Aryl H-Phosphonates. 12. Synthetic and 31P NMR Studies on the Preparation of Nucleoside H-Phosphonothioate and Nucleoside H-Phosphonodithioate Monoesters

Article abstract of DOI:10.1021/jo000729q

Transformation of nucleoside H-phosphonate monoesters into the corresponding H-phosphonothioate and H-phosphonodithioate derivatives and possible side-reactions that may accompany this process were studied using ³¹P NMR spectroscopy. These provided new insight into a possible mechanism involved in this transformation and constituted the basis for development of efficient methods for the preparation of nucleoside H-phosphonothioate and nucleoside H-phosphonodithioate monoesters using readily available H-phosphonate monoesters as starting materials.

Full text of DOI:10.1021/jo000729q

J . Org. Chem. 2000, 65, 7049-7054
7049
Ar yl H-P h osp h on a tes. 12. Syn th etic a n d 31P NMR Stu d ies on th e
P r ep a r a tion of Nu cleosid e H-P h osp h on oth ioa te a n d Nu cleosid e
H-P h osp h on od ith ioa te Mon oester s
J acek Cies´lak,† J adwiga J ankowska,† J acek Stawin´ski,*,†,‡ and Adam Kraszewski*,†
Institute of Bioorganic Chemistry, Polish Academy of Sciences, Noskowskiego 12/ 14,
61-704 Poznan´, Poland, and Department of Organic Chemistry, Arrhenius Laboratory, Stockholm
University, S-106 91 Stockholm, Sweden
akad@ibch.poznan.pl
Received May 13, 2000
Transformation of nucleoside H-phosphonate monoesters into the corresponding H-phosphonothioate
and H-phosphonodithioate derivatives and possible side-reactions that may accompany this process
were studied using 31P NMR spectroscopy. These provided new insight into a possible mechanism
involved in this transformation and constituted the basis for development of efficient methods for
the preparation of nucleoside H-phosphonothioate and nucleoside H-phosphonodithioate monoesters
using readily available H-phosphonate monoesters as starting materials.
In tr od u ction
During the past decade we have introduced and
investigated various aspects of H-phosphonothioates4,10-12
as a new type of synthetic intermediate that can supple-
ment and extend applications based on H-phosphonate
derivatives. In the context of nucleoside phosphorodithio-
ates synthesis via P(III) intermediates, H-phosphono-
thioate building blocks4,13 seem to be superior to phos-
phorothioamidite derivatives3 because of their higher
stability, ease of handling of the starting materials, and
efficiency in solid-phase synthesis.14 This stimulated
interest in H-phosphonothioates as convenient starting
materials for the preparation of biologically important
nucleotide analogues that can be difficult to prepare in
other ways, e.g., nucleoside phosphorodithioates,8,14,15
nucleoside phosphoramidothioates,16 nucleoside phos-
phorofluoridothioates,17 etc.
The pharmacological value of oligonucleotides with
natural phosphodiester internucleotide linkages is seri-
ously hampered by their rapid degradation in the pres-
ence of cell and serum nucleases.1 As a result of this, most
investigations have been focused on finding chemical
modifications that would confer stability to nucleases
while preserving the specificity and efficiency of com-
plexation of the modified oligonucleotide with target DNA
or RNA sequences.2 Recent studies have shown that
among nucleic acids analogues with modifications at the
phosphorus center, oligonucleoside phosphorodithioates3-6
possess most favorable biological and pharmacological
properties as potential antisense or antigene agents7 and
also as nucleic acids based clinical diagnostics.8,9
Nucleoside H-phosphonothioate monoesters can be
prepared either from suitably protected nucleosides using
various thiophosphonylation protocols10,18,19 or by non-
oxidative thiation of nucleoside H-phosphonates.11 This
latter approach seems to be particularly attractive in
† Polish Academy of Sciences.
‡ Stockholm University.
(1) Zon, G. Pharm. Res. 1988, 5, 539-549.
(2) Agrawal, S.; Zhao, Q. Curr. Opin. Chem. Biol. 1998, 2, 519-
528.
(3) Brill, W. K.-D.; Tang, J .-Y.; Ma, Y.-X.; Caruthers, M. H. J . Am.
Chem. Soc. 1989, 111, 2321-2322.
(4) Stawin´ski, J .; Thelin, M.; Zain, R. Tetrahedron Lett. 1989, 30,
2157-2160.
(10) Stawin´ski, J .; Thelin, M.; Westman, E.; Zain, R. J . Org. Chem.
1990, 55, 3503-3506. J ankowska, J .; Cies´lak, J .; Kraszewski, A.;
Stawin´ski, J . Tetrahedron Lett. 1997, 38, 2007-2010.
(11) Zain, R.; Stro¨mberg, R.; Stawin´ski, J . J . Org. Chem. 1995, 60,
8241-8244.
(12) Zain, R.; Stawin´ski, J . J . Chem. Soc., Perkin Trans. 2 1996,
795-799.
(5) Porritt, G. M.; Reese, C. B. Tetrahedron Lett. 1989, 30, 4713-
4716.
(6) Dahl, B. H.; Bjergarde, K.; Nielsen, J .; Dahl, O. Nucleosides &
Nucleotides 1991, 10, 553-554; Okruszek, A.; Sierzchała, A.; Fearon,
K. L.; Stec, W. J . J . Org. Chem. 1995, 60, 6998-7005.
(7) Caruthers, M. H.; Beaton, G.; Cummins, L.; Dellinger, D.; Graff,
D.; Ma, Y.-X.; Marshall, W. S.; Sasmor, H.; Shankland, P.; Van Wu, J .
V.; Yau, E. K. Nucleosides Nucleotides 1991, 10, 47-59. Cummins, L.;
Graff, D.; Beaton, G.; Marshall, W. S.; Caruthers, M. H. Biochemistry
1996, 35, 8734-8741. DeLong, R. K.; Nolting, A.; Fisher, M.; Chen,
Q.; Wickstrom, E.; Kligshteyn, M.; Demirdji, S.; Caruthers, M.; J uliano,
R. L. Antisense Nucl. Acid Drug Dev. 1997, 7, 71-77. Furrer, P.; Billeci,
T. M.; Donati, A.; Kojima, C.; Karwowski, B.; Sierzchała,, A.; Stec,
W.; J ames, T. L. J . Mol. Biol. 1999, 285, 1609-1622. Marshall, W. S.;
Beaton, G.; Stein, C. A.; Matsukura, M.; Caruthers, M. H. Proc. Natl.
Acad. Sci. U.S.A. 1992, 89, 6265-6269.
(8) Seeberger, P. H.; Yau, E.; Caruthers, M. H. J . Am. Chem. Soc.
1995, 117, 1472-1478.
(9) Brummer, O.; Wentworth, P.; Weiner, D. P.; J anda, K. D.
Tetrahedron Lett. 1999, 40, 7307-7310. Yang, X. B.; Fennewald, S.;
Luxon, B. A.; Aronson, J .; Herzog, N. K.; Gorenstein, D. G. Bioorg.
Med. Chem. Lett. 1999, 9, 3357-3362.
(13) Greef, C. H.; Seeberger, P. H.; Caruthers, M. H.; Beaton, G.;
Bankaitis-Davis, D. Tetrahedron Lett. 1996, 37, 4451-4454.
(14) Zain, R.; Bollmark, M.; Stawin´ski, J . Nucleosides Nucleotides
1997, 16, 1661-1662. Seeberger, P. H.; Caruthers, M. H.; Bankaitis-
Davis, D.; Beaton, G. Tetrahedron 1999, 55, 5759-5772.
(15) J ankowska, J .; Sobkowska, A.; Cies´lak, J .; Sobkowski, M.;
Kraszewski, A.; Stawin´ski, J .; Shugar, D. J . Org. Chem. 1998, 63,
8150-8156.
(16) Kers, I.; Stawin´ski, J .; Kraszewski, A. Tetrahedron Lett. 1998,
39, 1219-1222.
(17) Bollmark, M.; Stawin´ski, J . Tetrahedron Lett. 1996, 37, 5739-
5742.
(18) Kers, I.; Kers, A.; Stawin´ski, J .; Kraszewski, A. Tetrahedron
Lett. 1999, 40, 3945-3948.
(19) Seeberger, P. H.; J orgensen, P. N.; Bankaitis-Davis, D. M.;
Beaton, G.; Caruthers, M. H. J . Am. Chem. Soc. 1996, 118, 9562-
9566.
10.1021/jo000729q CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/20/2000

7050 J . Org. Chem., Vol. 65, No. 21, 2000
Cies´lak et al.
light of the growing availability of H-phosphonate mo-
noesters20,21 and has the added advantage that other
synthetically useful intermediates, H-phosphonodithioate
monoesters,5,22,23 can also be prepared from the same
substrate.24 Unfortunately, at present there is only one
protocol available for the transformation of H-phospho-
nate monoesters into the corresponding H-phosphono-
thioate derivatives.11 This is based on a P(III) interme-
diate (a nucleoside pivaloyl silyl phosphite12) whose
reactivity is difficult to modulate. To circumvent these
shortcomings, we searched for a more versatile method
for the H-phosphonates-H-phosphonothioates transfor-
mation that would permit control of reactivity of the
intermediate involved depending on structural variation
of the substrates and the reaction conditions required.
In this paper we describe 31P NMR studies that enabled
us to delineate the most important mechanistic aspects
of the reactions of activated H-phosphonate derivatives
with hydrogen sulfide (or its equivalent) under various
experimental conditions. On the basis of these observa-
tions, it became possible to direct these reactions, which
usually afforded various proportions of H-phosphono-
mono- and H-phosphonodithioates, to produce almost
exclusively one of these products. According to the
protocol developed, various nucleoside H-phosphonothio-
ates and nucleoside H-phosphonodithioates were ob-
tained in 85-95% yields in one-pot reactions from the
corresponding H-phosphonate monoesters.
Resu lts a n d Discu ssion
Reactions of nucleoside H-phosphonate monoesters of
type 1 with hydrogen sulfide in the presence of pivaloyl
chloride (or sulfhydrolysis of the in situ formed phospho-
nic-carboxylic anhydrides) usually afford as a major
product the corresponding H-phosphonodithioate mo-
noesters 3 with various proportions of an unreacted
H-phosphonate monoester, depending on the amount of
pivaloyl chloride used.11,25 This was assumed to be due
to relatively slow formation of the initial product, H-
phosphonothioate monoester of type 2, and its rapid
reaction toward H-phosphonodithioate derivative 3. To
change the unfavorable kinetics of this transformation,
we searched for an intermediate that would react fast
with hydrogen sulfide to produce H-phosphonothioate
monoesters 2 and in this way suppress the formation of
dithio derivatives 3. For this purpose we investigated the
reaction of nucleoside H-phosphonate monoesters 1 with
hydrogen sulfide in he presence of diphenyl phosphoro-
chloridate (DPCP) and sulfhydrolysis of various nucleo-
side aryl H-phosphonates of type 9.
(3 molar equiv), and progress of the reaction was moni-
tored by 31P NMR spectroscopy. The reaction was rapid,
and the first spectrum recorded (<3 min) revealed a
complete disappearance of the starting material 1a (δP
) 1.5 ppm) and the formation of nucleoside H-phospho-
1
3
nodithioate 3a (δP ) 85.4 ppm, J PH ) 528.3 Hz, J PH
)
13.9 Hz; dd) as the sole nucleotidic product.26 The
exclusive formation of dithio derivative 3a indicated that,
similarly to the pivaloyl chloride promoted reaction,11 the
second step (i.e., the transformation of H-phosphonothio-
ate 2a into 3a ) remained faster than the formation of
the initial monothio product 2a . However, somewhat
surprisingly we found that although sulfhydrolysis of
H-phosphonothioate 2a under analogous conditions pro-
duced the corresponding H-phosphonodithioate 3a , the
reaction was significantly slower than that of H-phos-
phonate 1a (ca 25 min vs <3 min for the completion).
This suggested that besides mixed anhydride 5, a puta-
tive common intermediate in transformations 1a f 3a
and 2a f 3a , there was probably yet another reactive
Rea ction s of Nu cleosid e H-P h osp h on a tes 1 w ith
DP CP in th e P r esen ce of Hyd r ogen Su lfid e. Nucleo-
side H-phosphonate 1a (1 molar equiv) and hydrogen
sulfide (3 molar equiv) in dichloromethane/pyridine
(9:1, v/v) were allowed to react in the presence of DPCP
(20) Stawin´ski, J .; Thelin, M. Nucleosides Nucleotides 1990, 9, 129-
135.
(21) J ankowska, J .; Sobkowski, M.; Stawin´ski, J .; Kraszewski, A.
Tetrahedron Lett. 1994, 35, 3355-3358.
(22) Seeberger, P. H.; Caruthers, M. H. Tetrahedron Lett. 1995, 36,
695-698.
(23) Stawin´ski, J .; Stro¨mberg, R.; Thelin, M. Nucleosides Nucleotides
1991, 10, 511-514.
(24) Stawin´ski, J .; Szabo´, T.; Thelin, M.; Westman, E.; Zain, R.
Collect. Czech. Chem. Commun. 1990, 55, 141-144.
(25) Stawin´ski, J .; Zain, R. Nucleosides Nucleotides 1995, 14, 711-
714.
(26) For assignments of signals in the 31P NMR spectra to particular
products or intermediates, see Experimental Section.

Aryl H-Phosphonates. 12. Synthetic and 31P NMR Studies
J . Org. Chem., Vol. 65, No. 21, 2000 7051
species involved in the former pathway27 that made the
formation of H-phosphonodithioate 3a from H-phospho-
nate 1a almost 1 order of magnitude faster. A possible
reactive species that could participate in above reaction
was assumed to be mixed phosphonic-thiophosphonic
anhydride 6. One can envisage its formation by the
reaction of the initial product of sulfhydrolysis, H-
phosphonothioate 2a , with activated H-phosphonate
derivative 4, generated from H-phosphonate 1a and the
condensing agent. This activation pathway for 2a is
available only when H-phosphonate 1a is subjected to
sulfhydrolysis in the presence of DPCP.
The intermediacy of 6 in transformation 1a f 3a and
its fast reaction to dithio derivative 3a implied that in
the presence of a limited amount of a condensing agent,
the reaction should afford equimolar amounts of 3a and
1a . Indeed, the 31P NMR spectra of the reaction mixture
from sulfhydrolysis of H-phosphonate 1a carried out in
the presence of 1.5 equiv of DPCP revealed the formation
of 1a (47%) and 3a (53%) in the expected ratio.
In separate experiments we found that when an
equimolar mixture of nucleoside 5′-H-phosphonate 728
and nucleoside 3′-H-phosphonothioate 2a was subjected
to sulfhydrolysis under analogous reaction conditions, it
was practically only 2a that underwent thiation to 3a .
The reaction was rapid (<3 min) and resulted in a
complete consumption of 3′-H-phosphonothioate 2a , while
5′-H-phosphonate 7 was converted to dithio derivative 8
only in ca. 10%. These results were in line with our
assumption concerning the involvement of phosphonic-
thiophosphonic anhydride 6 (vide supra) in the conver-
sion of H-phosphonates 1 into H-phosphonodithioates 3.
The postulated higher reactivity of this intermediate vs
thiophosphonic-phosphoric anhydride 5 can be rational-
ized on the grounds of a generally higher electrophilicity
of phosphorus centers in uncharged tetracoordinated
P(III) derivatives compared to those of the corresponding
P(V) compounds.29
The other point that deserves some comments is
chemoselectivity observed in the investigated reactions.
The initial product of activation of H-phosphonate 1a ,29
phosphonic-phosphoric anhydride 4, has two electrophilic
centers, P(III) and P(V), and since the former one is
softer,30 it should be attacked preferentially by soft
nucleophiles, e.g., HS-. In agreement with this, we
observed (31P NMR) exclusive formation of thiophospho-
nate derivatives and no products due to attack of sulfide
anion on the P(V) center in 4. The origin of chemoselec-
tivity in sulfhydrolysis of the postulated intermediate 6
is less obvious, but this can also be explained on the
ground of the HSAB principle.30 In this instance, the
presence of sulfur probably decreases electrophilicity of
the thiophosphonyl center in 6, but the higher acidity of
the phosphorus-bound hydrogen23 makes this center
softer by virtue of its higher tendency to form tervalent
species and thus more susceptible to attack by soft
nucleophiles.
The above results lend support to our earlier observa-
tion11 that, as a result of higher susceptibility of H-
phosphonothioate vs H-phosphonate derivatives to sulf-
hydrolysis, it can be difficult to stop sulfhydrolysis of
H-phosphonate monoesters promoted by condensing agents
at the stage of monothio derivative of type 2. However,
the observed tendency of the reactions investigated to
afford dithio derivatives 3a could be exploited for pre-
parative purposes. Indeed, using a synthetic protocol for
sulfhydrolysis of H-phosphonate monoesters based on
this reaction, various nucleoside H-phosphonodithioate
monoesters of type 3 were obtained in over 90% yields
(see Experimental Section).
Rea ction s of Ar yl Nu cleosid e H-P h osp h on a tes
w ith H2S a n d Hexa m eth yld isila th ia n e. Recently, we
have showed that nucleoside phenyl H-phosphonates
(generated in situ from suitably protected nucleosides
with diphenyl H-phosphonates) afforded the expected
nucleoside H-phosphonothioates of type 2 in high yields18
when treated with hydrogen sulfide in the presence of
triethylamine (TEA). However, since the indispensable
component of this reagent system is a strong base
(triethylamine),31 these reaction conditions are not com-
patible with base-sensitive substrates. Therefore, we
investigated the possibility of achieving similarly favor-
able kinetics under less basic reaction conditions by
modulation of the reactivity of H-phosphonates of type
9 via varying electron density in aromatic rings of the
aryl moieties.
To this end, nucleoside phenyl 9a and nucleoside
p-chlorophenyl 9b H-phosphonates (prepared in situ as
described previously32) were subjected to sulfhydrolysis
(3 molar equiv of H2S) in pyridine. As revealed by 31P
NMR spectroscopy, these compounds reacted with hy-
drogen sulfide at different rates (completion after ca. 60
min for phenyl derivative 9a and 30 min for p-chlorophe-
nyl derivative 9b) but afforded similar mixtures of
H-phosphonate 1a , H-phosphonothioate 2a , and H-
phosphonodithioate 3a (27%, 38%, and 35%, respectively,
for 9a and 23%, 37%, and 40% for 9b).
These results can best be rationalized on the ground
of a mechanism similar to that of sulfhydrolysis of 1a in
the presence of DPCP proposed above, involving phos-
phonic-thiophosphonic anhydride 6 as an intermediate,
but with slightly different rates for the formation of
monothio 2a and dithio 3a products.
Analogous reactions carried out on aryl H-phospho-
nates 9c-e33 bearing more acidic phenol moieties showed
that the increasing electrophilicity at the phosphorus
center in these compounds was counterproductive and
favored instead the formation of dithio derivative 3a .
Irrespective of the substrate 9c-e used, these reactions
always afforded mixtures containing almost equimolar
amounts of H-phosphonates 1 and dithio derivatives 3,
(27) Since in condensation reactions, the activation of H-phospho-
nothioate monoesters is usually a rate determining step,11 one can
argue that the observed differences in overall rates of 1a f 3a and 2a
f 3a are due to faster formation of 5 in the former pathway. However,
there in no obvious reactive species that could more effectively than
DPCP under the reaction conditions transform 2a into phosphoric-
thiophosphonic mixed anhydride 5, and thus this option remains less
likely.
(28) Reactivity of H-phosphonates 1a and 7 are similar. For this
particular experiment we chose nucleoside 5′-H-phosphonate 7 as its
thiation product 8 has a 31P NMR chemical shift and the splitting
pattern distinctive from that of 3a , and its use permitted the evaluation
of two reaction pathways, 7 f 8 and 2a f 3a .
(29) Stawin´ski, J . In Handbook of Organophosphorus Chemistry;
Engel, R., Ed.; Marcel Dekker: New York, 1992; pp 377-434.
(30) Pearson, R. G. J . Am. Chem. Soc. 1963, 85, 3533-3539.
(31) The presence of triethylamine made the sulfhydrolysis of 9a
proceed fast enough to suppress the formation of the secondary product,
H-phosphonodithioate 3a .
(32) Cies´lak, J .; Szymczak, M.; Wenska, M.; Stawin´ski, J .; Krasze-
wski, A. J . Chem. Soc., Perkin Trans. 1 1999, 3327-3331.
(33) These reactions were carried out in dichloromethane in the
presence of 10 equiv of pyridine as aryl H-phosphonates 9a -e cannot
be efficiently generated in neat pyridine.

7052 J . Org. Chem., Vol. 65, No. 21, 2000
Cies´lak et al.
Ta ble 1. 31P NMR Da ta of th e P r od u cts of Typ e 2 a n d 3
a n d Som e Rea ction s In ter m ed ia tesa
with complete absence of monothio products 2. Attempted
sulfhydrolysis of 9c-e with mixtures of hydrogen sulfide/
TEA [also in the presence of trimethylsilyl chloride
(TMSCl) to capture the initial product of sulfhydrolysis,
2a ] invariably led to H-phosphonate 1a and H-phospho-
nodithioate 3a in a 1:1 ratio.
Most striking was the absence of trimethylsilylated
derivatives of H-phosphonothioate 2a in the reaction
mixtures when sulfhydrolysis of 1a was carried out in
the presence of TEA and TMSCl. This suggested that in
the presence of TEA these reactions might proceed via
another pathway that did not involve formation of
monothio derivative 2a . This assumption was corrobo-
rated by the results of two experiments. First, when aryl
H-phosphonate 9e was treated in dichloromethane in the
presence of pyridine (10 equiv) with hydrogen sulfide and
compds
δP
1J HP (Hz)
3J HP (Hz)
2a
2b
2c
2d
2e
3a
3b
3c
3d
3e
7
8
9a -e
10
55.43, 55.85b
54.42, 54.50b
53.99, 55.29b
54.10, 54.88b
54.96, 56.55b
84.15
83.69
84.44
84.35
86.00
583.9, 579.3
583.0, 581.2
587.7, 578.4
582.8, 585.8
579.3, 587.6
546.9
545.0
546.9
547.8
551.5
12.6c
11.6c
11.6c
11.6, 13.4c
13.9c
14.8d
14.8d
13.9d
14.8d
15.8d
7.4e
4.41
637.7
530.2
86.39
9.3e
2.0-5.0f
130.12
9.27g
9.3, 10.2h
10.3h
11a
11b
11c
11d
11e
123.04, 124.07b
127.31, 127.91b
122.95, 123.72b
121.61, 123.88b
125.42, 126.10b
TMSCl (5 equiv), trimethylsilylated derivative of 2a was
10.3h
1
formed exclusively (δP ) 57.10 and 56.55 ppm, J PH
)
9.3h
3
9.3h
658 and 660 Hz; J PH ) 11.1 and 12.8 Hz, dd). Second,
H-phosphonate 9e in the presence of triethylamine
underwent rapid disproportionation34 to nucleoside bisar-
yl phosphite 10 and H-phosphonate monoester 1a , and
the former one underwent rapid conversion to dithio
derivative 3a upon the addition of hydrogen sulfide.
Since both transformations were clean and fast, we
attempted to develop them as synthetic procedures for
the preparation of nucleoside H-phosphonothioate and
nucleoside H-phosphonodithioate monoesters, respec-
tively. Considering that efficient formation of H-phospho-
nothioate monoesters of type 2 from the corresponding
H-phosphonate derivatives 9 required silylation of the
substrate prior to sulfhydrolysis, we investigated the
possibility of using 1,1,1,3,3,3-hexamethyldisilathiane35
(HMDST), a reagent that simultaneously can act as a
silylating agent and as a source of sulfide ion.36
a
Spectra recorded in dichloromethane containing pyridine (ca
10% v/v). Two diastereoisomers. c Two doublets of doublets.
b
Doublet of doublets. e Doublet of triplets. f Compounds 9a -e
d
resonate a few ppm downfield from H3PO4, and their 31P NMR
data have been published earlier.32 Doublet. Two doublets.
g
h
2,3,4,5,6-pentachlorophenol) can be efficiently generated
in pyridine from the corresponding nucleoside H-phos-
phonates 1 and the appropriate phenols in the presence
of DPCP,32 we tried to use them as intermediates in the
synthesis in of nucleoside H-phosphonodithioates 3.
Indeed, we found that these compounds when treated in
pyridine with HMDST (5 equiv) rapidly (<5 min) and
cleanly afforded the corresponding nucleoside H-phospho-
nodithioates of type 3. A synthetic usefulness of this
reaction was assessed on a preparative scale and nucleo-
side H-phosphonodithioates derived from thymidine, 2′-
deoxyadenosine, 2′-deoxycytidine and 2′-deoxyguanosine
(3a -d ) were obtained in excellent yields (>90%) after
silica gel chromatography (see Experimental Section).
When aryl H-phosphonates 9a -e (1 molar equiv) in
dichloromethane containing pyridine (10 molar equiv)
were treated with HMDTS (3 molar equiv), 31P NMR
spectroscopy showed rapid disappearance of the starting
materials (<3 min) and clean formation of products
resonating in the range of chemical shifts of tervalent
phosphorus compounds (usually two signals at ca. 125
ppm, see Table 1). On the basis of spectral data and the
observed chemical reactivity, we assigned to these species
structures of aryl nucleoside silyl phosphites 11. Al-
though all 11a -e were rapidly formed, they differed
significantly in their reactivity toward trimethylsilylsul-
fide ion generated from HMDST. Thus, phenyl and
p-chlorophenyl derivatives (11a and 11b) required more
than 24 h for a complete conversion to the corresponding
silylated derivatives of 2, 2,4-dichlorophenyl and 4-ni-
trophenyl derivatives (11c and 11d ) ca. 60 min, and for
2,4,6-trichlorophenyl derivative 11e, the reaction was
complete in less than 5 min. The latter reaction also
proved to be most efficient when carried out on a
preparative scale and afforded nucleoside H-phosphono-
thioates 2a -d in high yield (> 90%) after silica gel
chromatography.
It is worth noting that the methods developed seem to
be applicable also to ribonucleoside 3′-H-phosphonates,
as it was apparent from efficient conversion of uridine
3′-H-phosphonate 1e (bearing a bulky 2′-O-t-butyldim-
ethylsilyl protective group) into the corresponding H-
phosphonothioate 2e or H-phosphonodithioate 3e (yields
> 90%).
In conclusion, on the basis of 31P NMR studies, we
developed new protocols for transformations of nucleoside
H-phosphonate monoesters of type 1 into the correspond-
ing H-phosphonothioate 2 and H-phosphonodithioate 3
derivatives under mild reaction conditions. Nucleoside
H-phosphonothioates 2a -e were obtained in high yields
by reacting the in situ generated nucleoside aryl H-
phosphonates of type 9 with HMDST. For the preparation
of nucleoside H-phosphonodithioates 3a -e, two alterna-
tive procedures were developed. These consisted of the
reaction of H-phosphonate monoesters 1 with diphenyl
chlorophosphate (DPCP) in the presence of hydrogen
sulfide or sulfhydrolysis of the in situ produced nucleoside
diaryl phosphites of type 10. Since all these transforma-
tions were fast and efficient and could be carried out as
one-pot reactions, they can be recommended as versatile
and convenient methods for the preparation of nucleoside
H-phosphonothioates and nucleoside H-phosphonodithio-
ates from readily accessible nucleoside H-phosphonates.
Since nucleoside bisaryl phosphites of type 10 derived
from highly acidic phenols (e.g., 2,4,6-trichlorophenol or
(34) Kers, A.; Kers, I.; Stawin´ski, J .; Sobkowski, M.; Kraszewski,
A. Tetrahedron 1996, 52, 9931-9944.
(35) Eaborn, C. J . Chem. Soc. 1950, 3077-3089.
(36) The transformation of aryl H-phosphonates 9 to the correspond-
ing monothio derivatives 2 can also be effected using mixture of TMSCl
with hydrogen sulfide.

Aryl H-Phosphonates. 12. Synthetic and 31P NMR Studies
J . Org. Chem., Vol. 65, No. 21, 2000 7053
Hz). For 31P NMR data, see Table 1. Calcd for C37H48N3O8PS:
C, 61.23; H, 6.67; N, 5.79. Found: C, 61.19; H, 6.72; N, 5.74.
5′-O-Dim eth oxytr ityl-N6-ben zoyl-2′-d eoxya d en osin e 3′-
H-P h osph on oth ioate, Tr ieth ylam m on iu m Salt (2b). Yield,
0.15 g (90%); RTPH 1.91 (A), 1.99 (B), 1.39 (C); δH (CDCl3) 1.15
Exp er im en ta l Section
Ma ter ia l a n d Meth od s. 1H and 31P NMR spectra were
recorded at 300 and 121 MHz respectively. The 31P NMR
experiments were carried out at 25 °C in 5 mm tubes using
0.1 M concentrations of phosphorus-containing compounds in
appropriate solvents (0.6 mL), and 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
the following solvent systems: (A) CH3Cl/CH3OH 9:1 (v/v); (B)
iPrOH/NH3 aqueous concentrated/H2O 85:5:10 (v/v/v); (C)
iPrOH/NH3 aqueous concentrated/H2O 7:2:1 (v/v/v). TLC mo-
bilities (RTPH) are reported relative to 5′-O-dimethoxytritylthy-
midine 3′-H-phosphonate. Pyridine (LabScan Ltd.) was stored
over molecular sieves 4Å until the amount of water was below
20 ppm (Karl Fischer coulometric titration). Dichloromethane
(POCH, Poland) was dried with P2O5, distilled, and stored over
molecular sieves 4Å until the amount of water was below 10
ppm. 1,1,1,3,3,3-Hexamethyldisilathiane, diphenyl chlorophos-
phate, and trimethylsilyl chloride were commercial grade from
Aldrich. Stock solution of hydrogen sulfide (1 M) in dioxane
was prepared by passing H2S through the solvent until
saturation.
5′-O-Dimethoxytritylated nucleoside 3′-H-phosphonates 1a-e
were obtained according to published method.21 Nucleoside
aryl H-phosphonate 9a -e were generated in situ as described
previously.32 Nucleoside 5′-H-phosphonate 721 and nucleoside
5′-H-phosphonodithioate 824 were obtained analogously to
published procedures. The reference compounds used for the
assignment of certain 31P NMR resonances, were obtained as
follows: 5′-O-dimethoxytritylthymidin-3′-yl trimethylsilyl H-
phosphonothioate, by the reaction of 2a with trimethylsilyl
chloride in dichloromethane/pyridine;37 bis(2,4,6-trichlorophe-
nyl) 5′-O-dimethoxytritylthymidin-3′-yl phosphite 10, by treat-
ment of H-phosphonate 9e with diisopropylamine;32 2,4,6-
trichlorophenyl trimethylsilyl 5′-O-dimethoxytritylthymidin-
3′-yl phosphites 11a -e, by treatment of the corresponding
H-phosphonates 9a -e with trimethylsilyl chloride in dichlo-
romethane/pyridine.
3
3
(9H, t, J HH ) 7.2 Hz), 2.75 (6H, q, J HH ) 7.2 Hz), 2.81 (1H,
m), 2.91 (1H, m), 3.42 (2H, m), 3.77 (6H, s), 4.44 and 4.46 (1H,
m, m), 5.35 (1H, m), 6.60 (1H, m), 6.77-7.61 (18H, m), 8.05
3
and 8.10 (1H, 2d, J HP ) 582.8 and 581 Hz), 8.19 (1H, s), 8.73
(1H, s). For 31P NMR data see Table 1. Calcd for C44H51N6O7-
PS: C, 62.99; H, 6.13; N, 10.02. Found: C, 63.06; H, 6.18; N,
9.98.
5′-O-Dim et h oxyt r it yl-N4-b en zoyl-2′-d eoxycyt id in e 3′-
H-P h osp h on oth ioa te, Tr ieth yla m m on iu m Sa lt (2c). Yield,
0.22 g (90%); RTPH 1.77 (A), 1.88 (B), 1.39 (C); δH (CDCl3) 1.32
(9H, q, 3J HH ) 7.2 Hz), 2.37 (1H, m), 2.90 (1H, m), 3.08 (6H, q,
3J HH ) 7.2 Hz), 3.47 (2H, m), 3.80 (6H, s), 4.35 and 4.46 (1H,
3
m, m), 5.16 and 5.30 (1H, m, m), 6.31 (1H, t, J HH ) 6.0 Hz),
6.80-7.54 (18H, m), 7.88 (1H, d, 3J HH ) 7.2 Hz), 8.00 and 8.06
1
3
(1H, 2d, J HP ) 587.0 and 574.7 Hz), 8.24 (1H, d, J HH ) 7.2
Hz). For 31P NMR data see Table 1. Calcd for C43H51N4O8PS:
C, 63.38; H, 6.31; N, 6.88. Found: C, 63.44; H, 6.39; N, 6.79.
5′-O -D im e t h o x y t r it y l-N 2 -is o b u t y r y l-2′-d e o x y g u a -
n osin e 3′-H-P h osp h on oth ioa te, Tr ieth yla m m on iu m Sa lt
(2d ). Yield, 0.22 g (87%); RTPH 1.05 (A), 1.88 (B), 1.39 (C); δH
3
3
(CDCl3) 1.04 (3H, d, J HH ) 6.9 Hz), 1.13 (3H, d, J HH ) 6.9
3
3
Hz), 1.28 (9H, t, J HH ) 7.2 Hz), 2.31 (1H, heptet, J HH ) 6.9
Hz), 2.61 (1H, m), 3.04 (6H, q, J HH ) 7.2 Hz), 3.12 (1H, m),
3
3.22 (1H, m), 3.37 (1H, m), 3.73 (1H, m), 3.76 (6H, s), 4.21
and 4.22 (1H, 2m), 5.74 (1H, m), 6.17 (1H, m), 6.73-7.43 (13H,
m), 7.76 (1H, s), 7.93 and 8.06 (1H, 2d, 1J HP ) 586.1 and 581.3
Hz), 11.91 (1H, br s, exch. D2O). For 31P NMR data see Table
1. Calcd for C41H53N6O8PS: C, 59.99; H, 6.51; N, 10.24.
Found: C, 60.01; H, 6.59; N, 10.19.
5′-O-Dim et h oxyt r it yl-2′-O-d im et h ylt er t b u t ylsilylu r i-
d in e 3′-H -P h osp h on ot h ioa t e, Tr iet h yla m m on iu m Sa lt
(2e). Yield, 0.16 g (91%); RTPH 2.18 (A), 2.03 (B), 1.37 (C); δH
(CDCl3) 0.09-0.20 (6H, m), 0.89 and 0.91 (9H, 2s), 1.31 (9H,
3
3
t, J HH ) 7.2 Hz), 3.06 (6H, q, J HH ) 7.2 Hz), 3.54 (2H, m),
3.79 (6H, s), 4.35 and 4.49 (1H, 2 m), 4.43 (1H, m), 5.06 and
5.18 (1H, 2m), 5.16 and 5.22 (1H, 2d, 3J HH ) 8.4 Hz), 5.90 and
The assignments of signals in the 31P NMR spectra to
particular products or intermediates were done 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 a compound in question from reaction mixtures. Multiplicity
of some signals in 1H NMR spectra of products 2 are due to
P-diastereomers.
3
6.06 (1H, 2d, J HH ) 3.0 Hz and 6.0 Hz), 6.82-7.43 (13H,
3
m),7.85 and 8.08 (1H, 2d, J HH ) 8.4 Hz), 7.96 and 8.17 (1H,
2d, 3J HP ) 578.9 Hz and 586.9 Hz). For 31P NMR data see Table
1. Calcd for C42H60N3O9PSSi: C, 59.91; H, 7.18; N, 4.99.
Found: C, 59.94; H, 7.21; N, 5.02.
Gen er a l P r oced u r e for Syn th esis of Nu cleosid e H-
P h osp h on od ith ioa tes 3a -e. Meth od A. Nucleoside H-
phosphonate 1 (triethylammonium salt, 0.1 mmol) was ren-
dered anhydrous by repeated evaporation of added excess
pyridine and then dissolved in the same solvent (1 mL). To
this, hydrogen sulfide (3 molar equiv; 0.3 mL of 1 M stock
solution) was added, followed by diphenyl chlorophosphate (3
molar equiv). After 5 min the reaction mixture was quenched
with water and evaporated to viscous oil. The residue was
dissolved in a minimum volume of dichloromethane and
purified by a silica gel column chromatography using dichlo-
romethane/methanol/triethylamine (95:3:2 v/v/v) as an eluent.
Fractions containing triethylammonium salts of pure product
3 were collected and freeze-dried from benzene to afford white
solids.
Meth od B. Nucleoside H-phosphonate 1 (triethylammo-
nium salt, 0.1 mmol) and 2,4,6-trichlorophenol or 1,2,3,4,5-
pentachlorophenyl (3 molar equiv) were rendered anhydrous
by repeated evaporation of added excess pyridine. The residue
was dissolved in pyridine (1 mL) and treated with diphenyl
chlorophosphate (3 molar equiv). When the formation of
nucleoside bisaryl phosphite was complete (ca 30 min; 31P
NMR) 1,1,1,3,3,3-hexamethyldisilathiane (6 molar equiv) was
added. After 10 min the reaction mixture was concentrated,
and the residue was dissolved in dichloromethane and washed
with saturated aqueous NaHCO3. The organic layer was
separated, dried over anhydrous Na2SO4, and concentrated.
Nucleoside H-phosphonodithioates 3 (triethylammonium salts)
were isolated as described in Method A.
Gen er a l P r oced u r e for Syn th esis of Nu cleosid e 3′-H-
P h osp h on oth ioa tes 2a -e. Nucleoside H-phosphonate 1 (tri-
ethylammonium salt, 0.1 mmol) and 2,4,6-trichlorophenol (1.1
molar equiv) were rendered anhydrous by repeated evapora-
tion of added excess pyridine. The residue was dissolved in
dichloromethane (1 mL) containing pyridine (12 molar equiv)
and treated with diphenyl chlorophosphate (1.1 molar equiv).
When the formation of aryl nucleoside H-phosphonate was
complete (ca 5 min; 31P NMR), 1,1,1,3,3,3-hexamethyldisi-
lathiane (3 molar equiv) was added, and after 5 min the
reaction mixture was diluted with dichloromethane (15 mL)
and washed with saturated aqueous NaHCO3 (3 × 10 mL).
The organic phase was separated, dried over anhydrous Na2-
SO4, and concentrated to an oil. Nucleoside H-phosphonothio-
ates 2 were purified by a silica gel column chromatography
using dichloromethane/methanol/triethylamine (95:3:2 v/v/v)
as an eluent. Products 2 (ca. 1:1 mixture of diastereomers)
were obtained as white solids (triethylammonium salt) after
freeze-drying from benzene.
5′-O-Dim eth oxytr itylth ym id in e 3′-H-P h osp h on oth io-
a te, Tr ieth yla m m on iu m Sa lt (2a ). Yield, 0.21 g (96%); RTPH
3
1.82 (A), 1.24 (B), 1.07 (C); δH (CDCl3) 1.33 (9H, t, J HH ) 7.2
Hz), 1.36 and 1.38 (3H, br s), 2.38 (1H, m), 2.61 (1H, m), 3.08
3
(6H, q, J HH ) 7.2 Hz), 3.45 (2H, m), 3.79 (6H, s), 4.26 and
4.36 (1H, m), 5.32 (1H, m), 6.46 (1H, m), 6.81-7.60 (13H, m),
1
7.41 (1H, br s), 8.00 and 8.06 (1H, 2d, J HP ) 583.7 and 579.8
(37) Zain, R.; Stawin´ski, J . J . Org. Chem. 1996, 61, 6617-6622.

7054 J . Org. Chem., Vol. 65, No. 21, 2000
Cies´lak et al.
Compounds 3a -e produced by methods A and B were
indistinguishable by spectral methods (1H, 31P NMR), had
identical TLC mobilities (systems A-C), and gave similar
(within the experimental error) elemental analysis data.
5′-O-Dim eth oxytr itylth ym idin e 3′-H-P h osp h on od ith io-
a te, Tr ieth yla m m on iu m Sa lt (3a ). Yields: Method A 0.21
g (92%), Method B 0.22 g (96%). RTPH 1.73 (A), 2.00 (B), 1.39
5′-O-Dim e t h oxyt r it yl-N 2-isob u t yr yl-2′-d e oxygu a n o-
sin e 3′-H-P h osp h on od ith ioa te, Tr ieth yla m m on iu m Sa lt
(3d ). Yields: Method A 0.22 g (87%), Method B 0.23 g (91%).
3
RTPH 1.73 (A), 1.94 (B), 1.40 (C); δH (CDCl3) 1.07 (3H, d, J HH
3
3
) 6.8 Hz), 1.14 (3H, d, J HH ) 6.7 Hz), 1.15 (9H, t, J HH ) 7.2
3
Hz), 2.39 (1H, heptet, J HH ) 6.8 Hz), 2.69 (1H, m), 3.05 (1H,
3
m), 3.15 (6H, q, J HH ) 7.2 Hz), 3.34 (2H, m), 3.76 (6H, s),
3
4.33 (1H, m), 5.73 (1H, dq, 3J HP ) 14.9 Hz J HH ) 4.8 Hz), 6.20
(C); δH (CDCl3) 1.37 (3H, br s), 1.37 (9H, t, J HH ) 7.2 Hz),
3
2.38 (1H, m), 2.72 (1H, m), 3.25 (6H, q, J HH ) 7.2 Hz), 3.45
3
(1H, t, J HH ) 6.3 Hz), 6.74-7.42 (13H, m), 7.81 (1H, s), 8.70
3
(2H, m), 3.79 (6H, s), 4.37 (1H, m), 5.38 (1H, dm, J HP ) 14.9
(1H, d, J HP ) 547.7 Hz), 12.01 (1H, br m, exch. D2O). For 31P
1
Hz), 6,45 (1H, m), 6.38-7.41 (13H, m), 7.61 (1H, br s), 8.37
NMR data, see Table 1. Calcd for C41H53N6O7PS2: C, 58.83;
H, 6.38; N, 10.04. Found: C, 58.91; H, 6.43; N, 9.97.
1
(1H, br m, exch. D2O), 8.70 (1H, d, J HP ) 547.1 Hz). For 31P
NMR data, see Table 1. Calcd for C37H48N3O7PS2: C, 59.90;
H, 6.52; N, 5.66. Found: C, 59.81; H, 6.56; N, 5.64.
5′-O-Dim et h oxyt r it yl-2′-O-d im et h ylt er t b u t ylsilylu r i-
d in e 3′-H-P h osp h on oth ioa te (3e). Yields: Method A 0.17
g (96%), Method B 0.16 g (90%). RTPH 2.41 (A), 2.18 (B), 1.52
(C); δH (CDCl3) 0.16 (3H, s), 0.21 (3H, s), 0.91 (9H, s), 1.35
5′-O-Dim eth oxytr ityl-N6-ben zoyl-2′-d eoxya d en osin e 3′-
H-P h osp h on od ith ioa te, Tr ieth yla m m on iu m Sa lt (3b).
Yields: Method A 0.15 g (95%), Method B 0.15 g (90%). RTPH
3
3
3
(9H, t, J HH ) 7.5 Hz), 3.21 (6H, q, J HH ) 7.5 Hz), 3.58 (2H,
1.95 (A), 2.41 (B), 1.57 (C); δH (CDCl3) 1.36 (9H, t, J HH ) 7.2
3
3
m), 3.79 (6H, s), 4.45 (2H, m), 5.13 (1H, d, J HH ) 8.2 Hz),
Hz), 2.39 (2H, m), 3.19 (6H, q, J HH ) 7.2 Hz), 3.44 (2H, m),
3
3
3.77 (6H, s), 4.53 (1H, m), 5.45 (1H, dm, 3J HP ) 12.3 Hz), 6.59
(1H, m), 6.75-8.0 (18H, m), 8.19 (1H, s), 8.72 (1H, s), 8.77 (1H,
5.21 (1H, dm, J HP ) 15.7 Hz), 5.93 (1H, d, J HH ) 3.3 Hz),
3
6.82-7.43 (13H, m), 8.05 (1H, d, J HH ) 8.2 Hz), 8.20 (1H, br
m, exch. D2O), 8.76 (1H, d, J HP ) 551.3 Hz). For 31P NMR
1
1
d, J HP ) 547.1 Hz), 9.14 (1H, m, exch. D2O). For 31P NMR
data, see Table 1. Calcd for C42H60N3O8PS2Si: C, 58.78; H,
7.05; N, 4.90. Found: C, 58.80; H, 7.11; N, 4.84.
data, see Table 1. Calcd for C44H51N6O6PS2: C, 61.81; H, 6.01;
N, 9.83. Found: C, 61.91; H, 6.09; N, 9.76.
5′-O-Dim et h oxyt r it yl-N4-b en zoyl-2′-d eoxycyt id in e 3′-
H-P h osp h on od ith ioa te, Tr ieth yla m m on iu m Sa lt (3c).
Yields: Method A 0.22 g (90%), Method B 0.22 g (90%). RTPH
Ack n ow led gm en t. We are grateful to Prof. Prof. M.
Wiewio´rowski and P. J . Garegg for their kind interest
in this work. Financial support from the State Com-
mittee for Scientific Research, Republic of Poland (grant
3 T09A 118 14), the Swedish Natural Science Research
Council, and the Swedish Foundation for Strategic
Research is gratefully acknowledged.
3
1.91 (A), 2.35 (B), 1.59 (C); δH (CDCl3) 1.32 (9H, t, J HH ) 7.2
3
Hz), 2.41 (1H, m), 2.94 (1H, m), 3.09 (6H, q, J HH ) 7.2 Hz),
3
3.49 (2H, m), 3.80 (6H, s), 4.43 (1H, m), 5.29 (1H, dm, J HP
)
14.1 Hz), 6.32 (1H, t, 3J HH ) 6.0 Hz), 6.84-7.60 (18H, m), 7.88
3
3
(1H, d, J HH ) 7.2 Hz), 8.28 (1H, d, J HH ) 7.2 Hz), 8.76 (1H,
d, J HP ) 547.1 Hz). For 31P NMR data, see Table 1. Calcd for
1
C
43H51N4O7PS2: C, 62.15; H, 6.19; N, 6.74. Found: C, 62.07;
H, 6.24; N, 6.69.
J O000729Q