Stereochemistry of internucleotide bond formation by the H-phosphonate method. 7. Stereoselective formation of ribonucleoside (RP)- and (SP)-3′-H-phosphonothioate monoesters

Article abstract of DOI:10.1016/j.tetasy.2010.01.022

Ribonucleoside 3′-H-phosphonothioate monoesters exist in the form of (R_P)- and (S_P)-diastereomers. In order to obtain them in good yields and in high stereochemical purity, stereoselective strategies for their preparation were investigated. For the synthesis of the (R_P)-isomer, a stereoselective sulfhydrolysis of an activated nucleoside H-phosphonate was developed, while the monoesters with an (S_P)-configuration were prepared by asymmetric transformation of diastereomeric mixtures of nucleoside 3′-H-phosphonothioates using either a condensation with 9-fluorenemethanol, followed by β-elimination, or via pivaloylation-hydrolysis reaction sequence. A tentative assignment of the absolute configurations of the obtained diastereomers of 3′-H-phosphonothioate esters was carried out via a stereochemical correlation analysis.

Full text of DOI:10.1016/j.tetasy.2010.01.022

Tetrahedron: Asymmetry 21 (2010) 410–419
Contents lists available at ScienceDirect
Tetrahedron: Asymmetry
journal homepage: www.elsevier.com/locate/tetasy
Stereochemistry of internucleotide bond formation by the H-phosphonate
method. 7. Stereoselective formation of ribonucleoside (RP)- and
(SP)-30-H-phosphonothioate monoesters
b
a
Michal Sobkowski a, , Jacek Stawinski , Adam Kraszewski
*
a Institute of Bioorganic Chemistry, Polish Academy of Sciences, Noskowskiego 12/14, 61-704 Poznan, Poland
b Department of Organic Chemistry, Arrhenius Laboratory, Stockholm University, S-106 91 Stockholm, Sweden
a r t i c l e i n f o
a b s t r a c t
Ribonucleoside 30-H-phosphonothioate monoesters exist in the form of (RP)- and (SP)-diastereomers. In
order to obtain them in good yields and in high stereochemical purity, stereoselective strategies for their
preparation were investigated. For the synthesis of the (RP)-isomer, a stereoselective sulfhydrolysis of an
activated nucleoside H-phosphonate was developed, while the monoesters with an (SP)-configuration
were prepared by asymmetric transformation of diastereomeric mixtures of nucleoside 30-H-phosphono-
thioates using either a condensation with 9-fluorenemethanol, followed by b-elimination, or via pivaloy-
lation-hydrolysis reaction sequence. A tentative assignment of the absolute configurations of the
obtained diastereomers of 30-H-phosphonothioate esters was carried out via a stereochemical correlation
analysis.
Article history:
Received 5 January 2010
Accepted 28 January 2010
Available online 1 March 2010
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
tion of H-phosphonothioate diesters from diastereomerically pure
precursors, their P-epimerisation might be expected to occur at the
Ribonucleoside 30-H-phosphonates react with nucleosides and
other alcohols with high diastereoselectivity towards (SP)-diaste-
reomers (de >80%), which are substrates for stereospecific
transformations to various P-chiral derivatives, for example (RP)-
phosphorothioate diesters.1,2 This strategy has been successfully
applied in the preparation of all-(RP)-oligo(ribonucleotide phosp-
horothioate)s using a combination of chemical synthesis and
post-synthetic enzymatic processing.1 More recently, it was found
that a plausible source of asymmetric induction in these reactions
is the dynamic kinetic asymmetric transformation (DYKAT).3 Ste-
reocontrolled preparation of ribonucleoside phosphorothioate
diesters with an opposite [i.e., (SP)] sense of P-chirality may be
achieved by using ribonucleoside 30-H-phosphonothioate monoes-
ters as starting materials, since their condensations with alcohols
and nucleosides are stereoselective, while the main diastereomers
formed, (SP)-H-phosphonothioate diesters, could be oxidized ste-
reospecifically to (SP)-phosphorothioates.4
level of the H-phosphonothionic—carboxylic mixed anhydride, and
from the degree of stereospecificity, approximate rates of forma-
tion and collapse of reactive intermediates can be inferred. This
would be helpful in mechanistic studies in the early stages of H-
phosphonate diester formation, in order to obtain a better under-
standing of the underlying chemistry or to evaluate differences in
the reactivity of the chalcogen atoms in (SP)- and (RP)-nucleoside
H-phosphonothioates towards electrophiles.
In addition to the applications in basic studies, the availability of
separate (RP)- and (SP)-diastereomers of nucleoside H-phosphono-
thioates could also be useful in the stereospecific synthesis of stereo-
chemically pure hypermodified nucleoside derivatives, for example,
dinucleoside5,6 or alkyl nucleoside7 phosphoramidothioates, cur-
rently only available present only as pools of diastereomers.
Herein we report our synthetic and mechanistic studies on the
preparation of both P-diastereomers of ribonucleoside 30-H-phos-
phonothioate monoesters, and the determination of their absolute
configurations.
In contrast to nucleoside H-phosphonate monoesters, the corre-
sponding H-phosphonothioates are chiral at the phosphorus centre
and may be separated into individual P-epimers. Access to isolated
(RP)- and (SP)-diastereomers of ribonucleoside H-phosphonothio-
ates might be helpful in mechanistic and synthetic studies.
According to the DYKAT mechanism (cf. Fig. 2), during the forma-
2. Results and discussion
In the deoxyribo series, the diastereomers of (50-O-DMTr)thymi-
dine 9-fluorenemethyl 30-H-phosphonothioate8,9 can be separated
chromatographically as FAST [Rf 0.42,à (SP)] and SLOW [Rf 0.35,
* Corresponding author. Tel.: +48 61 8528 503; fax: +48 61 8520 532.
E-mail address: msob@ibch.poznan.pl (M. Sobkowski).
à
n-Hexane–dichloromethane–ethyl acetate 5:5:1.
0957-4166/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2010.01.022

M. Sobkowski et al. / Tetrahedron: Asymmetry 21 (2010) 410–419
411
(RP)] isomers, and subsequently converted via b-elimination to the
corresponding nucleoside H-phosphonothioate monoesters§ in a ste-
reospecific manner.10 Unfortunately, and somewhat unexpectedly,
the corresponding diastereomers of 9-fluorenemethyl 50-O-(DMTr)-
20-O-(TBDMSi)uridine 30-H-phosphonothioate 8b (Fig. 2) were chro-
matographically indistinguishable irrespective of the stationary and
mobile phase used. We assumed that a possible reason for this could
be the high lipophilicity of the compounds, due to the presence in
three highly hydrophobic groups, namely dimethoxytrityl (DMTr),
TBDMSi, and 9-fluorenemethyl units.
Indeed, in contrast to the deoxy series, the negatively charged
diastereomers of (50-O-DMTr-20-O-TBDMSi)-uridine 30-H-phospho-
nothioate monoester 5 had different TLC mobilities and could be
resolved into individual species by silica-gel column chromatogra-
phy using 0–30% gradient of acetonitrile in 95:5 toluene–TEA.
Although this allowed a significant enrichment of each of the dia-
stereomers (de 40–50%), the mixed fractions still contained sub-
stantial amounts of both diastereomers (ca. 50% of the recovered
compound).
the 4-RP/SP ratio during the initial stages of the process presumably
reflected the stereochemistry of Path A. Accordingly, it was found
that upon sulfhydrolysis of the derivatives 3a and 3b, the 4-RP/SP
ratio was ca. 3:1 when HMDST was used as a silylating/sulfhydro-
lytic agent and ca. 2:1, for the TMS-Cl/H2S mixture. Sulfhydrolysis
of 3a and 3b probably owes its stereoselectivity to the DYKAT
mechanism,3 and thus, the configuration of the main diastereo-
mers of H-phosphonothioates 4 and 5 was tentatively assigned as
(RP).
Concluding the above experiments, in order to achieve highest
stereoselectivity, HMDST should be used rather than the TMS-Cl/
H2S mixture. Both pathways (A and B, Fig. 1) yielded the (RP)-dia-
stereomer of H-phosphonothioate 4 as the main isomer, however,
the Path B was less stereoselective [for R = Pv (pivolyl or timethyl
acetyl)] and slower than the Path A and should be avoided during
the sulfhydrolysis of H-phosphonates of type 3.
Having established the above, two approaches were developed
for the stereoselective synthesis of the FAST diastereomer of uri-
dine H-phosphonothioate 5. In the first one, 20,50-di-O-protected
nucleoside 1 was reacted with diphenyl H-phosphonate in pyri-
dine17 to form uridine phenyl H-phosphonate 3a, which was sulf-
hydrolysed13 in situ with HMDST (2.5 equiv).|| Diester 3a was
consumed completely within ca. 1 h (31P NMR) to yield the silylated
H-phosphonothioate 4 (ca. 84%) and phenyl trimethylsilyl uridine
phosphite 6a (ca. 14%). To this mixture additional 2 equiv of HMDTS
were added to prevent the formation of the H-phosphonodithioate
derivative11,13,14 and the reaction mixture was left overnight to com-
plete the sulfhydrolysis of phosphite 6a. After quenching with water,
31P NMR showed the formation of uridine H-phosphonothioate
monoester 5 in ca. 90% yield and a 5-FAST/5-SLOW ratio of ca. 3:1.
In the second method for the preparation of 5-FAST, uridine H-
phosphonate 2 was converted into H-phosphonic–pivalic mixed
anhydride 3b in the reaction with pivaloyl chloride in acetonitrile
containing 6 equiv of pyridine. The subsequent reaction with
HMDST (3 equiv) yielded the desired silylated H-phosphonothioate
4 accompanied by ca. 10% of the tervalent intermediate 6b. Sulfhy-
drolysis of the latter compound was accelerated by the addition of
2,6-lutidinium hydrochloride11 and the whole process was com-
pleted within 30 min (31P NMR yield: 94%; 4-FAST/4-SLOW 3:1).
After work-up, the main product 5-FAST was isolated chromato-
graphically with de >80% and in a yield of ca. 60% (the remaining
material was eluted as a diastereomeric mixture; total yield 86%).
In order to improve the efficiency of this procedure, we took
advantage of the stereoselective formation of each of the diastereo-
mers of H-phosphonothioate 5 under different experimental condi-
tions, and achieved a significant increase in the content of the
desired isomer in the reaction mixture which simplified its chro-
matographic separation.
2.1. Synthesis of (RP)-uridine 30-H-phosphonothioate (5-FAST)
In previous papers on the preparation of nucleoside 30-H-phos-
phonothioates it was noted that one of the diastereomers of uri-
dine H-phosphonothioate was formed in significant excess during
sulfhydrolysis of an activated uridine H-phosphonate; however, lit-
tle attention was paid to stereochemistry of the reaction.11–14 Two
kinds of reactive intermediates were exploited for the syntheses,
either aryl nucleoside H-phosphonate of type 3a or nucleoside H-
phosphonic–pivalic mixed anhydride 3b. These were sulfhydroly-
sed using hexamethylenedisilathiane (HMDST), which was either
a commercial product or was generated in situ in the reaction of
H2S with TMS-Cl. Herein we found that in all cases the main diaste-
reomer formed was the one migrating faster on silica-gel (5-FAST,
Fig. 1) and concluded that sulfhydrolysis of reactive H-phospho-
nate 3 may proceed via two pathways. In Path A the activated H-
phosphonate 3 is presumably attacked by TMS2S (HMDST) or
TMS-SH, followed by rapid S?O silyl migration and formation of
O-silylated compound 4.– In the alternative Path B, H-phosphonate
3 is silylated initially to a tervalent species of type 6 which is sulfhy-
drolysed to the silyl ester 4. Path A is about twice as fast as Path B.
Typically, ca. 30–50% of sulfurization of H-phosphonate 3 proceeded
via the tervalent intermediate.
The stereochemical outcome of Path B was assessed by sulfhy-
drolysis (H2S in dioxane)16 of phosphite 6 (prepared in situ by sily-
lation of active H-phosphonate 3 with N,O-bis(trimethylsilyl)
acetamide (BSA) or TMS-Cl).16 It was found that the distribution
of diastereomers of H-phosphonothioate 4 formed did not depend
on the diastereomeric ratio of the intermediate 6, and that in the
sulfhydrolysis of the phenyl derivative 6a, the 4-RP/SP ratio was
ca. 2:1, while for the pivaloyl derivative 6b, it was ca. 3:2. The
underlying mechanism of the observed stereochemistry is cur-
rently under investigation.
2.2. Synthesis of (SP)-uridine 30-H-phosphonothioate (5-SLOW)
In contrast to the uridine 5-FAST diastereomer, there was no
method available for the stereocontrolled formation of its SLOW
counterpart as a major product. During these studies however,
we found two straightforward routes for asymmetric transforma-
tions favouring the formation of this diastereomer (Fig. 2).
In the first method, uridine H-phosphonothioate 5 (RP/SP ca.
1:1) was reacted with 9-fluorenemethanol using pivaloyl chloride
as a condensing agent (the reagent of choice for stereoselective
condensations of ribonucleoside 30-H-phosphonates18). The reac-
tion was performed in acetonitrile containing 3 equiv of pyridine.
Under these mild conditions the side reactions, which were ob-
served previously,19 were eliminated and the H-phosphonothioate
diester 8b was formed almost quantitatively, as a ca. 9:1 mixture of
||
Since the sulfhydrolysis of phosphites of type 6 (Path B) was
found to be ca. 2 orders of magnitude slower than that of Path A,
It is essential to use HMDTS from a freshly opened bottle. Upon multiple opening
and storage for several weeks (even at ꢀ25 °C), HMDTS undergoes significant
decomposition and must be used in uncontrolled excess to prevent H-phosphono-
dithioate formation.
The use of a polar solvent in sulfhydrolysis was critical for the yield of reaction.
The best results were obtained for acetonitrile, in methylene chloride some by-
products were formed, while in toluene nucleoside trimethylsilyl H-phosphonate and
nucleoside H-dithiophosphonate were the main products of the process.
§
Rf 0.15 for both diastereomers; toluene–MeCN–TEA 3:2:1.
–
The 31P chemical shift observed for 4 (ca. 57 ppm) was located in the region of H-
phosphonothioates with a double P@S bond, while for a O@P–S-TMS isomer, a dP of
ca. 15 ppm would be expected.15

412
M. Sobkowski et al. / Tetrahedron: Asymmetry 21 (2010) 410–419
DMTrO
DMTrO
DMTrO
DMTrO
H2O
B
B
B
B
O
O
H
O
O
O
OH OTBDMSi
O
O P
O
O
P
S
O
P
S
OTBDMSi
OTBDMSi
TMS
OTBDMSi
H
H
O-
1
R
(SP)-4
(SP)-5
SLOW
H2S/TMS-Cl
DPP
PvCl
δP 58.85
(RP)-3
or HMDST
1JPH 665.8 Hz
δP 57.96, 1JPH 591.6 Hz
path A
+
+
DMTrO
DMTrO
DMTrO
DMTrO
B
B
B
B
O
O
O
O
O
H2O
O
O
P
S
O
P
S
OTBDMSi
OTBDMSi
OTBDMSi
OTBDMSi
O P
O
-O
H
H
P
O
O
H
H
TMS
O-
R
2
(RP)-5
FAST
(RP)-4
δP 56.96
(SP)-3
more reactive
δP 56.41, 1JPH 591.6 Hz
1JPH 650.2 Hz
(main intermediate)
(main product)
silylation
H2S
DMTrO
B
O
O
path B
O
P
OTBDMSi
O
R
TMS
(RP)-6 and (SP)-6
6a, δP 126.1 & 130.1
6b, δP 127.4 & 132.2
NHBz
N
NHBz
3a, 6a, R = Ph
3b, 6b, R = Pv
N
N
N
N
AdeBz
=
=
CytBz
=
B = Ura or, if indicated in the text, AdeBz, CytBz, Guaibu
DMTr = 4,4'-dimethoxytrityl
TBDMSi = tertbutyldimethylsilyl
DPP = diphenyl H-phosphonate
Pv = pivaloyl (trimethylacetyl)
HMDST = hexamethyldisilathiane
ibu = isobutyryl
N
O
O
O
N
N
NH
NH
Guaibu
Ura =
N
O
N
NHibu
Figure 1. The assumed mechanism of stereoselective formation of uridine H-phosphonothioate (RP)-5 in the course of sulfhydrolysis of intermediate uridine phenyl H-
phosphonate 3b. The 31P NMR data are given for uridine derivatives in the crude reaction mixtures.
diastereomers [de 83%; dP 72.24 ppm (the main, downfield reso-
nating diastereomer, 8b-DOWN) and 71.17 ppm (8b-UP), respec-
tively]. The ratio of the diastereomers of H-phosphonothioate
diesters 8 did not depend on diastereomeric composition of the
starting H-phosphonothioate monoester 5, confirming that
the reaction was stereoselective rather than stereospecific. Upon
the addition of TEA (20% v/v) a rapid b-elimination10 of the fluore-
nemethyl moiety occurred, producing H-phosphonothioate 5 (yield
96%, 31P NMR) in an unaltered 9:1 ratio [de 82%; dP 55.05 (5-SLOW)
and 53.43 (5-FAST)]. This method is effective and requires only a
brief pre-drying of the starting H-phosphonothioate monoesters
and maintaining anhydrous conditions of the reaction.
the mixed anhydride 7 was a stereoselective reaction affording
5-SLOW as a dominant product. Thus, H-phosphonothioate 5 in
MeCN containing 3 equiv of pyridine was treated with PvCl
(1.2 equiv) to generate the mixed anhydride 7, which was then
hydrolyzed back to monoester 5 by the addition of water to the
reaction mixture. During the reaction no overactivation19 of H-
phosphonothioates occurred while a predominant formation of
the SLOW diastereomer was observed (ca. 65:35, irrespective of a
diastereomeric composition of the starting compound). These re-
sults seemed promising and prompted us to search for reaction
conditions to gain higher stereoselectivity and yield.
To this end, the acylation-hydrolysis procedure was carried out
in the presence of amines of different basicity and nucleophilicity
(Fig. 3). The mixed anhydride 7 was either pre-formed in situ
The second approach to generate uridine derivatives 5-SLOW as
a major product was based on the observation that hydrolysis of

M. Sobkowski et al. / Tetrahedron: Asymmetry 21 (2010) 410–419
413
TEA (if R = 9-fluorenemethyl)
DMTrO
DMTrO
DMTrO
BOTO
DMTrO
B
B
B
B
O
H
O
H
O
O
H
PvCl
OTBDMSi
H2O
O
P
O
P
O
O
P
O
O
P
O
OTBDMSi
OTBDMSi
OTBDMSi
S
S
S
O-
S
O-
Pv
R
R
(RP)-7
(SP)-9
(SP)-5
(SP)-8
main
path
ROH
SLOW
DOWN
(main product)
+
DMTrO
DMTrO
DMTrO
BOTO
DMTrO
B
B
B
B
O
O
O
O
O
O
PvCl
O
P
O
O
P
O
OTBDMSi
OTBDMSi
OTBDMSi
OTBDMSi
H
P
S
H
P
S
-O
S
H
S
O-
O
Pv
R
R
(RP)-5
(SP)-7
(RP)-8
(RP)-9
FAST
UP
B = Ura or, if indicated in the text, AdeBz, CytBz, Guaibu
Pv = pivaloyl
R = Et (8a, 9a), 9-fluorenemethyl (8b, 9b), nucleoside-5'-yl (8c, 9c)
Figure 2. Stereochemistry of reactions of nucleoside H-phosphonothioate derivatives.
and subsequently hydrolysed (‘preactivation’ procedure; Fig. 3,
lines 3 and 4),àà or PvCl was added to a solution of substrate 5 in
aqueous acetonitrile (‘aqueous activation’ procedure; lines 1, 2 and
5). In all cases, with an increasing amount of PvCl added, SLOW:FAST
ratio of 5 tended towards an equilibrium value of ca. 75:25 (lines 1, 3
and 5). Also some desulfurization occurred and its extent was pro-
portional to the quantity of PvCl used (lines 2 and 4).
Preactivation of uridine H-phosphonothioate 5 with PvCl appar-
ently favoured the predominant formation of one diastereomer of
the mixed anhydride 7§§ which upon hydrolysis afforded 5-SLOW,
and thus the amount of the condensing agent required for reaching
the endpoint of asymmetric transformation was lower (ca. 2 equiv,
Fig. 3A, line 3) than in the ‘aqueous activation’ procedures. The pro-
cedure, unfortunately, had two main drawbacks: the relatively high
degree of desulfurization (line 4) and a significant detritylation, par-
ticularly in the reaction mixtures containing pyridine as a base (ca.
30% for 4 equiv of PvCl and 10 equiv of pyridine; 31P NMR and
TLC). For these reasons, this approach was not pursued further.
For the ‘aqueous activation’ procedure, in which the condensing
agent was added to the aqueous MeCN solution of 5, a larger excess
of PvCl was necessary for the completion of the process when
highly enriched 5-FAST (de 84%) was used as the starting material.
When 5-SLOW (de 84%) was used as a substrate, the same end-
point of de 30% was reached after the addition of 5–6 equiv of PvCl
(Fig. 3A, line 5).
This procedure was particularly efficient in the presence of pyr-
idine (line 1), for which desulfurization was also at the lowest level
(line 2). With more basic pyridine derivatives, higher levels of
desulfurization were observed, while less basic derivatives were
less efficient in the asymmetric transformation (data not shown).
The non-nucleophilic tertiary amines appeared to be unsuitable
for this purposes due to the desulfurization of 5 and inefficient for-
mation of 5-SLOW.–– On the other hand, in the presence of N-meth-
yllimidazole (NMI) both desulfurization and interconversion of
diastereomers proceeded reluctantly, indicating that this strong
nucleophilic catalyst favoured the hydrolysis of PvCl over the acyla-
tion of H-phosphonothioate monoester 5.
When diphenyl chlorophosphate (DPCP; a reagent of choice for
standard nucleoside 30-H-phosphonothioate condensations19) was
used as the activating agent under similar reaction conditions
(MeCN/pyridine/H2O), only a marginal conversion of 5-FAST to 5-
SLOW was observed due to the rapid hydrolysis of DPCP. Diethyl-
chlorophosphate (DECP), which is less susceptible to hydrolysis,
was able to promote the asymmetric transformation of H-phos-
phonothioate 5 nearly as efficiently as PvCl and without any desul-
furization. The optimal pyridine content for this reaction was
established to be ca. 20%. These prevented detritylation of 4 and
kept hydrolysis of DECP at a low level.
àà
Generation of the mixed anhydride 7 was carried out in the presence of catalytic
Since this reaction was clean and the by-products formed could
amounts (0.5 equiv) of weakly basic N,N-dimethylaniline (DMA). In the presence of
stronger bases, for example, TEA,, no signals of the mixed anhydride 7 could be
detected in the 31P NMR spectra due to its rapid P-acylation (not observed in the oxo
be easily monitored (31P NMR) and removed during the work-up
series under similar conditions).19
The mixed anhydride 2 was always formed in ca. 3:1 ratio of diastereomers (31
P
The ‘aqueous activation’ procedure in the presence of DMA can be exploited for a
§§
––
NMR), irrespective of the diastereomeric composition of the starting monoester,
presumably due to their rapid equilibration under the reaction conditions.
rapid and efficient desulfurization of H-phosphonothioate monoesters, an alternative
to commonly used oxidative desulfurization procedures (see the Experimental part).

414
M. Sobkowski et al. / Tetrahedron: Asymmetry 21 (2010) 410–419
and silica-gel chromatography, it may be recommended as a sec-
ond best method for the preparation of uridine H-phosphonothio-
The assignment of the configuration of diastereomers of uri-
dine H-phosphonothioate monoester 5 was based on the stereo-
specific b-elimination of the 9-fluorenemethyl group from 9-
fluorenemethyl uridine H-phosphonothioate diester 8b. As it
was already shown (vide supra), diester 8b formed in a SP/RP ra-
tio of ca. 9:1 [Fig. 4, reaction (a)]. Treatment of such mixture
with TEA [reaction (b)] afforded uridine H-phosphonothioate 5
in the same ratio of 9:1, with predominance of the 5-SLOW dia-
stereomer. The full stereospecificity of this reaction was proved
by using as the starting material the diester 8b in other diaste-
reomeric ratios (55:45 and 7:3), which were accurately pre-
served in the reactions, proving that 5-SLOW derived from
(SP)-8b. Since the b-elimination of the 9-fluorenemethyl group
is a stereoretentive process, the configuration of 5-SLOW was as-
signed as (SP), while that of 5-FAST, as (RP). The same procedure
was applied to H-phosphonothioates of other three ribonucleo-
sides of type 5 (B = AdeBz, CytBz, Guaibu) and the same correlation
was found in all cases.
ate monoester 5-SLOW. Thus, when H-phosphonothioate
5
(FAST:SLOW ꢁ 2:1) was dissolved in 1% (v/v) aqueous acetonitrile
containing pyridine (20% v/v), and treated with DECP (6 equiv),
the diastereomeric ratio of H-phosphonothioate monoester 5
recovered changed to FAST:SLOW ꢁ 1:4. A typical work-up of such
reaction mixture and chromatographic separation afforded 5-
SLOW in de 84%. It is noteworthy that the process was very rapid
and did not require prior preparation of monoester 5. However,
the diastereomeric excess of 5-SLOW was lower than that obtained
by the 9-fluorenemethyl method.
Several mechanistic interpretations of the observed stereoselec-
tivity during the ‘aqueous activation’ of uridine H-phosphonothio-
ate 5 are possible. In the simplest scenario, the major diastereomer
of 5 (5-SLOW) was formed from the major diastereomer of the
mixed anhydride 7 and the reaction followed a dynamic thermody-
namic resolution (DYTR)20,21 type of asymmetric induction. The
reaction required several molar equivalents of PvCl to reach the
endpoint, apparently due to the competitive reactions of the con-
densing agent with water or pivalic acid (formed during hydrolysis
of PvCl and/or 7).22 Another possible mechanism would involve the
formation, P-epimerisation, and hydrolysis of the mixed anhydride
7 (cf. Fig. 1), according to a cyclic de-racemization type of asymmet-
ric induction.21 Some other mechanisms seem to be less probable
(e.g., the diverse reactivity of diastereomers of H-phosphonothio-
ate 5 towards PvCl). While the available data cannot delineate
which mechanism is actually operating, similar ratios of diastereo-
mers of the intermediate mixed anhydride 7 (74:26) and the
monoester 5 formed (77:23) point to the DYTR mechanism. How-
ever, when a more sterically demanding dimethoxytrityl group
was present in the 20-O position instead of the TBDMSi group, a
SLOW/FAST ratio of ca. 9:1 could be achieved via the same proce-
dure.|||| This result cannot be explained in terms of the DYTR and
points to the cyclic de-racemization process in this case.
The above methods can also be applied to achieve a comparable
diastereomeric enrichment of other fully protected ribonucleoside
H-phosphonothioate monoesters 5 (B = AdeBz, CytBz, Guaibu; Figs. 1
and 2). The ratios of diastereomers obtained and their chromato-
graphic and 31P NMR data are collected in Table 1. The lowest ste-
reoselectivity in the asymmetric transformations was observed for
the cytidine derivative, similarly as it was found in the oxo ser-
ies.1,18,23 Diastereomers of all the H-phosphonothioate monoesters
of type 5 had comparable differences in mobilities on silica-gel (dRf
ca. 0.12) and could be separated chromatographically.
Since the above correlation analysis for H-phosphonothioate
monoester 5 relied on the correct assignment of configuration of
diastereomers of 9-fluorenemethyl uridine H-phosphonothioate
diester 8b, we verified it in an independent procedure (Fig. 4,
non-shaded area). To this end, the diester 8b as previously ob-
tained in the ratio of 8-DOWN:8-UP ꢁ 9:1 [reaction (a)] was oxi-
dized in situ stereospecifically with 3H-2,1-benzoxathiol-3-one
1-oxide (BOTO)4 yielding two diastereomers of 9-fluorenemethyl
uridine 30-phosphorothioate 9b, also in a 9:1 ratio [reaction (c)].
In a separate experiment, the same phosphorothioate diester 9b
was prepared by the sulfurization of 9-fluorenemethyl uridine H-
phosphonate 10b [reactions (d) and (e)] and its configuration was
established as described previously.1,2,24 The (SP)-product obtained
in this way had identical 31P chemical shits and coupling constants
as the major diastereomer of 9b in the analyzed reaction mixture,
while the opposite (RP)-diastereomer, was the minor one. Since oxi-
dation with BOTO proceeds with retention of configuration,4 diaste-
reomer 8b-DOWN could be assigned as (SP), and 8b-UP, as (RP),
which is in agreement with the initial correlation. Since the same
procedure repeated for ethyl nucleoside diesters 8a–10a (B = AdeBz
,
CytBz, Guaibu, Ura) gave the same results (e.g., for B = Ura, see Fig. 5),
the formation of the (SP)-diastereomer as the main product in the
condensations of nucleoside 30-H-phosphonothioates may be con-
sidered as a general rule.
3. Conclusion
The H-phosphonothioate esters of protected thymidine and uri-
dine differ substantially in their ability to be isolated as individual
diastereomers: in the deoxy series only alkyl nucleoside diesters
are separable by silica-gel column chromatography, while in the
ribo series, only the monoesters. In order to enhance the yields
of diastereomers of protected nucleoside 30-H-phosphonothioate
with a desired configuration at the phosphorus atom, four stereo-
selective synthetic procedures were developed.
The ‘FAST’ diastereomers, showing higher chromatographic
mobility on silica-gel, were synthesised via sulfhydrolysis of acti-
vated derivatives of nucleoside 30-H-phosphonates with diastereo-
meric excess (de) 35–50%. The ‘SLOW’ diastereomers were
prepared using (i) stereoselective formation of 9-fluorenemethyl
nucleoside 30-H-phosphonothioate (de 60–80%) followed by stereo-
specific removal of the 9-fluorenemethyl group, or (ii) asymmetric
transformation of nucleoside 30-H-phosphonothioate in the course
of an activation—hydrolysis one pot process (de 35–50%). Using ste-
reochemical correlation analysis of 31P NMR spectra, the absolute
configuration of ‘FAST’ and ‘SLOW’ diastereomers of protected ribo-
nucleoside 30-H-phosphonothioates and of ‘DOWN’ and ‘UP’
2.3. A tentative assignment of the configuration of the
diastereomers of nucleoside 30-H-phosphonothioate 5
In the first step of the correlation analysis, the configuration of
the alkyl uridine H-phosphonothioate diesters of type 8 was as-
signed, and this subsequently used for the determination of the
configuration of uridine H-phosphonothioate monoester of type
5. The discussion below refers to uridine derivatives but it also
holds for the other ribonucleoside H-phosphonothioates.
The diastereoselectivity of formation of diribonucleoside H-phos-
phonothioates 8c (Fig. 2) is known and the (SP)-diastereomer was
identified as the main product.4 Thus, the same (SP)-configuration
was tentatively assigned to the main diastereomers of alkyl uridine
30-H-phosphonothioate diesters (8-DOWN) (note; a similar correla-
tion was found for alkyl ribonucleoside 30-H-phosphonates2,3).
||||
Stereoelectronic effects that possibly govern the stereoselective transformations of
nucleoside 30-H-phosphonates and their analogues, are a subject of separate studies.

M. Sobkowski et al. / Tetrahedron: Asymmetry 21 (2010) 410–419
415
Figure 3. Charts A–D: Diastereomeric composition of uridine 30-H-phosphonothioate monoester 5 during acylation–hydrolysis procedure in the presence of various amines
(in parentheses, pKa values in MeCN). The legend in chart A applies to all charts: Lines 1 and 5, de of 5-FAST during a stepwise addition of PvCl to a solution of 5 (5-FAST, de
84% or 5-SLOW, de 84%, respectively) in 90:6:4 (v/v/v) MeCN–amine–H2O; Line 2, % of desulfurization of 5 during this process; Line 3, de of 5-FAST during hydrolysis of the
mixed anhydride 7 formed in situ using a given amount of PvCl; Line 4, % of desulfurization of 5 during this process. Chart E: The same process promoted by diethyl
chlorophosphate (DECP) in the presence of pyridine.
Table 1
31P NMR data, Rf values, and ratios of diastereomers for nucleoside H-phosphonothioates 5 obtained under various conditions
Nucleosidea
Diastereomer
dP in DCM/ppm (1JPH
,
3JPH/Hz)
Rfb
Ratio of diastereomers of 5 (FAST/SLOW)c
4b + HMDST
1. 5 + FMOH + PvCl
5 + DECP in MeCN–H2O
2. +TEA
A
C
FAST
SLOW
FAST
SLOW
FAST
SLOW
FAST
54.25 (576.9; 12.8)
56.16 (587.0; 12.8)
54.27 (576.0; 12.8)d
54.35 (577.9; 12.8)d
54.91 (580.6; 15.6)
55.65 (588.4; 14.2)
53.80 (576.0; 12.8)
55.62 (582.4; 13.7)
0.54
0.37
0.59
0.44
0.36
0.25
0.42
0.32
3:1
2:1
3:1
3:1
1:9
1:4
1:7
1:9
1:3
1:2
1:3
1:4
G
U
SLOW
For reactions details, see Experimental. FMOH = (9H-fluoren-9-yl)methanol. Pv = pivaloyl (trimethacetyl).
a
Fully protected nucleosides (50-O-DMTr, 20-O-TBDMSi; B = AdeBz, CytBz, Guaibu, Ura).
b
Toluene–MeCN–TEA 3:2:1.
c
Integration of 31P NMR signals.
d
The signals overlapped in DCM and the ratio of diastereomers was established in toluene (FAST: dP 55.18, 1JPH 578.3 Hz, 3JPH 13.3 Hz; SLOW: dP 54.88, 1JPH 577.4 Hz, 3JPH
12.4 Hz).

416
M. Sobkowski et al. / Tetrahedron: Asymmetry 21 (2010) 410–419
NuO
5 + FMOH + PvCl
H
P
O
2
O
+ FMOH + PvCl
(a)
(d)
δP 2.15
Nu
O
P
Nu
O
P
Nu
O
P
TEA
BOTO
Nu
O
S8
(retention)
(retention)
(retention)
-
S
H
S
O
H
O
S
P
H
(b)
(c)
(e)
O-
O
O
O
FM
FM
FM
(Sp)-5
SLOW / DOWN
δP 55.05
(Sp)-8b
DOWN
(Sp)-9b
DOWN
δP 58.50
(Rp)-10b
δP 7.41
(minor product)
δP 72.24
(main product)
B = Ura or, if indicated in the text, AdeBz, CytBz, Guaibu
FMOH = 9-fluorenemethanol
PvCl = pivaloyl chloride
BOTO = 3H-2,1-benzoxathiol-3-one 1-oxide
Figure 4. Stereochemical correlation analysis for assignment of configuration of nucleoside H-phosphonothioates 5 and 8b. The chemical shifts given are of uridine
derivatives in DCM.
31
Figure 5. P NMR spectra of the reaction mixtures during esterification of uridine 30-H-phosphonothioate 5 with EtOH: (i) starting monoester 5 in DCM; (ii) reaction of 5
with PvCl; (iii) reaction of 5 with EtOH and PvCl; (iv) mixture (iii) after addition of BOTO; (v) phosphorothioate 9a obtained by sulfurization of ethyl uridine H-phosphonate;
(vi) mixtures (iv) and (v) combined.
diastereomers of alkyl ribonucleoside 30-H-phosphonothioates was
tentatively assigned as RP/SP, respectively.
eter. Commercial (Sigma–Aldrich, Alfa Aesar, Merck, POCh-Po-
land) reagents and were used as purchased unless otherwise
noted. For column chromatography, Silica Gel 60 (70–230 mesh)
was used. Anhydrous solvents (stored over molecular sieves 4 Å)
contained below 20 ppm of water (Karl Fischer coulometric titra-
tion, Metrohm 684 KF). Ethanol was distilled over magnesium.
Pivaloyl chloride was distilled and used within one month. 1 M
triethylammonium bicarbonate (TEAB) buffer was prepared by
4. Experimental
31P NMR spectra were recorded at 121 MHz on a Varian Unity
BB VT spectrometer. 1H and 13C NMR were recorded at 400 and
100 MHz, respectively, on a Bruker Avance II 400 MHz spectrom-

M. Sobkowski et al. / Tetrahedron: Asymmetry 21 (2010) 410–419
417
bubbling CO2 through aqueous TEA at 0 °C until pH 7.0 was
reached. All reactions were monitored using 31P NMR
spectroscopy.
H10); 6.78–7.45 (13.5H, m, Tr + 1/2 P–H); 7.49 (2H, t, J ꢃ7.6 Hz, m-
Bz); 7.57 (1H, t, J ꢃ7.6 Hz, p-Bz); 8.00 (2H, d, J 7.6 Hz, o-Bz); 8.26
(1H, s, H2); 8.69 (1H, s, H8); 8.75 (0.5H, 1/2 d, 1/2 P–H); 9.15 (1H,
br s, NH); 12.17 (1H, br s, NH); 13C NMR (CDCl3; DMTr resonances
not listed) dC ꢀ5.12 (Si–CH3); ꢀ4.52 (Si–CH3); 8.47 (NCH2CH3);
17.94 (Si–C(CH3)3); 25.60 (Si–C(CH3)3); 45.44 (NCH2CH3); 55.12
4.1. Typical procedures for the preparation of diastereomers
of 20-O-(t-butyldimethylsilyl)-50-O-(dimethoxytrityl)nucleoside
-30-yl H-phosphonothioate 5
(OCH3); 62.87 (C50); 73.57 (d, JP,C3 5.5 Hz, C30); 75.02 (d, JP,C2
0
0
4.4 Hz, C20); 82.98 (d, JP,C4 4.4 Hz, C40); 86.54, 88.59 (C10, C5);
141.65 (C2); 144.46 (C6); 158.40 (C4); 164.50 (C(O)Ph); 31P NMR
(DCM) dP 54.25 (dd, 1JH,P 576.9 Hz, 3JH,P 12.8 Hz).
0
4.1.1. Isomer(RP)-5 (FAST; B = Ura). Method A
The compound was prepared according to slightly modified
published procedures.13,14 20-O-(t-Butyldimethylsilyl)-50-O-(dime-
thoxytrityl)uridine 1 (1.32 g, 2 mmol) was rendered anhydrous
by evaporation of added pyridine (20 mL) under reduced pressure.
The residue was dissolved in 20 mL of the same solvent and diphe-
nyl phosphite (1.26 mL, 4 mmol) was added, while stirring vigor-
ously. After ca. 1 h, TLC showed complete phosphitylation and
HMDST (1.06 mL, 5 mmol) was added slowly (ca. 15 s) with stir-
ring. After ca. 1 h the sulfhydrolysis of diester 3a was complete
4.1.2.2. (RP)-5. (FAST; B = CytBz).
1H NMR (CDCl3) dH 0.20 (3H,
s, Si–CH3); 0.25 (3H, s, Si–CH3); 0.91 (9H, s, Si–tBu); 1.23 (9H, t, J
0
00
7.4 Hz, NCH2CH3); 3.00 (6H, q, NCH2CH3); 3.51 (1H, dd, J5 ,5 11.2,
J4 ,5 2.5 Hz, H50); 3.66 (1H, br d, H500); 3.79 (6H, s, 2 ꢂ OCH3); 4.41
0
0
1
(1H, br t, J3 ,4 8.5 Hz, H40); 4.45 (1H, br d, J2 ,3 4.0 Hz, H20); 5.05 ( H,
0
0
0
0
ddd, JP,3 13.7 Hz H30); 5.88 (1H, br s, H10); 6.84–7.42 (14.5H, m,
0
Tr + H5 + 1/2 P–H); 7.46 (2H, t, J ꢃ7.8 Hz, m-Bz); 7.58 (1H, t, J
ꢃ7.5 Hz, p-Bz); 7.87 (2H, d, J 7.5 Hz, o-Bz); 8.59 (1H, d, J5,6 7.5 Hz,
H6); 8.62 (0.5H, 1/2 d, 1/2 P–H); 12.03 (1H, br s, NH); 13C NMR
(CDCl3; DMTr resonances not listed) dC ꢀ4.74 (Si–CH3); ꢀ4.67 (Si–
CH3); 8.42 (NCH2CH3); 18.04 (Si–C(CH3)3); 25.78 (Si–C(CH3)3);
(
31P NMR) and the reaction mixture was left overnight with an-
other portion of HMDST (0.85 mL, 4 mmol). The reaction was
quenched with water, after which the solvent and volatile products
were removed in a rotary evaporator and the residue was worked
up (3 ꢂ DCM/1 M TEAB buffer) and concentrated.
45.40 (NCH2CH3); 55.09 (OCH3); 60.58 (C50); 73.04 (d, JP,C3 4.0 Hz,
0
C30); 75.56 (d, JP,C2 4.0 Hz, C20); 80.59 (d, JP,C4 6.7 Hz, C40); 91.60
0
0
4.1.2. Isomer (RP)-5 (FAST; B = Ura). Method B
(C10); 96.2 (br, C5); 143.86 (C2); 144.9 (br, C6); 158.50 (C4);
This method is based on a published procedure.11 20-O-(t-Butyl-
161.94 (C(O)Ph); 31P NMR (DCM) dP 54.43 (dd, JH,P 576.0 Hz, JH,P
1
3
1
3
dimethylsilyl)-50-O-(dimethoxytrityl)uridine H-phosphonate
2
12.8 Hz); 31P NMR (toluene) dP 55.18 (dd, JH,P 578.3 Hz, JH,P
(1.65 g, 2 mmol) was rendered anhydrous by the evaporation of
added toluene (20 mL) under reduced pressure. The residue was
dissolved in 20 mL of acetonitrile and 1 mL of pyridine (12 mmol),
and PvCl (0.76 mL, 6 mmol) was added. After 5 min HMDST
(1.27 mL, 6 mmol) was added rapidly, while stirring vigorously,
followed by 2,6-lutidinium hydrochloride (0.86 g, 6 mmol) and
the vigorous stirring of the suspension was continued until sulfhy-
drolysis was complete (ca. 30 min, 31P NMR). The reaction was
quenched with water, after which the solvent and volatile products
were removed in a rotary evaporator and the residue was worked
up (DCM/1 M TEAB buffer) and concentrated.
13.3 Hz).
4.1.2.3. (RP)-5. (FAST; B = Guaibu).
(3H, s, Si–CH3); 0.11 (3H, s, Si–CH3); 0.80 (9H, s, Si–tBu); 0.98 and
1.07 (2 ꢂ 3H, 2 ꢂ d, J 6.6 Hz, 2 ꢂ CHCH3) 1.25 (9H, t, J 7.5 Hz,
NCH2CH3); 2.31 (1H, septet, J 7.7 Hz CHCH3), 3.04 (6H, q, NCH2CH3);
1H NMR (CDCl3) dH ꢀ0.03
0
00
0
00
0
0
3.12 (1H, dd, J5 ,5 11.2 Hz, J4 ,5 2.5 Hz, H5 ); 3.49 (1H, br d, H5 ); 3.71
and 3.72 (2 ꢂ 3H, 2 ꢂ s, 2 ꢂ OCH3); 4.29 (1H, br d, J3 ,4 5.5 Hz, H40);
0
0
5.06 (1H, dd, J1 ,2 3.8 Hz, J3 ,2 4.6 Hz, H20); 5.56 (1H, ddd, JP,3
0
0
0
0
0
15.3 Hz, H30); 5.79 (1H, d, H10); 6.70–7.38 (13.5H, m, Ar + 1/2 P–
H); 7.79 (1H, d, H8); 8.63 (0.5H, 1/2 d, 1/2 P–H); 11.67 (1H, br s,
NH–ibu); 11.99 (1H, br s, NH); 13C NMR (CDCl3; DMTr resonances
not listed) dC ꢀ5.09 and ꢀ4.31 (2 ꢂ Si–CH3); 8.51 (NCH2CH3);
17.95 (Si–C(CH3)3); 18.63 and 18.73 (C(O)CH(CH3)2); 25.59 (Si–
C(CH3)3); 35.89 (C(O)CH(CH3)2); 45.59 (NCH2CH3); 55.09 (OCH3);
In both methods crude 5 contained ca. 3:1 ratio of RP:SP diaste-
reomers (31P NMR). A silica-gel column chromatography using 10–
30% gradient of MeCN in toluene containing 5% of TEA followed by
lyophilization from benzene yielded (Method B) 760 mg of (RP)-5
(de 84%, yield 57%), 440 mg of a mixture of diastereomers (ca.
1:1), and 50 mg of (SP)-5 (de 80%). Total yield 79% (Method A) or
86% (Method B).
61.97 (C50); 72.40 (d, JP,C3 3.6 Hz, C30); 74.60 (d, JP,C2 4.0 Hz, C20);
0
0
81.83 (d, JP,C4 4.3 Hz, C40); 89.51 (C10); 121.73 (C5); 147.31 and
0
147.96 (C2 and C4); 155.66 (C6); 179.09 [NHC(O)]; 31P NMR (DCM)
1H NMR (CDCl3) dH 0.17 (3H, s, Si–CH3); 0.18 (3H, s, Si–CH3);
0.88 (9H, s, Si–tBu); 1.27 (9H, t, J 7.5 Hz, NCH2CH3); 3.04 (6H, q,
dP 54.58 (dd, 1JH,P 577.9 Hz, 3JH,P 15.6 Hz).
NCH2CH3); 3.49 (1H, dd, J5 ,5 11.1 Hz, J4 ,5 2.3 Hz, H50); 3.55 (1H,
4.1.3. Isomer (SP)-5 (SLOW; B = Ura). Method A
0
00
0
0
dd, J4 ,5 2.1 Hz, H500); 3.75 (6H, s, 2 ꢂ OCH3); 4.33 (1H, br d, J3 ,4
A mixture of uridine H-phosphonothioate 5 (750 mg, 0.89 mmol)
and (9H-fluoren-9-yl)methanol (FMOH) (345 mg, 1.76 mmol) was
rendered anhydrous by evaporation of added toluene (10 mL) under
reduced pressure. The residue was dissolved in 10 mL of DCM con-
0
00
0
0
1
6.5 Hz, H40); 4.42 (1H, dd, J1 ,2 2.9 Hz, J3 ,2 4.5 Hz, H20); 5.05 ( H,
0
0
0
0
ddd, JP,3 13.7 Hz, H30); 5.16 (1H, d, JH6,H5 8.2 Hz, H5); 5.90 (1H, d,
0
H10); 6.80–7.42 (13.5H, m, Ar + 1/2 P–H); 8.03 (1H, d, H6); 8.68
(0.5H, 1/2 d, 1/2 P–H); 9.48 (1H, br s, NH); 12.19 (1H, br s, NH);
13C NMR (CDCl3; DMTr resonances not listed) dC ꢀ5.35 (Si–CH3);
ꢀ4.91 (Si–CH3); 8.07 (NCH2CH3); 17.65 (Si–C(CH3)3); 25.28 (Si–
C(CH3)3); 45.01 (NCH2CH3); 54.71 (OCH3); 60.97 (C50); 71.46 (d,
taining pyridine (240 lL, 3.0 mmol) and PvCl (166 lL, 1.32 mmol)
was added. After ca. 5 min 2 mL of TEA was added and the reaction
was monitored with 31P NMR or TLC until the signal of diester 8b dis-
appeared (ca 45 min). The crude reaction mixture contained ca. 10%
of (SP)-5 and 90% of (RP)-5 (31P NMR). Aftertypical work-up and rapid
column chromatography using 20–30% gradient of acetonitrile in
toluene containing 5% of TEA, followed by lyophilization from ben-
zene, 40 mg of mixture of diastereomers and 660 mg of (RP)-5 (de
96%) were obtained. Total recovery of 5, 93%.
JP,C3 4.7 Hz, C30); 75.18 (d, JP,C2 3.5 Hz, C20); 80.97 (d, JP,C4 5.2 Hz,
0
0
0
C40); 89.20 (C10); 101.33 (C5); 139.78 (C6); 149.82 (C2); 162.88
1
3
(C4); 31P NMR (DCM) dP 53.87 (dd, JH,P 576.0 Hz, JH,P 13.7 Hz).
4.1.2.1. (RP)-5. (FAST; B = AdeBz).
(3H, s, Si–CH3); 0.08 (3H, s, Si–CH3); 0.78 (9H, s, Si–tBu); 1.27 (9H,
1H NMR (CDCl3) dH ꢀ0.08
0
00
t, J 7.2 Hz, NCH2CH3); 3.03 (6H, q, NCH2CH3); 3.43 (1H, dd, J5 ,5
4.1.4. Isomer (SP)-5 (SLOW; B = Ura). Method B
10.8, J4 ,5 3.6 Hz, H50); 3.54 (1H, dd, J4 ,5 2.6 Hz, H500); 3.74 (6H, s,
Crude5(842 mg, 1.0 mmol,RP:SP = 55:45)wasdissolvedin10 mL
of MeCN containing pyridine (20%, v/v) and water (1.2% v/v), and
to this diethylchlorophosphate (DECP, 0.96 mL, 6 mmol) was
0
0
0
00
2 ꢂ OCH3); 4.50 (1H, br t, J3 ,4 3.2 Hz, H40); 5.02 (1H, br t, J1 ,2
0
0
0
0
1
5.2 Hz, J2 ,3 4.4 Hz, H20); 5.12 ( H, dt, JP,3 12.8 Hz H30); 6.19 (1H, d,
0
0
0

418
M. Sobkowski et al. / Tetrahedron: Asymmetry 21 (2010) 410–419
added slowly (ca. 15 s) with vigorous stirring. This yielded 5 as a dia-
stereomeric mixture containing ca. 20% of (RP)-5 and 80% of (SP)-5
(
ꢀ4.61 (2 ꢂ Si–CH3); 8.55 (NCH2CH3); 17.97 (Si–C(CH3)3); 18.90
and 18.97 (C(O)CH(CH3)2); 25.60 (Si–C(CH3)3); 36.16
(C(O)CH(CH3)2); 45.68 (NCH2CH3); 55.18 (OCH3); 61.76 (C50);
31P NMR). After work-up (1 ꢂ DCM/TEAB) the product was isolated
as described for the (RP)-isomer, yielding 90 mg of (RP)-5 (de 79%),
280 mg of mixture of diastereomers, and 290 mg of (SP)-5 (de 84%)
after lyophilization from benzene. Total yield 78%.
72.54 (d, JP,C3 4.8 Hz, C30); 74.93 (d, JP,C2 3.8 Hz, C20); 83.69 (d, JP,C4
0
0
0
3.7 Hz, C40); 90.72 (C10); 122.04 (C5); 147.41 and 148.12 (C2 and
C4); 155.42 (C6); 179.09 [NHC(O)]; 31P NMR (DCM) dP 54.93 (dd,
1JH,P 590.7 Hz, 3JH,P 14,66 Hz).
1H NMR (CDCl3) dH 0.08 (3H, s, Si–CH3); 0.14 (3H, s, Si–CH3);
0.87 (9H, s, Si–tBu); 1.30 (9H, t, J 7.5 Hz, NCH2CH3); 3.05 (6H, q,
NCH2CH3); 3.37 (1H, dd, J5 ,5 10.8, J4 ,5 2.0 Hz, H50); 3.65 (1H, dd,
4.2. Typical procedures for desulfurization of H-
phosphonothioate 5
0
00
0
0
J4 ,5 1.9 Hz, H500); 3.77 (6H, s, 2 ꢂ OCH3); 4.40 (1H, br d, J3 ,4
0
00
0
0
1
1
2.2 Hz, H40); 4.48 ( H, dd, J1 ,2 5.8 Hz, J3 ,2 4.7 Hz, H20); 5.16 ( H,
0
0
0
0
ddd, JP,3 13.7 Hz, H30); 5.21 (1H, d, JH6,H5 8.2 Hz, H5); 6.06 (1H, d,
Pivaloyl chloride (250
dine H-phosphonothioate 5 (421 mg, 0.5 mmol) in MeCN (10 mL)
containing dimethylaniline (DMA, 600 L) and water (100 L,
11 equiv). After 5 min DMA (500 L), water (40 L), and PvCl
(250 L) were added successively two times in 5 min intervals. The
lL, 4 equiv) was added to a solution of uri-
0
H10); 6.81–7.42 (13.5H, m, Ar + 1/2 P–H); 7.82 (1H, d, H6); 8.89
(0.5H, 1/2 d, 1/2 P–H); 9.25 (1H, br s, NH); 12.09 (1H, br s, NH);
13C NMR (CDCl3; DMTr resonances not listed) dC ꢀ4.97 (Si–CH3);
ꢀ4.70 (Si–CH3); 8.51 (NCH2CH3); 17.93 (Si–C(CH3)3); 25.58 (Si–
C(CH3)3); 45.54 (NCH2CH3); 55.17 (OCH3); 62.70 (C50); 74.66 (d,
l
l
l
l
l
31P NMR spectrum showed 65%, 88%, and >99% of desulfurization
of 5 after addition of each portion of PvCl. Triethylamine (1 mL)
wasaddedtopreventdetritylationandthesolventswereevaporated
under reduced pressure. The residue was dissolved in DCM and ex-
tracted with TEAB buffer. Silica gel chromatography (initial isocratic
3% MeOH in DCM, followed by 3?5% gradient of MeOH in DCM
containing 2% TEA) and lyophilization from benzene afforded 20-O-
(t-butyldimethylsilyl)-50-O-(dimethoxytrityl)uridin-30-yl H-phos-
phonate as a white amorphous solid (345 mg, 83%). Analytical data
were consistent with the literature data.25,26
JP,C3 5.0 Hz, C30); 75.12 (d, JP,C2 3.7 Hz, C20); 83.99 (d, JP,C4 4.3 Hz,
0
0
0
C40); 87.83 (C10); 102.22 (C5); 140.24 (C6); 150.49 (C2); 163.08
1
3
(C4); 31P NMR (DCM) dP 55.66 (dd, JH,P 582.4 Hz, JH,P 13.7 Hz).
4.1.4.1. (SP)-5. (SLOW; B = AdeBz).
(3H, s, Si–CH3); 0.05 (3H, s, Si–CH3); 0.75 (9H, s, Si–tBu); 1.32 (9H,
1H NMR (CDCl3) dH ꢀ0.26
0
00
t, J 7.3 Hz, NCH2CH3); 3.09 (6H, q, NCH2CH3); 3.47 (1H, dd, J5 ,5
10.3, J4 ,5 2.6 Hz, H50); 3.54 (1H, dd, J4 ,5 3.3 Hz, H500); 3.77 (6H, s,
0
0
0
00
2 ꢂ OCH3); 4.54 (1H, br s, H40); 5.05 (1H, br t, J2 ,3 4.7 Hz, H20); 5.30
0
0
(1H, br dd, J3 ,4 1.1 Hz, JP,3 13.8 Hz H30); 6.24 (1H, d, J1 ,2 6.8 Hz,
H10); 6.81–7.48 (13.5H, m, Tr + 1/2 P–H); 7.51 (2H, t, J ꢃ7.4 Hz, m-
Bz); 7.59 (1H, t, J ꢃ7.4 Hz, p-Bz); 8.03 (2H, d, J 7.4 Hz, o-Bz); 8.22
(1H, s, H2); 8.71 (1H, s, H8); 8.93 (0.5H, 1/2 d, 1/2 P–H); 9.20 (1H,
br s, NH); 12.06 (1H, br s, NH); 13C NMR (CDCl3; DMTr resonances
not listed) dC ꢀ5.37 (Si–CH3); ꢀ4.71 (Si–CH3); 8.51 (NCH2CH3);
17.75 (Si–C(CH3)3); 25.47 (Si–C(CH3)3); 45.49 (NCH2CH3); 55.14
0
0
0
0
0
4.3. Determination of the configuration of diastereomers of H-
phosphonothioate 8
3H-2,1-Benzoxathiol-3-one 1-oxide (BOTO) was prepared in situ
by treating 3H-l,2-benzodithiol-3-one 1,l-dioxide (Beaucage re-
agent, 0.15 mmol/100 lL of MeCN) with TEA (10 l
L).27 The precipi-
(OCH3); 63.26 (C50); 75.40 (m, C20, C30); 84.96 (d, JP,C4 4.0 Hz, C40);
tated sulfur was centrifuged and the supernatant was added to an
MeCN solution of alkyl nucleoside H-phosphonothioates 8a or 8b
(0.05 mmol/400 lL) obtained in situ as described above (Method
0
86.74, 87.51 (C10, C5); 141.50 (C2); 144.56 (C6); 158.43 (C4);
1
3
164.52 (C(O)Ph); 31P NMR (DCM) dP 56.79 (dd, JH,P 585.2 Hz, JH,P
13.7 Hz).
A). The diastereomeric composition of the resulting diesters 9a or
9b was analyzed by 31P NMR spectroscopy and their configurations
wereidentifiedby spikingthereactionmixtureswitha solutionof9a
or 9b, respectively, prepared as a SP/RP 1:4 diastereomeric mixture as
described previously.2
4.1.4.2. (SP)-5. (SLOW; B = CytBz).
s, Si–CH3); 0.16 (3H, s, Si–CH3); 0.90 (9H, s, Si–tBu); 1.29 (9H, t, J
1H NMR(CDCl3)dH 0.15(3H,
0
00
7.3 Hz, NCH2CH3); 3.04 (6H, q, NCH2CH3); 3.54 (1H, br d, J5 ,5 11.0,
H50); 3.65 (1H, dd, J4 ,5 2.7 Hz, H500); 3.80 (6H, s, 2 ꢂ OCH3); 4.47
0
00
1
(2H, m, H20, H40); 5.05 ( H, dt, JP,3 12.5 Hz, J2 ,3 ꢃ5.2 Hz, J4 ,3
Acknowledgements
0
0
0
0
0
ꢃ5.2 Hz H30); 6.05 (1H, d, J1 ,2 3.4 Hz, H10); 6.84–7.42 (14.5H, m,
Tr + H5 + 1/2 P–H); 7.49 (2H, t, J ꢃ7.8 Hz, m-Bz); 7.58 (1H, t, J
ꢃ7.9 Hz, p-Bz); 7.90 (2H, d, J 6.3 Hz, o-Bz); 8.36 (1H, d, J5,6 6.5 Hz,
H6); 8.82 (0.5H, 1/2 d, 1/2 P–H); 12.13 (1H, br s, NH); 13C NMR
(CDCl3; DMTr resonances not listed) dC ꢀ4.91 (Si–CH3); ꢀ4.73 (Si–
CH3); 8.51 (NCH2CH3); 18.03 (Si–C(CH3)3); 25.78 (Si–C(CH3)3);
0
0
The financial support from Polish Ministry of Science and High-
er Education is gratefully acknowledged.
References
45.50 (NCH2CH3); 55.19 (OCH3); 61.85 (C50); 73.03 (br d, JP,C3
1. Almer, H.; Stawinski, J.; Stromberg, R. Nucleic Acids Res. 1996, 24, 3811–3820.
2. Sobkowski, M.; Jankowska, J.; Stawinski, J.; Kraszewski, A. Nucleosides
Nucleotides Nucleic Acids 2005, 24, 1469–1484.
3. Sobkowski, M.; Kraszewski, A.; Stawinski, J. Tetrahedron: Asymmetry 2007, 18,
2336–2348.
4. Stawinski, J.; Thelin, M. Tetrahedron Lett. 1992, 33, 3189–3192.
5. Kers, I.; Stawinski, J.; Kraszewski, A. Tetrahedron Lett. 1998, 39, 1219–1222.
6. Kers, I.; Cieslak, J.; Jankowska, J.; Kraszewski, A.; Stawinski, J. Nucleosides
Nucleotides 1999, 18, 1245–1246.
7. Misiura, K.; Szymanowicz, D.; Olesiak, M.; Stec, W. J. Tetrahedron Lett. 2004, 45,
4301–4305.
8. Jankowska, J.; Cieslak, J.; Kraszewski, A.; Stawinski, J. Tetrahedron Lett. 1997, 38,
2007–2010.
9. Seeberger, P. H.; Jorgensen, P. N.; Bankaitis-Davis, D.; Beaton, G.; Caruthers, M.
H. J. Am. Chem. Soc. 1996, 118, 9562–9566.
10. Hiresova, R.; Putilova, L.; Kullberg, M.; Stawinski, J.. In Collection Symposium
Series; Hocek, M., Ed.; Institute of Organic Chemistry and Biochemistry,
Academy of Sciences of the Czech Republic: Prague, 2005; Vol. 7, pp 201–205.
11. Zain, R.; Stromberg, R.; Stawinski, J. J. Org. Chem. 1995, 60, 8241–8244.
12. Stawinski, J.; Zain, R. Nucleosides Nucleotides 1995, 14, 711–714.
13. Kers, I.; Kers, A.; Stawinski, J.; Kraszewski, A. Tetrahedron Lett. 1999, 40, 3945–
3948.
0
4.4 Hz, C30); 75.67 (br d, JP,C2 2.6 Hz, C20); 82.61 (br d, JP,C4 3.7 Hz,
0
0
C4’); 90.22 (C10); 96.4 (br, C5); 144.18 (C2); 144.9 (br, C6); 158.57
(C4); 161.96 (br C(O)Ph); 31P NMR (DCM) dP 54.43 (dd, JH,P
1
576.0 Hz, JH,P 12.8 Hz); 31P NMR (toluene) dP 54.88 (dd, JH,P
3
1
577.4 Hz, 3JH,P 12.4 Hz).
4.1.4.3. (SP)-5. (SLOW; B = Guaibu).
1H NMR (CDCl3) dH ꢀ0.16
(3H, s, Si–CH3); 0.06 (3H, s, Si–CH3); 0.77 (9H, s, Si–tBu); 0.84 and
0.97 (2 ꢂ 3H, 2 ꢂ d, J 6.8 Hz, 2 ꢂ CHCH3) 1.27 (9H, t, J 7.4 Hz,
NCH2CH3); 1.95 (1H, septet, J 6.6 Hz CHCH3), 3.05 (6H, q, NCH2CH3);
0
0
00
0
0
0
0
3.31 (1H, dd, J5 ,5 10.2 Hz, J4 ,5 3.5 Hz, H5 ); 3.45 (1H, dd, J4 ,5 1.9 Hz,
1
H500); 3.76 (6H, s, OCH3); 4.36 ( H, br d, J3 ,4 1.8 Hz, H4 ); 5.04 (1H, t,
0
0
0
J1 ,2 4.0 Hz, J3 ,2 7.1 Hz, H20); 5.50 (1H, ddd, JP,3 13.6 Hz, H30); 5.81
(1H, d, H10); 6.70–7.50 (13.5H, m, Ar + 1/2 P–H); 7.78 (1H, d, H8);
8.63 (0.5H, 1/2 d, 1/2 P-H); 11.76 (1H, br s, NH–ibu); 11.99 (1H, br
s, NH); 13C NMR (CDCl3; DMTr resonances not listed) dC ꢀ5.26 and
0
0
0
0
0

M. Sobkowski et al. / Tetrahedron: Asymmetry 21 (2010) 410–419
419
14. Cieslak, J.; Jankowska, J.; Stawinski, J.; Kraszewski, A. J. Org. Chem. 2000, 65,
21. Faber, K. Chem. Eur. J. 2001, 7, 5004–5010.
22. Sigurdsson, S.; Stromberg, R. J. Chem. Soc., Perkin Trans.
1688.
23. Sobkowski, M.; Jankowska, J.; Stawinski, J.; Kraszewski, A. Nucleosides
Nucleotides Nucleic Acids 2005, 24, 1033–1036.
24. Seela, F.; Kretschmer, U. Chem. Commun. 1990, 1154–1159.
25. Garegg, P. J.; Lindh, I.; Regberg, T.; Stawinski, J.; Stromberg, R.; Henrichson, C.
Tetrahedron Lett. 1986, 27, 4055–4058.
7049–7054.
2
2002, 1682–
15. Ferguson, I. G. T.; Glidewell, C. J. Chem. Soc., Dalton Trans. 1977, 2071–2074.
16. Zain, R.; Stawinski, J. J. Chem. Soc., Perkin Trans. 2 1996, 795–799.
17. Jankowska, J.; Sobkowski, M.; Stawinski, J.; Kraszewski, A. Tetrahedron Lett.
1994, 35, 3355–3358.
18. Sobkowski, M.; Stawinski, J.; Kraszewski, A. Tetrahedron: Asymmetry 2008, 19,
2508–2518.
19. Stawinski, J.; Thelin, M.; Zain, R. Tetrahedron Lett. 1989, 30, 2157–2160.
20. Beak, P.; Anderson, D. R.; Curtis, M. D.; Laumer, J. M.; Pippel, D. J.;
Weisenburger, G. A. Acc. Chem. Res. 2000, 33, 715–727.
26. Arnold, L.; Smrt, J.; Zajicek, J.; Schiesswohl, M.; Ott, G.; Sprinzl, M. Collect. Czech.
Chem. Commun. 1991, 56, 1948–1956.
27. Stawinski, J.; Thelin, M. J. Org. Chem. 1991, 56, 5169–5175.