ϫ9944.
Later experiments showed that the presence of phenyl H-phos-
phonate monoester (5Ϫ10%) in the reaction mixture contain-
ing 1, nucleoside 3a, and TBDPSCl does not lead to any detect-
able formation of the corresponding nucleoside H-phosphon-
ate derivatives, probably because of the significantly lower reac-
tivity of the silylated phenyl H-phosphonate relative to the
phenyl H-phosphonothioate derivative. See ref.[18]
The formation of a dinucleoside H-phosphonothioate diester
in this reaction is probably due to the lability of trimethylsilyl
esters and the partial generation of a nucleoside H-phosphono-
thioate monoester from the corresponding silylated species.
The produced monoester could then react with the silylated
intermediate of type 2 present in the reaction mixture to afford
unsymmetrical H-pyrophosphonothioate, from which, upon re-
action with nucleoside 3a, either the desired silylated product
4a or dinucleoside H-phosphonothioate diester would be
formed.
The reactions with added TEA were less efficient and usually
occurred to only about 50% completion, most likely due to
consumption of the silylating agent by the released phenol (31P
NMR experiments). Also, the presence of TEA stimulated si-
lylation of nucleosides 3 (TLC analysis).
hedron Lett. 2004, 45, 30.
C. H. Greef, P. H. Seeberger, M. H. Caruthers, G. Beaton, D.
[9]
Bankaitis-Davis, Tetrahedron Lett. 1996, 37, 4451Ϫ4454.
W. K.-D. Brill, J.-Y. Tang, Y.-X. Ma, M. H. Caruthers, J. Am.
[10]
Chem. Soc. 1989, 111, 2321Ϫ2322.
J. Jankowska, A. Sobkowska, J. Cieslak, M. Sobkowski, A.
[11]
Kraszewski, J. Stawinski, D. Shugar, J. Org. Chem. 1998, 63,
8150Ϫ8156.
[12]
P. H. Seeberger, E. Yau, M. H. Caruthers, J. Am. Chem. Soc.
1995, 117, 1472Ϫ1478.
[13]
I. Kers, J. Stawinski, A. Kraszewski, Tetrahedron Lett. 1998,
[21]
39, 1219Ϫ1222.
M. Bollmark, J. Stawinski, Tetrahedron Lett. 1996, 37,
5739Ϫ5742.
[14]
[15]
J. Cieslak, J. Jankowska, J. Stawinski, A. Kraszewski, J. Org.
Chem. 2000, 65, 7049Ϫ7054.
I. Kers, A. Kers, J. Stawinski, A. Kraszewski, Tetrahedron Lett.
[16]
1999, 40, 3945Ϫ3948.
P. H. Seeberger, P. N. Jorgensen, D. M. Bankaitis-Davis, G.
[17]
Beaton, M. H. Caruthers, J. Am. Chem. Soc. 1996, 118,
9562Ϫ9566.
[18]
[22]
A similar approach was recently used by Jones et al. for simul-
taneous phosphonylation and silylation of ribonucleosides by
phenyl H-phosphonate and tert-butyldimethylsilyl chloride.
Unfortunately, due to the low reactivity of phenyl H-phos-
phonate, the reaction had to be carried out in pyridine for 48 h,
and required the presence of a strong base, DBU, to effect the
transesterification. Such reaction conditions are not compat-
ible with N-acyl-protected nucleosides. See: Q. L. Song, W. M.
Wang, A. Fischer, X. H. Zhang, B. L. Gaffney, R. A. Jones,
Tetrahedron Lett. 1999, 40, 4153Ϫ4156.
[23]
[24]
Occasionally, some amount of the silylated nucleoside was
formed but it did not exceed 5%.
H. Schaller, G. Weimann, B. Lerch, H. G. Khorana, J. Am.
Chem. Soc. 1963, 85, 3821Ϫ3827.
Received August 3, 2004
5114
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 5111Ϫ5114