Stereochemistry of internucleotide bond formation by the H-phosphonate method. Part 3: Investigations on a mechanism of asymmetric induction

Article abstract of DOI:10.1016/j.tetasy.2007.09.017

The stereochemistry of H-phosphonate diester bond formation (including internucleotide ones) with ribonucleoside H-phosphonates as substrates has been investigated using ³¹P NMR spectroscopy. It was found that the reactions investigated owe the

Full text of DOI:10.1016/j.tetasy.2007.09.017

Available online at www.sciencedirect.com
Tetrahedron: Asymmetry 18 (2007) 2336–2348
Stereochemistry of internucleotide bond formation
by the H-phosphonate method. Part 3: Investigations
on a mechanism of asymmetric induction
Michal Sobkowski,a,* Adam Kraszewskia and Jacek Stawinskia,b
aInstitute of Bioorganic Chemistry, Polish Academy of Sciences, Noskowskiego 12/14, 61-704 Poznan, Poland
bDepartment of Organic Chemistry, Arrhenius Laboratory, Stockholm University, S-106 91 Stockholm, Sweden
Received 17 August 2007; accepted 14 September 2007
Abstract—The stereochemistry of H-phosphonate diester bond formation (including internucleotide ones) with ribonucleoside H-phos-
phonates as substrates has been investigated using 31P NMR spectroscopy. It was found that the reactions investigated owe their stereo-
selectivity to a dynamic kinetic asymmetric transformation. The absolute configurations of the compounds involved in the reaction
pathways were tentatively assigned on the basis of their 31P NMR chemical shifts and their correctness was verified for the H-phosphonic–
pivalic mixed anhydrides and H-phosphonate aryl esters.
Ó 2007 Elsevier Ltd. All rights reserved.
1. Introduction
2. Results and discussion
During condensation of deoxyribonucleoside 30-H-phos-
phonate monoesters with nucleosides both diastereomers
of the nascent diesters are usually formed in approximately
equal amounts. In contrast, in an analogous reaction
with ribonucleoside 30-H-phosphonates, the formation of
(DP)1–3 (SP) diastereomers is favoured, and thus the reaction
is stereoselective.4,5 This stereoselectivity was found to be
dependent mainly on structural elements of the reacting
H-phosphonate (the nucleobase and its protective groups,
sugar moiety protective groups), and also on the solvent
composition.4–6 Recently published papers have provided
some practical solutions to attain a relatively high stereose-
lectivity, although no attempt has been made to elucidate
the underlying mechanism of the observed phenomenon.5,6
The synthesis of ribonucleoside H-phosphonate diesters in-
volves activation of nucleoside 30-H-phosphonate monoes-
ters of type 1 with pivaloyl chloride with the formation of
diastereomeric mixtures of mixed anhydrides 2-DP(RP) and
2-LP(SP), which upon reaction with a hydroxylic compo-
nent afford the corresponding diesters 3 (Fig. 1). These
reactions are usually stereoselective and produce P-epi-
meric mixtures of H-phosphonates with the significant
advantage of only one diastereomer, identified as having
DP(SP) configuration.4,5 However, due to the high reactiv-
ity of the mixed anhydrides of type 2, the assignment of 31P
NMR signals 2A and 2B (Fig. 1b) to individual diastereo-
mers 2-DP(RP) and 2-LP(SP) cannot be done directly.
The observed stereoselectivity during the ribonucleoside H-
phosphonate condensation may have various origins,
which could arise from one of the following phenomena:
(i) dynamic thermodynamic resolution (DYTR),7,8 (ii) di-
astereoconvergent transformation,8 or (iii) dynamic kinetic
asymmetric transformation (DYKAT)8,9 (for the relevant
energy profiles see Fig. 2).
Since the stereoselectivity of H-phosphonate condensation
was expected to have its sources in the initial stages of the
H-phosphonate diester bond formation, we investigated
the stereochemical aspects of the activation of H-phospho-
nate monoesters and reactivity of the intermediates
produced.
Due to a similar stereoselectivity being observed for the esterification
with nucleosides and simple aliphatic alcohols,6 the model reactions were
performed with ethanol as a hydroxylic component.
*
Corresponding author. Tel.: +48 61 8528 503; fax: +48 61 8520 532;
e-mail: msob@ibch.poznan.pl
0957-4166/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2007.09.017

M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
2337
DMTrO
DMTrO
B
B
O
O
ROH
O
P
O
P
OTBDMSi
OTBDMSi
O
H
H
O
O
O
Pv
DMTrO
O
R
B
2-DP(RP)
3-LP(RP)
O
PvCl
+
+
O
OTBDMSi
P
DMTrO
DMTrO
H
B
B
O
O
O
ROH
O
P
O
P
OTBDMSi
OTBDMSi
1
H
O
O
H
O
O
DMTr = 4,4'-dimethoxytrityl
TBDMSi = tertbutyldimethylsilyl
B = protected nucleobase
Pv = pivaloyl
Pv
R
2-LP(SP)
3-DP(SP)
(main product)
ROH = nucleoside or alcohol
3a, R = Me
3b, R = Et
3c, R = nucleoside
O
O
DMTrU
DMTrU
O
P
H
O
Et
O
P
H
O
Pv
3b
2
DP(SP)
A
B
LP
(
RP)
(ii)
(i)
86 : 14
70 : 30
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
Figure 1. (a) A synthetic route from nucleoside H-phosphonate monoester 1 to diester 3. (b) Exemplary 31P NMR spectra of the reaction of the mixed
anhydride 2 with ethanol in DCM containing 3 equiv of 2,6-lutidine; (A) the mixed anhydride 2 (B) after addition of 2 equiv of ethanol.
According to the first mechanism (dynamic thermodynamic
resolution, Fig. 2a), the mixed anhydride diastereomers 2-
DP and 2-LP are configurationally stable enough to permit
their stereospecific esterification into the corresponding
diesters 3. Thus, the major DP(SP) epimer of product 3
should descend from the main epimer 2A of the intermedi-
ate mixed anhydride 2, and the minor LP(RP) epimer of
product 3, from the minor epimer 2B of intermediate 2.
for intermediate 2 was 70:30 (de = 40%) while that for
ethyl nucleoside H-phosphonate diester 3b, 86:14 (de =
72%; Fig. 1b). The significant difference in the diastereo-
meric ratio of intermediate 2 and product 3 indicated that
the thermodynamic stability of the diastereomers of 2 does
not dominate the observed stereoselectivity.
In the second scenario—diastereoconvergent transformation
(Fig. 2b), esterification of either of the diastereomers of
intermediate 2 [designated as 2-LP(SP) and 2-DP(RP)] was
expected to take a different course as shown in Figure 3.
However, integration of the 31P NMR signals of intermedi-
ate 2 and product 3 revealed that the ratio of diastereomers

2338
M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
(a) DYTR
(b) Diastereoconvergent Transformation
(c) DYKAT
5b
other tbp's
of type 5
5a
5c
5c
2-LP
2-DP
2-DP
2-LP
2-LP
3-DP
2-DP
3-LP
3-LP
3-DP
3-LP
3-DP
Figure 2. Possible energy profiles for stereoselective transformation of the mixed anhydride 2 into diester 3. (a) dynamic thermodynamic resolution; (b)
diastereoconvergent transformation (the dashed lines show alternative energies of intermediates 2); (c) dynamic kinetic asymmetric transformation.
RO
RO
B
B
RO
RO
O
O
B
B
O
O
OPv
P
O
P
OPv
O-
P
OR
OR
O
O
O
path b
Y
H
OR
OR
O
O
-O
P
H
OPv
H
H
OPv
2-LP(SP)
O
O
b
c
2-DP(RP)
R1
R1
R1OH
R1OH
5a
5b
PvO-
path c
RO
RO
B
B
O
O
RO
B
O
OR
OR
O
O
OPv
PvO-
OR
O
H
P
O
P
H
H
P
O
O-
O
O
O
R1
R1
R1
3-DP(SP)
3-LP(RP)
5c
(major)
(minor)
R1 = alkyl or nucleoside
=
=
protecting group
nucleobase
R
B
Pv = pivaloyl
Figure 3. Putative mechanism of induction of stereoselectivity due to differences of phosphorane intermediates.
The main path of esterification of the mixed anhydride 2
having an LP(SP)-configuration should proceed via phos-
phorane intermediate 5a (tbp with a pivaloyl residue in
the apical position) that can preferentially collapse to dies-
ter 3-DP(SP) with inversion of configuration.
investigated, one would have to assume that the activation
barrier leading to tbp 5c was higher (probably due to steric
factors) than that leading from 2-DP(RP) to tbp 5b (Fig. 2b)
and the latter cannot collapse to diester immediately10 but
undergoes pseudorotation to 5a. The assignment of the
absolute configuration to 2A and 2B (Fig. 1b) is a compli-
cated task in this scenario (this ambiguity is indicated by
solid and dashed lines in Fig. 2b); however, for a reaction
with high stereoselectivity one might expect the predomi-
nance of 2-LP versus 2-DP and thus, the assignment of
signals of 2 should be similar as for a DYTR mechanism
discussed above [i.e., major 2A = 2-LP(SP) and minor
2B = 2-DP(RP)].
For the other diastereomer of 2 [DP(SP)], two reaction
pathways involving tbp 5b (path b) and tbp 5c (path c)
might be available. Path c should produce the LP(RP)
diastereomer of diester 3 (inversion of configuration), while
path b, the same product as the one obtained from 2-
LP(SP) [i.e., 3-DP(SP); retention of configuration]. To
explain the observed stereoselectivity in the transformation

M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
2339
The major drawback of this hypothesis is that path b in
Figure 3 would involve energetically unfavourable tbp 5b
(with apical negatively charged oxygen)à and there is no
obvious explanation why this route could predominate
over path c.§ However, if both paths b and c (and other
possible ones) would be unfavourable, the epimer 2-DP
could be relatively stable, and if it could interconvert into
2-LP, the intermediacy of pseudorotation could be omitted
and the energy diagram would be identical as in the third
proposed mechanism, dynamic kinetic asymmetric transfor-
mation (Fig. 2c).7,8 In this scenario it is required that the
energy profile of the transformation is subject to the
Curtin–Hammett principle.11 For the investigated reac-
tions, the conditions require a rapid equilibrium between
diastereomers of the mixed anhydride 2, followed by the
esterification of 2-LP and 2-DP to form the corresponding
diastereomers of product 3 with different rates.
rapid interconversion of the corresponding diastereomers
(Eq. 1).k
2-DPðRPÞ þ PvOꢀ ꢀ 2-LPðSPÞ þ PvOꢀ
ð1Þ
2.1. Esterification of the mixed anhydrides of type 2
Since the fast equilibrium between the isomers of the mixed
anhydride 2-DP ꢀ 2-LP supported the DYKAT scenario,
we tried to assess the relative rates of the equilibration
and esterification as well as to find out which diastereomer
of this intermediate (2A or 2B, Fig. 1b) was more reactive
and thus was a precursor from which the main diastereo-
mer of the product, that is, the corresponding H-phospho-
nate diester 3, was formed. To this end, the generated
mixed anhydride 2 (B = U) was treated with various ali-
phatic alcohols (MeOH, EtOH, i-PrOH, t-BuOH; 5 equiv),
to see whether gradual changes in the ratio of diastereo-
mers of 2 during the course of the reaction could be regis-
tered. Unfortunately, the formation of the respective
H-phosphonate diesters of type 3 was completed before the
first 31P NMR spectrum could be recorded (ca. 1 min),
even for sterically hindered alcohols (e.g., t-BuOH) and
in the absence of a nucleophilic catalyst. The rates of the
reactions remained beyond the time suitable for kinetic
measurements by 31P NMR spectroscopy also when the
above alcohols were used in sub-stoichiometric amounts
(0.5 equiv). Some insight into the composition of the reac-
tion mixture at the initial stages of the reaction was gained
by quenching experiment (addition of aqueous ACN). This
revealed that the half-life of the reaction was ca. 13 s and
the ratio of diastereomers of the produced H-phosphonate
diester 3 was constant throughout the reaction (Fig. 4A).
Since such an equilibrium between the diastereomers of the
mixed anhydrides (2-DP ꢀ 2-LP) was crucial for the DY-
KAT mechanism, its existence had to be proven experimen-
tally. To this end, H-phosphonic–pivalic mixed anhydride 2
(B = U) was prepared in situ from a suitably protected
uridine 30-H-phosphonate 1 and pivaloyl chloride (1.2 equiv).
The 31P NMR spectroscopy (1H decoupled spectra)
revealed two signals at 1.39 and 1.47 ppm in a ratio ca.
1:2, assigned to a diastereomeric mixture of 2. The mixed
anhydride 2 was quite stable under the reaction conditions
[0.1 M solution in dichloromethane (DCM) containing
3 equiv of 2,6-lutidine, rt] and products of its degradation
appeared in the 31P NMR spectrum after ca. 1 h. No
changes in the ratio of diastereomers of 2 were observed
after several hours. However, the addition of triethylam-
monium adamantanecarboxylate (0.5 equiv) to the reaction
mixture caused an immediate (ca. 1 min; time required for
recording the first 31P NMR spectrum) appearance of two
new signals at 1.55 and 1.67 ppm (ca. 30%) due to the
mixed nucleoside H-phosphonic–adamantanecarboxylic
anhydride.– In separate experiments, it was found that
the ratio of the 31P NMR signals of these two mixed anhy-
drides proportionally changed with the amount of the car-
boxylate (pivalate or adamantanecarboxylate) added. A
similar equilibrium was observed when acetate anions were
added to the reaction mixture containing the mixed anhy-
dride 2 (data not shown); this allowed us to conclude that
the mixed anhydrides of type 2 might undergo a fast
exchange of the carboxylate moieties. As the solution of
the in situ generated 2 always contained some quantity of
pivalate anions (i.e., due to hydrolysis of PvCl by residual
water), these may react with the mixed anhydride 2 causing
When a nucleoside with a free 30-hydroxyl function [50-O-
(4,40-dimethoxytrityl)thymidine, 3 equiv] was used instead
of simple alcohols as a hydroxylic component, the esterifi-
cation became slow enough (ca. 8 min for completion) to
record a series of the 31P NMR spectra (Fig. 4B). From
the collected data depicted in Figure 4, it is clear that the
equilibrium between diastereomers of 2 (2-DP ꢀ 2-LP)
had to be significantly faster than their esterification. As
a consequence, no essential changes in the ratio of dia-
stereomers of intermediate 2 and the products, H-phos-
phonate diesters, could be observed over the course of
the reaction.
2.2. Experiments on kinetic quenching of the mixed
anhydrides of type 2
The data collected so far showed that the equilibration
between diastereomers of the mixed anhydride 2 is very
fast; this made the measurement of kinetic parameters of
particular P-epimers impossible. Since the efforts to reduce
the rate of equilibration 2-DP ꢀ 2-LP (e.g., by trimming
à There are several other stereoretentive pathways leading from 2-DP(RP)
to 3-DP(SP); however in each case the intermediacy of unfavourable tbps
is unavoidable.
§ One should keep in mind that the above discussed pseudorotational
mechanism, despite its rather low likelihood, could not be completely
rejected and may contribute to some extent to the observed
stereoselectivity.
– The assignment of new signals to diastereomers of adamantanecarboxylic
mixed anhydride was confirmed by spiking the reaction mixture with an
authentic sample of this compound (prepared from H-phosphonate 1
with adamantanecarbonyl chloride).
k The equilibrium between diastereomers of the mixed anhydride 2 may be
also promoted by chloride anions present in the reaction mixture.
Due to extremely high reactivity of mixed anhydrides of type 2, all
attempts to isolate these species in a pure form failed and the
experiments were performed with the DCM solutions of compound 2
generated in situ.

2340
M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
A
B
100%
100%
80%
60%
40%
20%
0%
80%
60%
40%
20%
0%
0
10
20
30
40
50
60
70
80
90
100
0
1
2
3
4
5
6
7
8
Time [seconds]
Time [minutes]
Figure 4. Reaction of uridine H-phosphonate 1 (B = U) with (A) EtOH (quenched with aqueous ACN) and (B) 50-O-dimethoxytrityl-thymidine,
promoted by pivaloyl chloride (1.5 equiv) in the presence of 3 equiv of 2,6-lutidine (31P NMR data). ---j--- % of monoester 1; ---s--- % of the mixed
anhydride 2; ---h--- % of the product [H-phosphonate diester 3b (A) or 30-30 dinucleoside H-phosphonate (B), sum of diastereomers]; —e—the fraction
of the downfield signal of the mixed anhydride 2; —m— the fraction of DP diastereomer of the product [DP(SP) H-phosphonate diester 3b (A) or DP(RP)
30-30 dinucleoside H-phosphonate (B)].
down the concentration of pivalic or chloride anions that
might be involved in the catalysis of the equilibration 2-
DP ꢀ 2-LP) were unsuccessful, experiments on kinetic
quenching were attempted. We expected that the compari-
son of data from the regular and quenching experiments
could help us in the assignment of the configuration of dia-
stereomers 2A and 2B.
sonable to assume that the very reactive alcohol used in
large excess esterified both diastereomers of the mixed
anhydride 2 with substantially higher rates than the rate
of the equilibrium 2-DP ꢀ 2-LP, and thus under such con-
ditions the major LP(SP)-epimer of diester 3a was formed
from the major diastereomer 2A, and consequently 3a-
DP(RP), from 2B (DYTR kinetics, Figs. 2 and 6a). As
the esterification proceeds with the inversion of configura-
tion, P-epimer 2A should thus have DP(SP)-configuration,
while P-epimer 2B, LP(RP)-configuration.
In preliminary experiments, stereospecific sulfurization4 of
the mixed anhydride 2 (B = U; the ratio 2A: 2B ꢁ 2:1)
yielded a configurationally stable mixture of P-epimers of
phosphorothioic–pivalic mixed anhydride [dP 50.0 and
51.2 ppm], whose ratio depended significantly on the basi-
city of the medium. In the presence of 2,6-lutidine, this
ratio was ca. 1:1, while when sulfurization was carried
out in the presence of triethylamine (TEA), the upfield dia-
stereomer dominated (ratio ca. 1:2). Since strong bases are
known to speed up the sulfurization of H-phosphonates,12
one could speculate that the stereochemistry observed in
the presence of TEA resulted from kinetic quenching of
the mixed anhydride 2.
In contrast to the above kinetic quenching experiment,
under standard reaction conditions (small excess of an
alcohol), the equilibration 2-DP ꢀ 2-LP was presumably
significantly faster than the esterification and therefore
the major diastereomer DP(SP) of the produced diester 3
should originate from the minor diastereomer 2B of the
mixed anhydride 2, having an LP(RP)-configuration
(Fig. 6b).
For other nucleotide derivatives (B = ABz, CBz, Gibu) the
kinetic quenching experiments gave less clear-cut results
(Fig. 5A, C and G), but also in these cases the trend in
changing the ratio of diastereomers of product 3a was con-
gruent with the postulated DYKAT mechanism.
Unfortunately, phosphorothioic–pivalic mixed anhydrides
were found to be of little use due to their limited stability.
Nevertheless, the kinetic quenching observed during the
fast sulfurization of the mixed anhydride 2 encouraged us
to search for conditions under which a similar kinetic
quenching could be achieved for the esterification of 2.àà
To achieve this goal, a DCM solution (0.3 mL) of the
in situ prepared mixed anhydride 2 (B = U) was injected
into methanol (5 mL, 2500 equiv). Analysis of the 31P
NMR spectra revealed that the esterification was quantita-
tive and that the intensity of signals of the produced methyl
nucleoside H-phosphonate 3a (B = U) has been reversed,
compared to the standard reaction in which only a few
equivalents of methanol were used (Fig. 5U). It seems rea-
2.3. Transesterification of aryl nucleoside H-phosphonates
of type 4
The above experiments with H-phosphonic–pivalic mixed
anhydrides of type 2 implied that the DYKAT mechanism
was a source of stereoselectivity during the condensation of
ribonucleoside H-phosphonates. Unfortunately, the very
high rate of equilibration of the intermediate mixed anhy-
drides prevented a quantification of reactivities of their
individual diastereomers, and the assumed DYKAT kinet-
ics could not be proved directly. Thus, we have focused our
attention on other types of active derivatives, also bearing
electron-withdrawing ligands, namely aryl nucleoside
àà In mechanistic terms this meant that dynamic kinetic process (DYKAT)
would be changed into dynamic thermodynamic one (DYTR).7

M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
2341
(U)*
(A)
(i)
(ii)
(i)
(ii)
DMTrABzPHPv
DMTrUPHPv
+ 5 equiv.
+ 2500 equiv.
MeOH
+ 5 equiv.
MeOH
+ 2500 equiv.
MeOH
MeOH
8.75
8.50
8.25
10.25
10.00
9.75
9.00 8.75 8.50 8.25 8.00
10.33 10.17 10.00 9.83 9.67
(C)
(G)**
(i)
(ii)
(i)
(ii)
DMTrCBzPHPv
DMTrGibuPHPv
+ 5 equiv.
MeOH
+ 2500 equiv.
MeOH
+ 5 equiv.
MeOH
+ 2500 equiv.
MeOH
9.00 8.75 8.50 8.25 8.00 7.75
10.67 10.50 10.33 10.17 10.00 9.83
9.00
8.75
8.50
10.25
10.00
9.75
9.50
Figure 5. 31P NMR spectra of methyl nucleoside H-phosphonates of type 3 prepared (i) under standard conditions [using 5 equiv of MeOH (10 lL)]; (ii)
by quenching the reaction with 5 mL (ca. 2500 equiv) of MeOH. (A) B = ABz, (C) B = CBz, (G) B = Gibu, (U) B = U. Different chemical shifts in panels (i)
and (ii) are due to differences in solvent compositions.6 * The reversal of product distribution for uridine derivative was confirmed by evaporating the
solvent, dissolving the residue in DCM and mixing with an aliquot of reaction (i). ** For guanosine derivative the signals of diastereomers in DCM have
apparently inverted positions.6
H-phosphonates13,14 of type 4 (B = U, Fig. 7), the reacti-
vity of which can be controlled and adjusted.15,16 We have
chosen three aryl groups, phenyl, p-chlorophenyl and
p-nitrophenyl, that clearly diversified the reactivity of the
phosphorus centre in H-phosphonate diesters 4.
the diastereomers of intermediate 4 exist in an equilibrium
(Eq. 2), similarly as it was found for the mixed anhydrides
2 (Eq. 1).
4-DPðRPÞ þ ArOH ꢀ 4-LPðSPÞ þ ArOH;
ð2Þ
From a stereochemical point of view, this approach con-
sisted of three events: the formation of two diastereomers
of the mixed anhydride 2 (i.e., 1!2, Fig. 7A, discussed in
previous sections), formation of aryl nucleoside H-phos-
phonate (i.e., 2!4) and its transesterification with an alco-
hol (i.e., 4!3). It is believed (and partially proven13) that
We found that the esterification of the mixed anhydride 2
with phenol, p-chlorophenol and p-nitrophenol (3 equiv)
was complete within <1–2 min (the time required for
recording the first 31P NMR spectrum). The ratios of the
produced aryl nucleoside H-phosphonate diester diastereo-
mers were 80:20, 81:19 and 71:29, for 4c, 4b and 4a, respec-

2342
M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
tively, and these were apparently equilibrium values as no
(a) DYTR
(b) DYKAT
changes were observed when reactions were left for several
hours (31P NMR). Sacrificing the advantages of pseudo-first
order kinetics, we were able to slow down these reactions by
reducing the excess of phenols from 3 equiv to 1.2 equiv.
This manoeuvre allowed us to observe a gradual appear-
ance of aryl H-phosphonates 4 at the expense of intermedi-
ate 2 during the 10 min required for the completion of the
reaction. The ratio of the produced diastereomers of diest-
ers 4 was constant during the progress of the reaction,
and was the same as for the reactions with 3 equiv of
phenols. Furthermore, the ratio of the diastereomers of 2
remained unchanged, and this precluded drawing tentative
conclusions regarding a correlation of the 31P NMR chem-
ical shifts of the signals and configuration at the phosphorus
centre during the formation of aryl diesters 4 (Fig. 8).
slow
fast
2-LP(SP)
minor
2-LP(SP)
minor
2-DP(RP)
major
2-DP(RP)
major
fast
fast
slow
fast
+ MeOH
+ EtOH
(a few equiv.)
(large excess)
3a-LP(RP)
main product
3b-LP(RP)
3a-DP(SP)
3b-DP(SP)
main product
Figure 6. Postulated mechanisms of stereoselective esterification of the
mixed anhydride 2. (a) DYTR, with a large excess of methanol. The rate
of equilibration is significantly lower than the rates of esterification. (b)
DYKAT, with several equivalents of ethanol. The rate of esterification of
2-LP(SP) is comparable to the rate of equilibration and significantly higher
than the rate of esterification of 2-DP(RP).
DMTrO
DMTrO
DMTrO
A
B
B
B
O
O
O
ArOH
ROH
O
P
O
P
O
P
OTBDMSi
OTBDMSi
OTBDMSi
O
H
H
O
O
H
DP(RP)
LP(SP)
DP(SP)
O
O
O
Ar
Pv
O
R
PvCl
1
DMTrO
DMTrO
DMTrO
B
B
B
O
O
ArOH
ROH
O
P
O
P
O
P
OTBDMSi
OTBDMSi
OTBDMSi
H
O
O
H
H
O
LP(SP)
DP(RP)
LP(RP)
O
O
O
Pv
Ar
R
4a, Ar = p-nitrophenyl
4b, Ar = p-chlorophenyl
4c, Ar = phenyl
3a,R =M e
3b,R =E t
2A and 2B
B
DP(SP)
LP
(
RP
)
55%
45%
71%
29%
(iv)
3b (DMTrUPHEt)
(iii)
(iI)
4a (DMTrUPHPNP)
2 (DMTrUPHPv)
70%
30%
(i)
1 (DMTrUPH
)
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
Figure 7. (A) Stereochemistry of reactions involving aryl nucleoside H-phosphonates. (B) An example of 31P NMR traces of the above reactions leading to
ethyl uridine H-phosphonate 3b, using p-nitrophenyl (PNP) intermediate 4a (minor signals about 1.5 ppm in spectrum (iii) are due to an equilibrium
between 2 and 4a). For abbreviations, see Figure 1.

M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
2343
100%
80%
60%
40%
20%
0%
the first spectrum could be recorded with no changes ob-
served in the later spectra. In order to gain more insight
into the composition of the reaction mixture during transe-
sterification, the reaction was repeated and the aliquots
taken 10–90 s after the addition of methanol, were
quenched with aqueous acetonitrile (ACN). In this way,
the reactive H-phosphonates 2 and 4a were hydrolyzed rap-
idly to H-phosphonate monoester 1, while H-phosphonate
diester 3b remained intact and its contents and diastereo-
meric composition could be analyzed. This revealed a grad-
ual decrease in the fraction of the DP(SP)-diastereomer of
product 3 during the progress of the reaction (56!43%),
however, a correlation of these changes with those in the
ratio of diastereomers of 4a, was not possible.
0
2
4
6
8
10
12
Time [minutes]
Figure 8. Reaction of the mixed anhydride 2 (generated in situ) with p-
chlorophenol (1.2 equiv) in DCM containing 3 equiv of 2,6-lutidine (31
P
Similar, but clearer results (longer reaction time) were
obtained for the reaction of aryl H-phosphonate 4a with
ethanol (Fig. 9B). In this instance, the minor diastereomer of
4a was consumed very quickly and after 2 min. in the 31P
NMR spectra only the major diastereomer of 4a could be
detected. Moreover, simultaneous changes in the ratio of
diastereomers of the forming ethyl nucleoside H-phospho-
nate diester 3b were observed. Initially (i.e., quenching after
10 s) the ratio of the product diastereomers was 70:30 while
after 5 min, it decreased to 60:40. Such a trend indicated
that it could be the minor diastereomer of the intermediate
that was the precursor of the major isomer of product 3b,
and that the consumption of the minor diastereomer of 4a
was probably faster than its regeneration from the less reac-
tive major diastereomer. This would explain why the minor
diastereomer of 4a disappeared quickly from the spectrum,
and also why the ratio of diastereomers of product 3b
decreased. It should be noted that this interpretation is in
line with the preliminary conclusions regarding reactivities
of the mixed anhydrides of type 2 (vide supra) and strongly
supports the DYKAT scenario for their esterification.
NMR data). ---h--- % of the product (p-chlorophenyl nucleoside H-
phosphonate 4b; sum of diastereomers); ---s--- fraction of the downfield
signal of the mixed anhydride 2a; —m— fraction of DP(RP) diastereomer
(downfield signal) of the product (p-chlorophenyl nucleoside H-phospho-
nate 4b).
The initial experiments on transesterification of aryl H-
phosphonates of type 4 (the last stage in Fig. 7A) were per-
formed in order to find reactions having rates in a range
that would enable tracking their progress with 31P NMR
spectroscopy. Very fast reactions could not be monitored
by 31P NMR (at least 1 min was required to register the
spectrum), while the reactions lasting longer than one hour
had to be excluded due to the possible subsequent side
products formation (e.g., further transesterification17,18 or
disproportionation13,19 of H-phosphonate diesters under
basic conditions, or detritylation due to released HCl,
PvOH and acidic phenols).
To this end, three aryl diesters 4a, 4b and 4c were treated
with various alcohols and the half-life times of the reac-
tions were estimated by 31P NMR (Table 1). The transesteri-
fication of phenyl H-phosphonate diester 4c was too slow
to be useful for our purposes and was abandoned. The
convenient range of the reaction rates of several minutes
was found for the transesterification of p-nitrophenyl
uridine H-phosphonate 4a with various alcohols, and for
p-chlorophenyl uridine H-phosphonate 4b with methanol.
These combinations of the reactants were investigated in
more detail.
Reactions of aryl H-phosphonate 4a with sterically hin-
dered alcohols showed notably higher stereoselectivities
than those with MeOH and EtOH, indicating that the
equilibration rate of diastereomers of 4a was significantly
higher than the rate of their transesterification (DYKAT
scenario). While for isopropanol, an increase of the ratio
of diastereomers of 4a was still perceptible (Fig. 9C), for
tert-butanol it remained constant (Fig. 9D). In both these
cases, the ratio of diastereomers of the products did not
vary over the course of the reactions.
2.4. Transesterification of p-nitrophenyl nucleoside
H-phosphonates 4a
The best experiment with ethanol (Fig. 9B) was repeated
for H-phosphonates bearing other nucleobases (Fig. 10)
and the general trend in changes of the ratio of diastereo-
mers was found to be similar in all the cases. Thus, the con-
clusions concerning the kinetics of transesterification
drawn for p-nitrophenyl uridine H-phosphonate seem to
be valid for all four H-phosphonates 4a.§§
The progress of the reactions of p-nitrophenyl ester 4a with
alcohols is shown in Figure 9. In the case of methanol
(Fig. 9A), the transesterification was practically over before
Table 1. Estimated half-life times of transesterification of aryl uridine H-
phosphonates 4a–c (dmtUPHAr) with 5 equiv of alcohols (interpolation of
31P NMR data)
Ar
+MeOH
+EtOH
+i-PrOH
+t-BuOH
§§ For cytidine derivative (Figs. 10C and 11C), the signals of diester 3
(B = CBz) overlapped, precluding their individual integration. The final
fraction of DP(SP) diastereomer (shown in the figures) was determined
by changing the solvent of the quenched reaction mixture from DCM to
toluene where the signals were well separated.6
t1/2
t1/2
t1/2
t1/2
Ph
pCl–Ph
pNO2–Ph
150 min
35 min
10 s
330 min
90 min
20 s
500 min
180 min
55 s
630 min
270 min
170 s

2344
M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
A (MeOH)
B (EtOH)
100%
100%
80%
60%
40%
20%
0%
80%
60%
40%
20%
0%
0
1
2
3
4
5
6
7
8
9
0
0.5
1
1.5
2
2.5
3
Time [minutes]
Time [minutes]
C (iPrOH)
D (tBuOH)
100%
80%
60%
40%
20%
0%
100%
80%
60%
40%
20%
0%
0
1
2
3
4
5
6
7
8
9
10
0
5
10
15
Time [minutes]
20
25
30
Time [minutes]
Figure 9. Reactions of p-nitrophenyl uridine H-phosphonate 4a (B = U) with (A) MeOH, (B) EtOH, (C) i-PrOH and (D) t-BuOH (5 equiv) in DCM
containing 3 equiv of 2,6-lutidine (31P NMR data). Line description for Figures 9–11, ---s--- the fraction of the downfield signal of aryl nucleoside
H-phosphonate of type 4; ---h--- % of the product (alkyl nucleoside H-phosphonate of type 3; a sum of diastereomers); —m— the fraction of DP(SP)
diastereomer of product 3. Grey-filled data points derived from reactions quenched with water.
2.5. Esterification of p-chlorophenyl nucleoside
H-phosphonates 4b
H-phosphonate diesters of type 4 have an LP(SP)-configu-
ration, are substantially more reactive than DP(RP)-diaste-
reomers, and are usually precursors of the major DP(SP)
diastereomers of the produced alkyl nucleoside H-phos-
phonate diesters of type 3 according to the DYKAT
mechanism (cf. Fig. 5A).–– However, when the rate of
equilibration (Eq. 2) was not fast enough to keep the ratio
of DP/LP diastereomers constant, the transesterification
produced the respective epimers of 3 in a smaller or even
in the reversed ratio due to a higher participation of the
DYTR scenario. It may be expected that by increasing of
the transesterification rate of the moderately reactive aryl
esters further, this might result in achieving a kinetics of
an almost pure DYTR.
According to the preliminary data (Table 1), the reactions
of p-chlorophenyl nucleoside H-phosphonates of type 4b
with methanol were expected to give at least equally distinct
results as for p-nitrophenyl ester 4a. Indeed, we observed
not only clear-cut changes in the ratio of diastereomers of
aryl H-phosphonate 4b (similarly as for 4a, vide supra),
but also a substantial decrease in the fraction of the
DP(SP)-epimer of the produced H-phosphonate diester 3
(Fig. 11).§§ This second phenomenon was probably caused
by the fast consumption, in the early stages of the reaction,
of the minor LP(SP)-diastereomer of aryl H-phosphonate
4b, evidently the more reactive substrate for 3-DP(SP). As
a result, after a relatively short period of time the reaction
mixture contained a large excess of the DP(RP)-epimer of
intermediate 4b, which apparently underwent faster transe-
sterification with reactive methanol [yielding 3-LP(RP)] than
transformation into its epimer 4b-LP(SP) [according to Eq.
2]. Fast transesterification of 4b-DP(RP) with methanol led
to an accumulation of 3-LP(RP) that ultimately became
the main epimer of the product.
2.6. Experiments on kinetic quenching of aryl nucleoside
H-phosphonates of type 4
The above conclusions were confirmed by performing
transesterification of aryl H-phosphonates 4a-b (B = ABz,
CBz, Gibu, U) under the conditions of kinetic quench-
ing with a large excess of methanol. Thus, when aryl
–– Attempts to obtain quantitative kinetic data were unsuccessful due to
complexity of the reaction mixtures.
Concluding the above set of experiments, the acquired data
suggested that the minor diastereomers of aryl nucleoside

M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
2345
U
100%
A
100%
80%
60%
40%
20%
0%
80%
60%
40%
20%
0%
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
8
9
Time [minutes]
Time [minutes]
C
G
100%
100%
80%
60%
40%
20%
0%
80%
60%
40%
20%
0%
0
1
2
3
4
5
6
7
8
9
10
0
1
2
3
4
5
6
7
8
9
Time [minutes]
Time [minutes]
Figure 10. Reactions of p-nitrophenyl nucleoside H-phosphonates of type 4a with ethanol (31P NMR data). (U) B = U, (A) B = ABz, (C) B = CBz and (G)
B = Gibu. ---s--- downfield signal of 4a; ---h--- % of product 3 (DP + LP); —m— 3-DP(SP). For detailed description, see the legend for Figure 9.
H-phosphonates 4a,b were treated with 2500 equiv of
methanol, a clear-cut reversal of stereoselectivity was
obtained [i.e., the predominant formation of the LP(RP)
diastereomer of produced 3a, Table 2, column 5] for both
aryl derivatives and all four nucleobases. In these cases,
the diastereomeric composition of the produced H-phos-
phonate diester 3a (Table 2, column 5) practically mirrored
the population of diastereomers of the intermediate aryl H-
phosphonates 4 (Table 2, column 2). Apparently, the minor
diastereomer of intermediate 4-LP(SP) yielded the minor
diastereomer of product 3a-DP(SP), while the major
diastereomer 4-DP(RP) yielded the major diastereomer of
the product, 3a-LP(RP). This could be rationalized by
the kinetic quenching experiments of the mixed anhydrides
of type 2 (vide supra).
Table 2 contains additional data (some of them discussed
previously), containing a comparison of stereochemistries
of the reactions of H-phosphonates 2, 4a and 4b. Thus,
kinetic quenching of the mixed anhydride 2 with methanol
caused, in general, a loss of stereoselectivity (Table 2, cf.
column 2 and 5) that favoured the DP(SP) epimer of prod-
uct 3 in the regular reactions (columns 3 and 4). A similar
stereochemical outcome was found for the reactions of p-
nitrophenyl H-phosphonate intermediate 4a with limited
amounts of alcohols (columns 3 and 4), and for 4b + EtOH
(5 equiv). Most probably, the lack of stereoselectivity in
these cases was due to an intermediate character of kinetics
of the reactions, having the energetic profiles inbetween
DYKAT and DYTR (cf. Fig. 2).
assuming
a high rate of transesterification of aryl
H-phosphonates 4 with methanol, and a moderate rate of
their P-epimerization (DYTR kinetics).
A similar reactivity of nucleoside 30-H-phosphonates acti-
vated by pivaloyl chloride and that of the corresponding
aryl esters13–16 allows us to extend the assignment of the
configuration made for the diastereomeric aryl H-phospho-
nates 4 to pivalic mixed anhydrides 2, that is, to assign the
DP(RP)-configuration to the major diastereomer A and the
LP(SP), to the minor diastereomer B of the mixed anhy-
dride 2 (Fig. 2). Importantly, this assignment is in full
agreement with the preliminary conclusions drawn from
It should be noted that the stereoselectivities obtained in
the regular esterification of the H-phosphonic–pivalic
mixed anhydrides 2 were considerably higher than those
found for aryl H-phosphonates 4a/b (Table 2, columns 3
and 4), despite the reactivities of 2 and 4a towards alcohols
being comparable [t1/2 = ꢂ13 s (Fig. 4A) and ꢂ20 s (Table
1), respectively]. In contrast, the stereoselectivity of transe-
sterifications of 4a and 4b was similar (Table 2), despite

2346
M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
U
A
100%
80%
60%
40%
20%
0%
100%
80%
60%
40%
20%
0%
0
50
100
150
200
0
50
100
150
200
Time [minutes]
Time [minutes]
C
G
100%
100%
80%
60%
40%
20%
0%
80%
60%
40%
20%
0%
0
50
100
150
200
0
50
100
150
200
Time [minutes]
Time [minutes]
Figure 11. Reactions of p-chlorophenyl nucleoside H-phosphonates of type 4b with methanol (31P NMR data). (U) B = U, (A) B = ABz, (C) B = CBz and
(G) B = Gibu. ---s--- downfield signal of 4b; ---h--- % of product 3 (DP + LP); —m— 3-DP(SP). For detailed description, see the legend for Figure 9.
Table 2. Diastereomeric composition of diesters 3a obtained from substrates 2, 4a and 4b by treatment with 5 equiv of EtOH, or 5 or 2500 equiv of MeOH
Substrate
Substrate,
% of LP(SP)
2
Product 3b, % of DP(SP)
(substrate + 5 equiv of EtOH)
3
Product 3a, % of DP(SP)
(substrate + 5 equiv of MeOH)
4
Product 3a, % of DP(SP)
(substrate + 2500 equiv of MeOH)
5
1
2 (B = ABz
)
26.4
38.9
34.6
30.0
27.6
34.0
30.5
25.2
19.5
27.8
22.7
19.0
81.9
74.6
80.5
81.0
54.1
49.0
62.0
54.7
80.2
75.7
80.0
81.0
48.1
54.8
51.0
42.0
32.5
39.0
43.5
41.3
57.9
47.1
51.7
35.7
29.7
36.4
37.6
29.6
20.8
28.4
25.3
22.7
2 (B = CBz
)
2 (B = Gibu
2 (B = U)
)
4a (B = ABz
)
)
)
4a (B = CBz
4a (B = Gibu
4a (B = U)
4b (B = ABz
)
)
)
ca. 55a
ca. 53a
4b (B = CBz
4b (B = Gibu
4b (B = U)
—
b
ca. 55a
a Due to low rate of transesterification (several hours) a significant decomposition of H-phosphonate diesters took place before the completion of the
reaction.
b Not determined due to decomposition and overlapping of the signals.
significant differences in their reactivities, (ca. two orders of
magnitude, cf. Table 1). Thus, the equilibration rate is pre-
sumably governed by the nucleophilic character of the
interchanging ligand (PvOꢀ > pNO2–PhOꢀ ꢁ pCl–PhOꢀ),
while the esterification rate depends on the leaving group
ability (PvOꢀ ꢁ pNO2–PhOꢀ > pCl–PhOꢀ). As a result,
aryl H-phosphonate esters of type 4 are apparently more
prone to the DYTR scenario than the mixed anhydrides
2.kk
kk However, for the reactions with sterically hindered alcohols, i-PrOH and
particularly t-BuOH, these differences were significantly smaller and the
aryl H-phosphonates 4 reacted also with high stereoselectivity (for
example, see Fig. 9C and D).

M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
2347
3. Conclusion
and used within one month. Phenol (Merck), p-chlorophe-
nol (Aldrich), p-nitrophenol (Aldrich), adamantanecarbon-
yl chloride (Fluka), acetic acid (POCh) were of commercial
grade and used without purification. Nucleoside H-phos-
phonates17 were obtained according to the published
method.
The aim of this work was to elucidate mechanistic and ste-
reochemical aspects of the condensation of ribonucleoside
H-phosphonates with nucleosides and alcohols. The results
of the experiments performed supported a mechanism
involving dynamic kinetic asymmetric transformation (DY-
KAT) which consists of a very fast equilibrium between
diastereomers of the intermediate mixed anhydrides or aryl
esters, followed by rate-diversified esterification of each
active diastereomer. However, some participation of a dia-
stereoconvergent transformation-type mechanism involving
pseudorotation of the phosphorane intermediates could
not be excluded. For several reactions in which a large ex-
cess of methanol was used (kinetic quenching), a reversal of
stereoselectivity was found, indicating a kinetics of dynamic
thermodynamic resolution (DYTR) under such conditions.
The relative ease of reaching this kinetic quenching
suggested that the rates of the equilibria involved and the
esterification reactions are of the same order of magnitude.
In that way the Curtin–Hammett principle might be
followed only partly and the diastereomeric composition
of the mixed anhydride (or aryl diester) could influence
the ratio of diastereomers of the final diesters formed.
Immediately prior to the reaction, all nucleosidic deriva-
tives were rendered anhydrous by dissolving in DCM
(0.5 mL/0.05 mmol) followed by toluene (3 mL/0.05 mmol)
and the evaporation of these solvents under reduced pres-
sure (pyridine was avoided in order not to contaminate
the mixture with a nucleophilic catalyst). After drying
under vacuum (15 min, 0.5 Torr), the flask was filled with
air which has been dried by passing through Sicapent
(Merck).
4.1. General procedure for the condensation of H-phospho-
nates of type 1 with alcohols and phenols
Nucleoside H-phosphonate 1 (0.05 mmol) was dissolved in
0.5 mL of the appropriate solvent (DCM or ACN contain-
ing 3 equiv of 2,6-lutidine) and dry alcohol (5 equiv) or
phenol (3 equiv) was added, followed by pivaloyl chloride
(1.2 equiv). For kinetic analysis, the reaction mixture was
transferred immediately to an NMR tube and sets of 8
scans (ca. 40 s acquisition time) for each consecutive
spectrum were registered.
Under standard reaction conditions, amongst the pair of
diastereomers of H-phosphonic–pivalic mixed anhydride
2 and aryl H-phosphonate 4 intermediates, the minor dia-
stereomers were identified as the more reactive species
yielding the major [DP(SP)] diastereomers of the produced
H-phosphonate diesters. As a consequence, the configura-
tion at the phosphorus centre in these more reactive and
thermodynamically less stable intermediates was assigned
as LP(SP).
4.2. General procedure for condensation of H-phosphonates
with nucleosides
Nucleoside H-phosphonate 1 (0.05 mmol) and nucleoside
[20,30-O-dibenzoyl-uridine or 50-O-(4,40-dimethoxytrityl)-
thymidine; 0.06 mmol] were mixed together, dried and dis-
solved in 0.5 mL of DCM containing 3 equiv 2,6-lutidine
followed by pivaloyl chloride (1.2 equiv), and the 31P
NMR spectrum was recorded.
Interestingly, a literature survey indicates that the DYKAT
mechanism may be an underestimated common phenome-
non taking place in the transformations of several other
types of P-chiral nucleotide derivatives. For example, DY-
KAT might explain the diastereoselective synthesis of
LP(RP) dinucleoside methylphosphonates20,21 or the rather
unusual retention of configuration during the SN2(P) reac-
tion of chlorooxazaphospholidines.22,23
4.3. General procedure for in situ preparation of the mixed
anhydrides of type 2
Nucleoside H-phosphonate 1 (0.05 mmol) was rendered
anhydrous by the evaporation of the added toluene
(3 mL) under reduced pressure. After drying under vacuum
(15 min, 0.5 Torr), the flask was filled with air which has
been dried by passing through Sicapent (Merck). The resi-
due was dissolved in 0.5 mL of the appropriate solvent
(DCM, ACN, or toluene) and 2,6-lutidine (3.0 equiv) and
pivaloyl chloride (1.2 equiv) were added successively. The
31P NMR spectra showed that under such conditions, the
mixed anhydrides of type 2 were stable for several hours
without the appearance of detectable degradation
products.
4. Experimental
31P NMR spectra were recorded at 121 MHz on a Varian
Unity BB VT spectrometer. 31P NMR experiments were
carried out in 5 mm tubes using 0.1 M concentrations of
phosphorus-containing compounds in the appropriate
solvents (0.5 mL) and the spectra were referenced to 2%
H3PO4 in D2O (external standard). Dichloromethane
(POCh, Poland) and acetonitrile (Merck) were refluxed
over P2O5 and distilled. Toluene (POCh) and 2,6-lutidine
(Fluka) were distilled and after discarding ca. 30% of fore-
runnings the fractions containing below 20 ppm of water
were collected. All solvents were stored over molecular
4.4. General procedure for esterification of the mixed
anhydrides of type 2 or aryl nucleoside diesters of type 4
˚
sieves 4 A and contained below 20 ppm of water (Karl
Typical procedure: to a solution of 2 or 4 (generated as
described above) dry alcohol was added in the appropriate
amount with vigorous stirring, and the mixture was ana-
lyzed by 31P NMR spectroscopy. For the experiments
Fischer coulometric titration, Metrohm 684 KF coulome-
ter). Anhydrous triethylamine (POCh) was distilled and
kept over CaH2. Pivaloyl chloride (Fluka) was distilled

2348
M. Sobkowski et al. / Tetrahedron: Asymmetry 18 (2007) 2336–2348
requiring the addition of <5 lL of reagents their 1 M solu-
Acknowledgements
tions in DCM or ACN were used.
The financial support from Polish Ministry of Science and
Higher Education and the Swedish Research Council, is
gratefully acknowledged.
The composition of the reaction mixtures in the early
stages of the fast reactions was determined by hydrolytic
quenching. To this end, after defined time periods 170 lL
aliquots of the reaction mixtures were taken and added
to 336 lL of ACN–pyridine–H2O, 300:18:18, and the
mixture was analyzed by 31P NMR spectroscopy.
References
1. Sobkowski, M.; Stawinski, J.; Kraszewski, A. Nucleosides
Nucleotides Nucleic Acids 2005, 24, 1301–1307.
2. Sobkowski, M.; Stawinski, J.; Kraszewski, A. Nucleosides
Nucleotides Nucleic Acids 2006, 25, 1363–1375.
3. Sobkowski, M.; Stawinski, J.; Kraszewski, A. Nucleosides
Nucleotides Nucleic Acids 2006, 25, 1377–1389.
4. Almer, H.; Stawinski, J.; Stro¨mberg, R.; Thelin, M. J. Org.
Chem. 1992, 57, 6163–6169.
4.5. General procedure for kinetic quenching experiments
The solution (0.3 mL) of intermediate 2 or 4 (0.5 mmol;
generated as described above) was added dropwise by a
syringe to a septum-sealed flask containing vigorously stir-
red alcohol (5 mL) and 2,6-lutidine (120 lL). After 5 min, a
sample of 0.5 mL was used for recording a 31P NMR spec-
trum. The remaining solution was concentrated under
reduced pressure (without heating), dissolved in DCM
(1 mL) and analyzed by 31P NMR spectroscopy.
5. Almer, H.; Stawinski, J.; Stro¨mberg, R. Nucleic Acids Res.
1996, 24, 3811–3820.
6. Sobkowski, M.; Jankowska, J.; Stawinski, J.; Kraszewski, A.
Nucleosides Nucleotides Nucleic Acids 2005, 24, 1033–1036.
7. Beak, P.; Anderson, D. R.; Curtis, M. D.; Laumer, J. M.;
Pippel, D. J.; Weisenburger, G. A. Acc. Chem. Res. 2000, 33,
715–727.
4.6. 50-O-Dimethoxytrityl-20-O-t-butyldimethylsilyl-uridin-
30-yl phosphorothioic–pivalic mixed anhydride
8. Faber, K. Chem. Eur. J 2001, 7, 5004–5010.
9. Trost, B. M.; Bunt, R. C.; Lemoine, R. C.; Calkins, T. L. J.
Am. Chem. Soc. 2000, 122, 5968–5976.
10. Westheimer, F. H. Acc. Chem. Res. 1968, 1, 70–78.
11. Seeman, J. I. Chem. Rev. 1983, 83, 83–134.
12. Jankowska, J.; Sobkowska, A.; Cieslak, J.; Sobkowski, M.;
Kraszewski, A.; Stawinski, J.; Shugar, D. J. Org. Chem. 1998,
63, 8150–8156.
13. Cieslak, J.; Szymczak, M.; Wenska, M.; Stawinski, J.;
Kraszewski, A. J. Chem. Soc., Perkin Trans. 1 1999, 3327–
3331.
To a stirred solution of 50-O-dimethoxytrityl-20-O-t-butyl-
dimethylsilyl- uridin-30-yl H-phosphonate (413 mg,
0.5 mmol) in DCM (5 mL) containing 2,6-lutidine
(120 lL) pivaloyl chloride (1.2 equiv, 0.6 mmol, 76 lL)
was added. After 5 min, elemental sulfur was added
(3 equiv, 1.5 mmol, 48 mg), optionally followed by triethyl-
amine (10% v/v). After 15 min 1 M TEAB buffer (pH 7.2)
was added and after work-up the organic layer was
collected and evaporated (TLC and 31P NMR analysis
indicated stability of the compound during work-up).
Attempted purification of the mixed anhydride on a
silica-gel column failed due to the decomposition of the
product. TLC (DCM–MeOH, 9:1 v/v): Rf 0.19; (DCM–
14. Stawinski, J.; Kraszewski, A. Acc. Chem. Res. 2002, 35, 952–
960.
15. Sobkowski, M.; Jankowska, J.; Stawinski, J.; Kraszewski, A.
Stereochemistry of internucleotide bond formation by the
H-phosphonate method. 2. Transesterification of aryl
ribonucleoside H-phosphonate diesters with alcohols. 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 183–187.
16. Sobkowski, M.; Jankowska, J.; Stawinski, J.; Kraszewski, A.
Nucleosides Nucleotides Nucleic Acids 2005, 24, 887–890.
17. Jankowska, J.; Sobkowski, M.; Stawinski, J.; Kraszewski, A.
Tetrahedron Lett. 1994, 35, 3355–3358.
MeOH, 8:2 v/v): Rf 0.54 (diastereomers not resolved); 31P
3
0
NMR (121 MHz, DCM): d 50.07 (d, JP;H3 ¼ 10:08 Hz),
3
0
51.23 (d, JP;H3 ¼ 11:91 Hz).
4.7. Adamantanecarboxylic acid, triethylammonium salt
18. Kers, A.; Kers, I.; Stawinski, J.; Sobkowski, M.; Kraszewski,
A. Synthesis 1995, 4, 427–430.
Adamantanecarboxylic acid (9.0 g, 50 mmol) and triethyl-
amine (7.7 mL, 55 mmol) were mixed together in ACN
(100 mL). The solvent was removed under reduced pressure
and the remaining residue was recrystallized from ACN to
yield white crystals (12.8 g; 91%). 1H NMR (300 MHz,
CDCl3): d 1.17 (t, J = 7.30 Hz, 9H, N+CH2CH3), 1.71
(m, 6H, adamantane), 1.90 (m, 6H, adamantane), 2.00
(m, 3H, adamantane), 2.95 (q, J = 7.29 Hz, 6H,
N+CH2CH3), 10.33 (br s, 1H, HN+).
19. Kers, A.; Kers, I.; Stawinski, J.; Sobkowski, M.; Kraszewski,
A. Tetrahedron 1996, 52, 9931–9944.
20. Loschner, T.; Engels, J. Tetrahedron Lett. 1989, 30, 5587–
5590.
21. Engels, J.; Jager, A. Angew. Chem. 1982, 94, 2010–2015.
22. Iyer, R. P.; Yu, D.; Ho, N. H.; Tan, W. T.; Agrawal, S.
Tetrahedron: Asymmetry 1995, 6, 1051–1054.
23. Oka, N.; Wada, T.; Saigo, K. J. Am. Chem. Soc. 2003, 125,
8307–8317.