Synthesis and structural studies of Sp and Rp diastereomers of deoxyxylothymidyl-3′ -O-acetylthymidyl (3′, 5′)-O-(2-cyanoethyl)phosphorothioate in solution and in the solid state

Article abstract of DOI:10.1002/1099-0690(200104)2001:8<1491::AID-EJOC1491>3.0.CO;2-3

The synthesis and full assignment of the molecular structure of the diastereomers of deoxyxylothymidyl-3′-O-acetylthymidyl (3′,5′)-O-(2-cyanoethyl)phosphorothioate (1) in the liquid phase based on ¹H, ¹³C, 1D and 2D homo- and heteronuclear PFG (Pulse Field Gradient) NMR spectroscopic experiments are reported. The pseudorotation parameters of deoxyribose and deoxyxylose were analyzed by means of the PSEUROT program. The absolute configuration R_p and S_p of the phosphorus centers was deduced from ¹H ROESY experiment. The ³¹P CP/MAS technique was used to establish the ratio between the amorphous and crystalline phases, and to test the progress of crystallization. The differences in the ³¹P δ_ii principal elements of the chemical shift tensor, in particular the δ₃₃ parameter, for crystalline and amorphous phases were calculated. Finally, the X-ray data for the S_p diastereomer are reported. The influence of weak C-H···S hydrogen bonding on the molecular packing of both R_p-1 and S_p-1, and the significance of this type of intermolecular interaction for phosphorothioate systems is discussed.

Full text of DOI:10.1002/1099-0690(200104)2001:8<1491::AID-EJOC1491>3.0.CO;2-3

FULL PAPER
Synthesis and Structural Studies of SP and RP Diastereomers of Deoxyxylo-
thymidyl-3Ј-O-acetylthymidyl (3Ј,5Ј)-O-(2-Cyanoethyl)phosphorothioate in
Solution and in the Solid State
Marek J. Potrzebowski,*[d] Xian-Bin Yang,[a] Konrad Misiura,[a] Wieslaw R. Majzner,[b]
[c]
[a]
[a]
´
Michal W. Wieczorek, Slawomir Kazmierski, Sebastian Olejniczak, and
Wojciech J. Stec[a]
1
Keywords: Deoxyxylo dinucleotide / Pseudorotation / H and 13C NMR / 31P CP/MAS / X-ray
The synthesis and full assignment of the molecular structure
between the amorphous and crystalline phases, and to test
of the diastereomers of deoxyxylothymidyl-3Ј-O-acetylthymi- the progress of crystallization. The differences in the 31P δii
dyl (3Ј,5Ј)-O-(2-cyanoethyl)phosphorothioate (1) in the liquid
phase based on 1 H, 13C, 1D and 2D homo- and heteronuclear
principal elements of the chemical shift tensor, in particular
the δ33 parameter, for crystalline and amorphous phases were
PFG (Pulse Field Gradient) NMR spectroscopic experiments calculated. Finally, the X-ray data for the SP diastereomer are
are reported. The pseudorotation parameters of deoxyribose
and deoxyxylose were analyzed by means of the PSEUROT
program. The absolute configuration RP and SP of the phos-
phorus centers was deduced from 1H ROESY experiment.
The 31P CP/MAS technique was used to establish the ratio
reported. The influence of weak C−H···S hydrogen bonding
on the molecular packing of both Rp-1 and SP-1, and the sig-
nificance of this type of intermolecular interaction for phos-
phorothioate systems is discussed.
Introduction
tion process competing with the substitution.[10] Oligonucle-
otides containing incorporated 2Ј-deoxyxylonucleosides
were first prepared by Shabarova,[11] and independently by
Seela et al.[12] Both groups of researchers emphasized lower
affinity of oligonucleotides containing incorporated 2Ј-de-
oxyxylonucleosides towards complementary DNA or
RNA,[13] but their enhanced stability against nucleases.
However, to our best knowledge, no detailed structural
studies on oligodeoxyxylonucleotides have been reported.
This paper shows the power of a multi-technique approach
in the structure elucidation of the stereoisomers of dinucle-
otide analogues in both the liquid and solid phases. Those
assignments are crucial for the unambiguous determination
of the absolute configuration of deoxyxylonucleoside 3Ј,5Ј-
phosphorothioates incorporated in long-chain oligonucleot-
ides.[6]
The utility of NMR spectroscopy for nucleic acid struc-
ture determination has long been recognized.[1] There is a
number of studies showing the application of multi-dimen-
sional NMR spectroscopic techniques to the structure elu-
cidation of nucleotides in the liquid phase.[2] In contrast,
the literature regarding the NMR spectroscopic studies of
nucleotides in the solid state is not as extensive.[3Ϫ5] In this
paper, we report the synthesis, the solution- and solid-state
NMR spectroscopic studies, as well as the X-ray analysis of
RP- and SP-deoxyxylothymidyl-3Ј-O-acetylthymidyl (3Ј,5Ј)-
O-2-(cyanoethyl)phosphorothioate (RP-1 and SP-1) con-
taining β- sugars.[6] Although an early interest in the
chemistry of xylonucleosides was kindled by hopes of their
use as intermediates in the synthesis of oligonucleotides,[7,8]
most efforts were paid to their use in the synthesis of new
therapeutics, such as AZT.[9] 3Ј-O-Activated xylonucleos-
ides, if used as the substrates for oligo(nucleoside phos-
phorothioate)s synthesis, have shown limited application
owing to the low nucleophilicity of phosphorothioate ions
towards the secondary 3Ј-carbon atom and to the elimina-
Results and Discussion
Synthesis
Diastereomers of 1 were synthesized by using the phos-
Polish Academy of Sciences, Center of Molecular and phoramidite method,[14] as depicted in Scheme 1.
[a]
Macromolecular Studies,
Condensation of 5Ј-O-trityldeoxyxylothymidine 3Ј-O-(2-
´
Sienkiewicza 112, 90-363 Lodz, Poland
[b]
[c]
[d]
cyanoethyl)-N,N-diisopropylphosphoramidite (2) with 3Ј-
O-acetylthymidine (3) in the presence of 1-H-tetrazole was
performed in a solution of acetonitrile. The intermediate O-
(2-cyanoethyl)phosphite 4 was not isolated, but treated with
elemental sulfur resulting in the formation of fully pro-
tected dimer 5 as a mixture of two diastereomers in a ratio
of 56:44 (31P NMR determination). The crude diastereo-
Technical University, Institute of Technical Biochemistry,
´
Stefanowskiego 4/10, 90-924 Lodz, Poland
Technical University, Institute of General Food Chemistry,
´
Stefanowskiego 4/10, 90-924 Lodz, Poland
Center for Molecular and Macromolecular Studies, Polish
Academy of Sciences
´
Sienkiewicza 112, 90-362 Lodz, Poland
Fax: (internat.) ϩ 48-42/684-7126
E-mail: marekpot@bilbo.cbmm.lodz.pl
Eur. J. Org. Chem. 2001, 1491Ϫ1501
WILEY-VCH Verlag GmbH, 69451 Weinheim, 2001
1434Ϫ193X/01/0408Ϫ1491 $ 17.50ϩ.50/0
1491

M. J. Potrzebowski et al.
FULL PAPER
Scheme 1
mers were separated and purified by column chromato- investigations presented in this paper fully proved the cor-
graphy through silica gel column. Thus, fast-eluted (FAST- rectness of earlier enzymatic assignments.
5) and slow-eluted (SLOW-5) diastereomers were isolated in
1H, 13C NMR Studies in Solution
37% and 23% yields, respectively. The diastereomeric purity
of FAST-5 and SLOW-5 was confirmed by means of 31P
In this project we were attracted by the prospect of ana-
NMR spectroscopy. Each diastereomer of FAST-5 and lyzing the conformation of pentose rings and elucidating
SLOW-5 was deprotected by treatment with 80% acetic acid the absolute configuration of the phosphorothioyl residue
to provide FAST-1 and SLOW-1, respectively. In our pre- of SP-1 and RP-1. For this purpose, NMR spectroscopy
liminary studies, we tentatively assigned the SP configura- seemed to be the method of choice. In addition to one-
1
tion to FAST-1 and the RP configuration to SLOW-1, by dimensional H and 13C NMR spectra, several two-dimen-
means of enzymatic assays on fully deprotected dimer dias- sional experiments were carried out in order to obtain the
tereomers.[6] The results of NMR spectroscopic and X-ray complete assignment of proton and carbon chemical shifts,
1492
Eur. J. Org. Chem. 2001, 1491Ϫ1501

Diastereomers of Deoxyxylothymidyl-3Ј-O-acetylthymidyl (3Ј,5Ј)-O-(2-Cyanoethyl)phosphorothioate
FULL PAPER
and to unambiguously elucidate the structure of SP-1 and
RP-1. In some experiments, we have taken advantage of the
Pulse Field Gradient (PFG) system in order to reduce the
time of measurement and to improve the quality of the
1
spectra (e.g. reduction of T1 noise). The H and 13C NMR
chemical shifts, as well as proton-proton J coupling con-
1
1
stants established by means of H-1H PFG COSY, H-13C
PFG HMQC, and 1H-13C PFG HMBC experiments for SP-
1 and RP-1 are given in Table 1. The atom-numbering sys-
tem for both diastereomers of 1 is shown in Scheme 2. The
accuracy of δiso and J assignments was confirmed by com-
paring experimental and theoretical spectra obtained by
means of the WINDAISY program.[15]
1
Table 1. H - and 13C NMR spectral data
Scheme 2
those published in the literature, we can conclude that for
both diastereomers, the deoxyribose rings adopt an S-type
conformation (phase angle ca 180°), while deoxyxyloses ad-
opt an N-type conformation (P ca 40°). The accurate
pseudorotational parameters which best fit the experi-
mental data can be found with the computer program PSE-
UROT.[18] The vicinal coupling constants taken from the
WINDAISY simulation were used as input data. For the
deoxyxylose ring, the constants in the Karplus equation
were changed as shown elsewhere.[16] The standard routine
was adopted in our calculations. The puckering amplitudes
for North- and South-type conformers were held constant,
then raised in 1° steps between 30° and 42°, whilst the phase
angle and percentage populations were allowed to vary
freely, until the best solution was found. Five parameters
PN, ΦN, PS, ΦS and fS (sum of fS and fN is equal to 1) fully
characterize the conformations of the pentose rings. From
our calculations, it was established that for the deoxyribose
ring of SP-1, the S-type conformer is the major component
(fS ϭ 0.74 and PS ϭ 142.0; ΦS ϭ 36.5; PN ϭ 50.6, ΦN
ϭ
33.0, RMS ϭ 0.13) while the N-type conformer is the ex-
clusive component for of deoxyxylose ring (fN ϭ 1 and
PN ϭ 29.0; ΦN ϭ 36.0; RMS ϭ 0.36). Very similar values
were obtained for the RP-1 diastereomer. The S-type con-
former is the major component of deoxyribose (fS ϭ 0.82
and PS ϭ 143.7; ΦS ϭ 34.0; PN ϭ 36.3, ΦN ϭ 35.5, RMS ϭ
0.23) while the N-conformer is the only component of de-
oxyxylose (fN ϭ 1 and PN ϭ 30.2; ΦN ϭ 36.0; RMS ϭ
0.28).
Important structural information was also obtained by
analysis of NOE effects, which allow for direct proton-pro-
˚
ton connectivity with distances up to 5 A, through dipolar
Conformation of different pentose rings in nucleosides interactions. For intermediate-sized molecules with an ap-
and nucleotides including, deoxyribose and xylose, was dis- proximate molecular weight of 1 kD, the NOE measure-
cussed in detail by Leeuw and Altona.[16] The relationship ments are often hampered by the vanishing cross-relaxation
between vicinal coupling constants and pseudorotation rates. Since the measurement of transverse cross-relaxation
parameters (phase angle P and puckering amplitude Φm) is (ROE) does not suffer from this problem, the ROESY ex-
known.[17] Comparing the data for SP-1 and RP-1 with periments were performed for diastereomers SP-1 and RP-1
Eur. J. Org. Chem. 2001, 1491Ϫ1501
1493

M. J. Potrzebowski et al.
FULL PAPER
1
Figure 1. 500.13 MHz H NMR ROESY spectra of a) SP-1 and b) RP-1; both spectra recorded in [D6]DMSO; for details see Exp. Sect.
(Figure 1, a, and 1, b, respectively). The dipolar connectiv- other methods to investigate the absolute configuration at
ity between the protons are shown in contour plots. Em- phosphorus in the diastereomers of 1 (vide infra).
ploying the ROESY technique, the chemical shifts of isol-
31P CP/MAS NMR Studies in the Solid State
ated protons, eg. H(6) and H(6Ј) as well as the CH3 groups
of the thymine residues, were assigned unambiguously.
Solid-state NMR spectroscopy is the technique that pro-
Attempts have been made to determine the absolute con- vides a link between NMR spectroscopic data in the solu-
figuration of the phosphorothioyl residue on the basis of tion or liquid phases and results obtained from single-crys-
NOE distance constraints. The O-cyanoethyl group directly tal X-ray or neutron diffraction studies. A comparison of
bonded to phosphorus was considered as a ‘‘marker’’ which the isotropic chemical shifts and further structural results
could help in the assignment of the stereochemistry of that characterize the geometry of molecules in the crystal
phosphorus. Unfortunately, in the case of the SP-1 diastere- lattice offers the possibility of drawing conclusions regard-
omer, the signals for the OCH2 group overlaps with those ing changes in conformation and configuration, as well as
of the 5Ј protons of deoxyribose. For RP-1, the signals for the nature of the intermolecular contacts in both phases.
the OCH2 group overlap partially with those for the 4Ј pro- Since SS NMR uses a powder specimen (i.e. not only one
tons. However, despite these limitations, a detailed analysis selected single crystal) this method can provide immediate
of both spectra in the region of the OCH2 resonances shows information on polymorphs, solvates, inclusion complexes,
some differences. The cross-peak (inside the triangle) be- etc. that may exist in the crystalline state. In this work, we
tween the OCH2 and the 5Ј protons of deoxyxylose seen for have used 31P CP/MAS to optimize the crystallization con-
SP-1 is not present in the contour plot of RP-1. On the basis ditions, indicative of the presence of different crystallo-
of molecular modeling and analysis of distances between graphic modifications, and to establish the ratio of the crys-
the OCH2 group and other centers, it is clear that only for talline to amorphous phases.
the diastereoisomer with an S-configuration at phosphorus
The room-temperature 31P CP/MAS spectra of SP-de-
can the OCH2····H5Ј linkage of deoxyxylose be shorter than oxyxylothymidyl-3Ј-O-acetylthymidyl (3Ј,5Ј)-O-(2-cyanoe-
˚
5 A. Another piece of evidence that supports the SP-config- thyl)phosphorothiate (SP-1) crystallized from a mixture of
uration of FAST-1 is the cross peak between the signals for solvents (see exp. Section) are displayed in Figure 2. The
OCH2 and H(6) of thymine (also inside the triangle). On spectra show a set of spinning sidebands from the large
the other hand, the lack of other cross-peaks predicted for chemical shielding anisotropy (CSA). From inspection of
SP diastereoisomer does not enable us to make a final as- spectrum 2a, it is apparent that the powdered sample of SP-
signment of the absolute configuration at phosphorus by 1 contains crystalline and amorphous phases in an approx-
this approach. These ambiguities prompted us to employ imate ratio of 1:10, represented by sharp and very broad
1494
Eur. J. Org. Chem. 2001, 1491Ϫ1501

Diastereomers of Deoxyxylothymidyl-3Ј-O-acetylthymidyl (3Ј,5Ј)-O-(2-Cyanoethyl)phosphorothioate
FULL PAPER
signals, respectively. As seen from spectrum 2b, the propor- complete elimination of signals from the amorphous phase.
tion of both phases is changed with improved crystallization This spectrum was further used for theoretical calculations.
conditions. In this work, we were interested in the analysis The principal components of the 31P chemical shift tensors
of the principal elements of the chemical shifts of the phos- δii were established from the intensities of the spinning side-
phorothioyl residue of the dinucleotide in the crystalline bands, employing the WINMAS program, based on the
form, and in the correlation between δii parameters and mo- BergerϪHerzfeld algorithm.[19,20] The calculated values of
lecular geometry. For this purpose, a good quality spectrum the principal tensor elements δii and of the shielding para-
of the crystalline phase was required. Taking advantage of meters are given in Table 2. The accuracy of the calculations
the differences between the 31P T2 relaxation times of the was confirmed by a comparison of the experimental and
crystalline and amorphous phases, the 31P CP/MAS experi- theoretical spectra (Figure 3).
ment was performed with a ‘‘dead time delay’’ of 370 µs
(one over rotor period at spinning rate 2700 Hz). Figure 2
(c) presents the spectrum of sample SP-1 with an almost
Figure 3. a) 121.49.46 MHz 1H-31P CP/MAS experimental spec-
trum of SP-1 recorded under the conditions given in Figure 2c; b)
simulated spectrum calculated employing the WINϪMAS pro-
gram, version 940108, BrukerϪFranzen Analytik
Figure 2. 121.49.46 MHz H-31P CP/MAS experimental spectra of
1
In contrast, attempts at the crystallization of the RP-1
diastereomer provided only an amorphous powder. Pains-
taking attempts to obtain crystals of RP-1 failed. Thus the
δii elements were only established for the amorphous form.
SP-1; the spectra have 4 K data points with 10 Hz line broadening,
a contact time of 1 ms, 0.5 K scans and νrot ϭ 2.7 kHz; a) spectrum
recorded after first crystallization; b) spectrum after second crystal-
lization with a changed proportion of solvents in the mixture; c)
spectrum b recorded with a dead time delay of 1/νrot
Table 2. 31P chemical shifts and NMR parameters for SP-1 and RP-1
Compound[a]
δ11(ppm)
δ22(ppm)
δ33(ppm)
∆δ(ppm)
Ω(ppm)
η
κ
δiso
(ppm)
SP-1
152
151
154
131
128
130
Ϫ81
Ϫ91
Ϫ92
223
230
233
233
242
246
0.14
0.20
0.16
0.81
0.81
0.85
67.8
62.6
63.6
(crystalline phase)
SP-1
(amorphous phase)
RP-1
(amorphous phase)
[a]
Estimated errors in δ11, δ22, δ33 and ∆δ are Ϫ/ϩ5 ppm; errors in δiso are Ϫ/ϩ0.2 ppm. The principal components of the chemical shift
tensor are defined as follows; δ11 Ͼ δ22Ͼ δ33. The isotropic chemical shift is given by δiso ϭ (δ11ϩδ22ϩδ33)/3
Eur. J. Org. Chem. 2001, 1491Ϫ1501
1495

M. J. Potrzebowski et al.
FULL PAPER
The experimental and theoretical spectra are shown in Fig- logues of dinucleotides. Crystallographic data and experi-
ure 4, and calculated parameters are collected in Table 2. mental details for SP-1 are listed in Table 3. Atomic coord-
The very detailed analysis of δii shows several structural inates, bond lengths and angles, and torsional angles have
features of SP-1 and RP-1. It is interesting to note the con- been deposited (CCDC 147326). The thermal ellipsoid plots
siderable differences between the isotropic values of the and the numbering schemes of SP-1 are shown in Figure 5.
crystalline and of the amorphous phases (ഠ 5 ppm). For all
modifications, only small changes in the δ11 and δ22 para-
meters are seen. More significant differences are established
Table 3. Crystal data and experimental details for SP-1
for δ33 and cover the range up to 10 ppm. The similar be-
Molecular formula
Molecular mass
C25H32N5O12PS
657.59
havior of phosphorothioyl compounds, and the origin of
δ33 changes were exhaustively discussed in a recent paper.[21]
We have concluded that the effect is related to the formation
of weak PϭS····HϪC hydrogen bonds and to the decrease
in the electronic shielding of the phosphorus atom, as the
phosphorothioyl group becomes more polarized during
bond formation. Similar conclusions also hold for samples
RP-1 and SP-1.
Crystallographic system
Space group
Monoclinic
P21
˚
a (A)
9.662(2)
14.279(5)
11.296(3)
94.10(3)
1554.4(8)
2
˚
b (A)
˚
c (A)
β (°)
3
˚
V (A )
Z
Dcalcd. (g/cm3)
1.405
20.11
µ [cmϪ1
]
Crystal dimensions (mm)
0.2 ϫ 0.3 ϫ 0.8
150
Maximum 2θ (°)
˚
Radiation, λ (A)
Cu-Kα, 1.54184
ω/2θ
Scan mode
Scan width (°)
hkl ranges:
0.70 ϩ 0.14.tanθ
h ϭ Ϫ12/12
k ϭ Ϫ17 17
l ϭ 0/14
0.9066
EAC correction
min.
max.
ave.
0.9996
0.9643
No. of reflections:
unique
6379
refine with I Ͼ 2σ(I)
observed with I Ͼ 2σ(I)
No. of parameters refined
6279
6053
537
Ϫ3
˚
Largest diff. peak (eA
)
0.344
Ϫ3
˚
Largest diff. hole (eA
)
Ϫ0.158
Robs
wRobs
0.0355
0.0911
weighting coeff.* m/n
extinction coeff.** k
Sobs
0.0706/0.0467
0.0087(5)
1.063
shift/esd max
Absolute structure
0.001
SP1, RC1Ј, SC3Ј, RC4Ј
,
RC1Јa, RC3Јa, RC4Јa
Ϫ0.017(12)
0.0563
,
Flack parameter χ
Rint
Tmeas.
F(000)
293(2)
688
* weighting scheme w ϭ [ σ2(Fo2)ϩ(m*P)2ϩn*P]Ϫ1 where P ϭ
[max(0,Fo2) ϩ 2Fc2]/3
** extinction method SHELXL, extinction expression Fc* ϭ kFc[1
ϩ 0.001xFc2λ3/sin(2θ)]Ϫ1/4
Crystallographic data (excluding structure factors) for
the structure(s) reported in this paper have been deposited
with the Cambridge Crystallographic Data Centre as
supplementary publication no. CCDC-147326 Copies
of the data can be obtained free of charge on
application to CCDC, 12 Union Road, Cambridge CB2
1EZ, UK [Fax: (internat.) ϩ 44Ϫ1223/336Ϫ033; E-mail:
deposit@ccdc.cam.ac.uk].
Figure 4. a) 121.49.46 MHz 1H-31P CP/MAS experimental spec-
trum of RP-1; the spectrum has 4 K data points with 10 Hz line
broadening, a contact time of 1 ms, 0.5 K scans and νrot ϭ 2.7 kHz;
b) simulated spectrum of RP-1 calculated employing the WIN-
MAS program, version 940108, BrukerϪFranzen Analytik
X-ray Crystallography Studies
The results obtained from solid-state NMR showing that
Compound SP-1 crystallizes in a monoclinic system in
diastereomer SP-1 has
a tendency to form crystals space group P21 with the unit cell consisting of two molec-
prompted us to grow single crystals of suitable quality for ules of the dinucleotide. One of the molecules is an asym-
X-ray studies. To the best of our knowledge, there are no metric unit of the crystal. This observation is consistent
literature reports of X-ray studies of phosphorothioate ana- with the 31P CP/MAS experiment, in which a single line
1496
Eur. J. Org. Chem. 2001, 1491Ϫ1501

Diastereomers of Deoxyxylothymidyl-3Ј-O-acetylthymidyl (3Ј,5Ј)-O-(2-Cyanoethyl)phosphorothioate
FULL PAPER
Figure 5. Molecular structure of SP-1 showing 50% probability displacement ellipsoids; atoms with σof ϭ 0.26 are omitted for clarity
in the isotropic part of spectrum was detected. Molecular tion of N(3)ϪH as donors and C(4)ϭO(4) as acceptors;
packing is shown in Figure 6. As seen, both molecules are and ii) Hydrogen bonding between the acyl block of deoxy-
bonded through intermolecular interactions; i) Symmet- ribose and the hydroxyl group at C5 of deoxyxylose [C(31)-
rical hydrogen bonding between thymines with participa- O(32)·····H(6Јa)ϪC(5Јa)]. The values of inter- and intramo-
Figure 6. Crystal packing diagram for SP-1; atoms with σof ϭ 0.26 are omitted for clarity
Eur. J. Org. Chem. 2001, 1491Ϫ1501
1497

M. J. Potrzebowski et al.
FULL PAPER
lecular hydrogen bond lengths and bond angles presented Table 5 shows several intra- and intermolecular Pϭ
in Table 4 are within the expected range for hydrogen bonds S·····HϪC interactions. The most important intermolecular
of this type.
contacts seem to be interaction of sulfur with: i) The CH3
group of thymine, ii) The CH3 group of the acyl block, and
iii) The C(4)ϪH of deoxyribose.
Table 4. CϪH· · ·O, NϪH· · ·O and CϪH· · ·N hydrogen-bonding
˚
geometry (A,°) for SP-1
Distances and angles for deoxyribose and deoxyxylose
are typical for this class of compounds. Endocyclic torsion
angles and sugar thiophosphate backbone angles are shown
in Table 6. A comparison of the γ2 torsion angle for both
pentoses reveals their similar absolute values and opposite
signs. By definition, the N-type conformations are charac-
terized by positive values of C1ϪC2ϪC3ϪC4-; in the S-
type, this torsion angle has a negative value.
DϪH· · ·A
DϪH
H· · ·A D· · ·A
DϪH· · ·A
C1ЈϪH1Ј· · ·O2
C2ЈϪH2Ј1· · ·N1
C2ЈϪH2Ј1· · ·C6
C2ЈϪH2Ј2· · ·O4i
C3ЈϪH3Ј· · ·O5Ј
C3ЈϪH3Ј· · ·O32
0.98(2)
0.90(4)
0.90(4)
0.96(3)
1.01(3)
1.01(3)
2.37(2) 2.765(2)
2.47(4) 2.502(3)
2.75(4) 3.157(3)
2.66(3) 3.442(3)
2.78(2) 2.932(2)
2.32(3) 2.654(3)
2.77(3) 3.581(3)
2.04(2) 2.860(2)
2.42(2) 2.742(3)
2.69(3) 2.826(3)
2.36(3) 2.688(2)
2.53(3) 2.822(3)
1.71(4) 2.737(3)
1.99(3) 2.747(2)
2.41(3) 2.732(3)
2.52(4) 3.365(4)
2.75(4) 2.877(3)
2.70(4) 2.877(3)
2.70(4) 3.262(4)
2.70(6) 3.577(4)
2.71(5) 3.464(4)
2.68(5) 3.450(5)
103(2)
82(2)
109(3)
139(2)
89(2)
98(2)
C5ЈϪH5Ј1· · ·O5Јaii 0.94(3)
145(2)
178(2)
99(2)
N3ϪH3· · ·O4aiii
C6ϪH6· · ·O4Ј
0.82(2)
0.95(2)
0.98(4)
0.94(3)
Table 6. a) Torsion angles of the endocyclic sugars; b) torsion
angles for the sugar phosphate backbone for the SP-1 diastereomer
C7ϪH73· · ·O4
C1ЈaϪH1Јa· · ·O2a
88(2)
100(2)
96(2)
C5ЈaϪH5Јa· · ·O3Јa 1.02(3)
a)
Torsion angles
Definition
C5ЈaϪH6Јa· · ·O32iv 1.03(4)
173(3)
170(4)
101(2)
147(3)
88(3)
N3aϪH3a· · ·O4i
C6aϪH6a· · ·O4Јa
0.77(3)
0.92(3)
'Deoxyribose Deoxyxylose
C7aϪH72a· · ·O5Јaii 0.96(4)
C7aϪH72a· · ·O4a
C7aϪH73a· · ·O4a
C32ϪH323· · ·O2ai
C12ϪH121· · ·O4v
C13ϪH131· · ·O2vi
C13ϪH132· · ·O5Јa
0.96(4)
1.03(5)
1.03(5)
1.18(6)
0.87(5)
0.82(5)
C4ϪO4ϪC1ϪC2
O4ϪC1ϪC2ϪC4
C1ϪC2ϪC3ϪC4
C2ϪC3ϪC4ϪO4
C3ϪC4ϪO4ϪC1
Ϫ28.6
37.7
6.3
89(3)
γ0
γ1
γ2
γ3
γ4
115(3)
130(4)
145(4)
157(4)
Ϫ25.8
33.7
Ϫ32.2
16.3
Ϫ30.7
15.6
7.7
C13‘‘ϪH133· · ·O2vi 1.13(10) 2.42(9) 3.419(13) 147(7)
b)
Torsion angles
Symmetry codes: (i) 1 Ϫ x, Ϫ0.5 ϩ y, 2 Ϫ z; (ii) Ϫ x, Ϫ0.5 ϩ y, 1
Ϫ z; (iii) 1 Ϫ x, 0.5 ϩ y, 2 Ϫ z; (iv) Ϫ x, 0.5 ϩ y, 1 Ϫ z; (v) 1 Ϫ
x, Ϫ0.5 ϩ y, 1 Ϫ z; (vi) Ϫ1 ϩ x, y, Ϫ1 ϩ z.
Definition
A DNA[a]
B DNA[a]
α
β
γ
δ
ε
ξ
O3ЈϪPϪO5ЈϪC5Ј
PϪO5ЈϪC5ЈϪO3Ј
164.95
150.22
Ϫ50
172
41
Ϫ46
Ϫ147
36
Our attention has been focused on possible PϭS·····HϪC
intermolecular contacts which were suggested in the previ-
ous section. The role of CH····S hydrogen bonding and the
influence of weak contacts on the molecular packing of or-
O5ЈϪC5ЈϪC4ЈϪC3Ј 47.60
C5ЈϪC4ЈϪC3ЈϪO3Ј Ϫ35.70
79
157
C4ЈϪC3ЈϪO3ЈϪP
C3ЈϪO3ЈϪPϪO5Ј
161.09
106.83
Ϫ146
Ϫ78
155
Ϫ96
χxylo O4ЈϪC1ЈϪN1ϪC2 Ϫ159.94
ganic crystals was very recently discussed by Borrmann et χribo O4ЈϪC1ЈϪN1ϪC2 Ϫ137.50
Ϫ154
Ϫ98
al.[22] As reported by Potrzebowski and co-workers, such
[a] Average values of the backbone torsion angles in polynucleotides
(taken from ref.2).
interactions are not unusual for organophosphorothioyl
compounds.[21] The geometric criteria for these interactions
were discussed elsewhere.[21,22] The nonlinear character of
such weak hydrogen bonds is proved. According to the liter-
Finally, the absolute S -configuration of the chiral phos-
ature, we have considered contacts within the range of the phorus center was established by the Flack method [with
˚
C···S distance 3.55Ϫ4.10 A and CϪH···S angles in the χ ϭ Ϫ0.017(12)]. It should be emphasized that this result
range 80.0Ϫ180.0°. Inspection of X-ray data given in is consistent with data obtained from NMR spectroscopic
Table 5. Possible CϪH· · ·S hydrogen bonds for SP-1
XϪH· · ·Y
XϪH
H· · ·Y
X· · ·Y
(XϪH· · ·Y)
Symmetry op.
C4ЈϪH4Ј S1
0.964(24)
0.980(30)
0.953(24)
1.016(42)
0.940(30)
0.839(26)
0.997(33)
0.959(43)
1.034(45)
1.027(47)
0.705(129)
3.167(24)
3.091(30)
3.146(24)
3.021(44)
3.734(28)
2.989(25)
3.533(34)
3.455(41)
3.420(43)
3.422(44)
3.096(49)
4.083(2)
3.528(2)
3.948(2)
3.928(3)
4.047(2)
3.332(2)
3.672(3)
3.672(3)
3.672(3)
3.864(4)
3.393(4)
3.393(4)
159.2(18)
108.6(19)
142.8(18)
149.1(33)
102.8(20)
107.0(19)
90.0(20)
1Ϫx,Ϫ0.5 y, 1 Ϫ z
C5ЈϪH5Ј2 S1
C6ϪH6 S1
C7ϪH71 S1
C2ЈAϪH2ЈA S1
C3ЈAϪH3ЈA S1
C7AϪH71A S1
C7AϪH72A S1
C7AϪH73A S1
C32ϪH323 S1
C12ϪH124 S1
C12ϪH122 S1
Ϫx,Ϫ0.5 ϩ y, 1 Ϫ z
Ϫx, Ϫ0.5 ϩ y, 1 Ϫ z
Ϫx,Ϫ0.5 ϩ y, 1 Ϫ z
1 Ϫ x, Ϫ0.5 ϩ y, 1 Ϫ z
95.4(26)
95.8(26)
108.0(29)
109.2(67)
100.7(34)
1.057(57)
3.034(52)
˚
˚
Criteria: XϪH ϭ 0.69Ϫ1.37(A),H· · ·Y ϭ 1.18Ϫ4.00 (A), (XϪH· · ·Y) ϭ 80.0Ϫ180.0°
1498
Eur. J. Org. Chem. 2001, 1491Ϫ1501

Diastereomers of Deoxyxylothymidyl-3Ј-O-acetylthymidyl (3Ј,5Ј)-O-(2-Cyanoethyl)phosphorothioate
FULL PAPER
studies in solution (vide supra) and with results of our earl-
ier studies on stereodifferentiation in the enzymatic de-
Experimental Section
gradation of SP- and RP-deoxyxylothymidylthymidyl
Syntheses: N,N-Diisopropylethylamine, dichloromethane and ace-
tonitrile were dried by distillation from calcium hydride. The pro-
gress of each reaction was monitored by TLC.
(3Ј,5Ј)phosphorothioates.[6]
5Ј-O-Trityldeoxyxylothymidine 3Ј-O-[N,N-Diisopropyl-O-(2-cyano-
ethyl)phosphoramidite] (2): To a stirred solution of dry 5Ј-O-trityl-
deoxyxylothymidine[24] (485 mg, 1 mmol) and anhydrous N,N-diiso-
propylethylamine (700 µL, 4 mmol) in dry dichloromethane (5 mL)
under an argon atmosphere at ambient temperature was added
N,N-diisopropyl-O-(2-cyanoethyl)phosphoramidochloridite (335
µL, 1.5 mmol) by syringe over a period of 2 min. The obtained
Conclusions
In this work we have presented complete assignment of
the structure of SP-deoxyxylothymidyl-3Ј-O-acetylthymidyl
(3Ј,5Ј)-O-(2-cyanoethyl)phosphorothioate (SP-1) by X-ray
crystallography and NMR spectroscopy in solution and in
the solid state. To the best of our knowledge, it is the first solution was stirred for 3.0 h. Methanol (0.1 mL) was then added,
and stirring was continued for 10 min. The solvent was removed by
evaporation. Ethyl acetate (200 mL) was added to the residue and
the solution was washed with ice-cold aqueous sodium bicarbonate
(10%, 2 ϫ 30 mL) and brine (2 ϫ 20 mL). The organic phase was
dried with anhydrous MgSO4 and evaporated to dryness. The crude
product was purified by column chromatography on silica gel (silica
gel: 230Ϫ400 mesh, 40 g, eluted with CH2Cl2/AcOEt/Et3N, 90:8:2
v/v/v, to provide the title compound (482 mg, yield 71%): Rf 0.43
(chloroform/ethanol ϭ 19:1, v/v). Ϫ 31P NMR (CDCl3) ϭ 151.9
(49%), 148.5 (51%) (ref.[25] 151.8, 148.5 for dimethoxytrityl ana-
report showing such results for this class of compounds.
The conformational analysis in liquid and in the solid state,
and the resemblance of the structures in both phases is of-
ten a matter of debate and controversy. Altona and Sundar-
alingman analyzed large sets of nucleotides and nucleosides,
and have shown that in the case of ribose rings, the con-
formational preferences in the solid state are generally very
similar to those established in solution.[23] It has been fur-
ther concluded that intramolecular forces play a decisive
role in the determination of the geometry of the ribose logue).
rings. However, this is not the case with compound SP-1.
SP- and RP-5Ј-O-Trityldeoxyxylothymidyl-3Ј-O-acetylthymidyl
(3Ј,5Ј)-O-(2-Cyanoethyl)phosphorothioate (SP-5 and RP-5): To
The phase parameters established on the basis of the X-ray
data of SP-1 were found to be 79.7° for the deoxyxylose
ring, and 221.7° for the deoxyribose ring. The P values in
the solid phase are considerably larger than those obtained
from measurements in solution (30° for the deoxyxylose
ring, fN ϭ 1, and 142° for the deoxyribose ring, fS ϭ 0.74).
Instead, these results suggest that in case of 1, the intermol-
ecular contacts and the crystal packing effects are the dom-
inating factors that determinate the conformation of the
sugar rings in the solid state. The number of intermolecular
hydrogen bonds found in the crystal lattice of SP-1 is shown
in Table 4 and 5. The another point is the geometry of the
backbone of SP-1, which is characterized by several torsion
angles as shown in Table 6b. This table also contains the
average values of the backbone torsion angles for A DNA
and for B DNA. From the analysis of this data, it is appar-
ent that a modification of the deoxyxylose has a dramatic
effect on the geometry of the whole molecule. This differ-
ence can be one of the reasons for the lower affinity of
oligonucleotides containing xylo nucleosides to single-
stranded DNA or RNA than that of unmodified oligonu-
cleotides.[13]
a
stirred solution of 2 (431 mg, 0.5 mmol) and 1H-tetrazole (42 mg,
0.6 mmol) in anhydrous acetonitrile (1.0 mL) at room temperature
under an argon atmosphere, was added 3Ј-O-acetylthymidine
(114 mg, 0.4 mmol). Stirring was continued for 2 h, and elemental
sulfur (32 mg, 1 mmol) was added. To this solution was added
dichloromethane (2 mL) and stirring was continued overnight. Sul-
fur was removed by filtration. The solvent was removed by evapora-
tion, and acetonitrile (50 mL) was added to the residue. The precip-
itated sulfur was removed by filtration, and the solvent was evapor-
ated to dryness. The residue was dissolved in chloroform (50 mL)
and the resulting solution was washed with NaHCO3 (10%,
15 mL), brine (15 mL), and finally with water (2 ϫ 15 mL). The
CHCl3 layer was dried over anhydrous MgSO4, the solvent was
evaporated to dryness. Crude product (530 mg) was purified by col-
umn chromatography on silica gel [silica gel: 230Ϫ400 mesh, 45 g,
eluted with CH2Cl2 (100 mL), then a gradient of 0Ϫ2% methanol
in chloroform] to provide the separated diastereomers as follows:
(1) Fast-eluted diastereomer SP-5 (105 mg): Rf 0.49 (chloroform/
1
ethanol, 19:1, v/v). Ϫ H NMR (CDCl3): δ ϭ 1.76 (s, 3 H), 1.90
(s, 3 H), 2.07 (s, 3 H), 2.1Ϫ5.3 (m, 18 H), 6.1Ϫ6.4 (m, 2 H), 7.0Ϫ8.0
(m, 16 H), 9.89 (br. s, 1 H), 9.94 (br. s, 1 H). Ϫ δP (CDCl3) 67.6.
Ϫ MS: FAB, -ve, m/z ϭ 899 [M Ϫ 1]Ϫ. Ϫ C44H47N5O12PS: calcd.
C 58.66, H 5.26, N 7.77; found C 58.27, H 5.62, N 7.46; (2) A
mixture of diastereomers (62 mg); and (3) Slow-eluted diastereomer
Furthermore, comparing the NMR spectroscopic data in
1
1
the solution for SP-1 and RP-1, we found a very similar H
RP-5, 54 mg: Rf 0.45 (chloroform, ethanol 19:1, v/v). Ϫ H NMR
and 13C data set for both diastereomers. Thus we assume
that a spatial arrangement of the major segments of the
dinucleotide (with the evident exception of the PϭS unit) is
very similar for both compounds. Finally, the intra- and
intermolecular hydrogen bonds seem to play a crucial role
in the crystallization process. The importance of weak Pϭ
S····HϪC hydrogen bonds as one of the factors that deter-
mine the molecular packing is an open question. However,
in our opinion, these weak interactions should not be neg-
lected.
(CDCl3): δ ϭ 1.75 (s, 3 H), 1.86 (s, 3 H), 2.10 (s, 3 H), 2.1Ϫ5.3 (m,
18 H), 6.1Ϫ6.4 (m, 2 H), 7.0Ϫ7.6 (m, 16 H), 9.56 (br. s, 1 H), 9.62
(br. s, 1 H). Ϫ 31P NMR (CDCl3): δ ϭ 66.7. Ϫ MS: FAB, -ve, m/
z ϭ 899 [M Ϫ 1]Ϫ. Ϫ C44H47N5O12PS: calcd. C 58.66, H 5.26, N
7.77; found C 58.49, H 5.52, N 7.99. The total yield of 5 was 62%.
SP- and RP-Deoxyxylothymidyl-3Ј-O-acetylthymidyl (3Ј,5Ј)-O-(2-
Cyanoethyl)phosphorothioate (SP-1 and RP-1): Fast-eluted diastere-
omer SP-5 (100 mg) was treated with acetic acid (3 mL, 80%) at
room temperature for 3 d. The solvent was removed by evapora-
tion. The residue was purified by column chromatography (silica
Eur. J. Org. Chem. 2001, 1491Ϫ1501
1499

M. J. Potrzebowski et al.
FULL PAPER
gel, 230Ϫ400 mesh, 5.0 g: eluted with a gradient of 0Ϫ10% meth-
anol in chloroform to provide SP-1 (55 mg, 72%): Rf 0.22 (chloro-
form, ethanol 19:1, v/v). Ϫ δP 67.7. Ϫ MS: FAB, ϩve, m/z ϭ 658
bis(dineopentoxyphosphorothioyl) disulfide set at 84.0 ppm. The
principal elements of the 31P chemical shift tensor and shielding
parameters were calculated employing the WINMAS program. The
[M ϩ 1]ϩ. Ϫ C25H32N5O12PS: calcd. C 45.66, H 4.90, N 10.65, S details describing the method and accuracy of calculations are dis-
4.88; found C 45.58, H 4.90 N 10.34 , S 4.65. SP-1 was crystallized cussed elsewhere in detail.[19,20]
from ethanol/chloroform/ hexane (1:1:1, v/v/v) to provide crystals,
X-ray Analysis: The crystal and molecular structure of SP-1 was
m.p. 189Ϫ190° C.
determined using data collected at room temperature on a CAD4
diffractometer with graphite monochromated Cu-Kα radiation.[26]
The compound crystallizes in a monoclinic system, in the space
group P21, with the unit cell consisting of two molecules. The lat-
tice constants were refined by a least-squares fit of 25 reflections
in the θ range 19.01°Ϫ28.87°. The decline in the intensity of three
control reflections (3, Ϫ5, 4; 1, Ϫ7, Ϫ2; Ϫ3, Ϫ4, 1) was 1.7% dur-
ing 94.0 h of exposure time. An empirical absorption correction
Starting from slow-eluted diastereomer RP-5 and using the same
procedure, RP-1 was obtained in 68% yield: Rf 0.18 (chloroform,
ethanol 19:1, v/v). Ϫ 31P NMR: δ ϭ 66.8. Ϫ C25H32N5O12PS:
calcd. C 45.66, H 4.90, N 10.65, S 4.88; found C 45.36, H 4.83, N
10.30, S 4.55. Attempts to crystallize RP-1 from several solvents
failed.
NMR Measurements in Solutions: The samples of SP-1 or RP-1
(5 mg) were dissolved in [D6]DMSO (0.5 mL). All spectra were re-
corded on a Bruker Avance DRX 500 spectrometer, operating at
500.1300 MHz for 1H, 125.2578 MHz for 13C, and 202.46 MHz for
31P. For all experiments, original Bruker pulse programs were used.
The chemical shift of the DMSO signal was used as a reference
(δ ϭ 2.49 for 1H and δ ϭ 39.5 for 13C). Phosphoric acid (85%) was
used as an external standard for 31P spectra. The spectrometer was
equipped with a Pulse Field Gradient Unit (50 G/cm). The inverse
broadband probe-head was used. The COSY90 spectra were ob-
tained from 1024 experiments, each with 4 scans. The relaxation
delay was 1.5 sec. The spectral width was 10 ppm (5000 Hz) in both
dimensions. The data size in F2 was 4 K. Digital quadrature detec-
tion (DQD) was applied. Two 10 µs length z-gradient pulses, with
a strength of approximately 5 G/cm each, were applied with a 1-ms
delay for gradient recovery. The FID were apodized with a sine-
bell function in both dimensions. Final data were zero filled twice
in both dimensions and symmetrized about the diagonal. The
ROESY spectra were recorded in a 2 K ϫ 1 K (F2 ϫ F1) data
matrix. Digital quadrature detection was applied and 32 scans were
accumulated in each experiment. The experiment was run in the
phase-sensitive mode with a 3650 ms cw pulse for ROESY spin
lock. The spectral width was 4500 Hz (9 ppm) in both dimensions.
Data were processed with a sine bell shaped apodization function
in both directions and TPPI in F1. No zero filling was applied. The
PFG-HMQC spectra were acquired in a 1 K ϫ 4 K [F1(13C) ϫ
F2(1 H)] data matrix. Three 1 ms length z-gradient pulses, with
strengths of approximately 25 G/cm, 15 G/cm, and 20 G/cm, in se-
quence were applied with 1 ms delay for gradient recovery. The
spectral width was 4000 Hz (8 ppm) in F2 (1 H) and 25 kHz
(200 ppm) in F1 (13C). A Garp decoupling sequence was employed.
Final data were processed with a sine function in F1 and qsine in
the F2 dimension. The PFG-HMBC experiment was acquired in
0.5 K ϫ 4 K (F1 ϫ F2) data matrix and 8 scans for each experi-
ment. Three 1-ms length z-gradient pulses, strength of about 25 G/
cm, 15 G/cm and 20 G/cm, in sequence were applied with 50 µs
delay for gradient recovery. Final data were processed with sine-
bell and qsine-bell functions in F1 and F2 dimensions respectively.
was applied by the use of the ψ-scan method (EAC program).[27]
A
total of 6279 reflections with I Ն 2σ(I) were used to solve the struc-
ture by direct methods and to refine it by full-matrix least-squares
using
F2.
Hydrogen
atoms
at
the
disordered
ϪOϪCH2ϪCH2ϪCϵN group with gof ϭ 0.26(2) were placed
geometrically, and allowed to ride on the preceding C atom with a
CϪH distance free to refine and fixed thermal parameters equal to
1.3 times the equivalent isotropic thermal parameter of the parent-
atom. All other hydrogen atoms were found on a difference-Fourier
map and refined isotropically. Anisotropic thermal parameters were
applied for all non-hydrogen atoms. The final refinement converged
to R ϭ 0.0355 for 6053 reflections with I Ն 2σ(I). The absolute
configuration at the chiral atoms were established as SP1, RC1Ј, SC3Ј
,
RC4Ј, RC1Јa, RC3Јa, RC4Јa. The absolute structure was determined by
the Flack method,[27] with the resulting χ ϭ Ϫ0.017(12).
Data Collection and Cell Refinement: Absorption correction:
EAC.[28] Structure solution: SHELXS-86.[29] Structure refinement:
SHELXL-93.[30]
Acknowledgments
This project was financially assisted by the State Committee for
Scientific Research, grants 3 T09A 026 19 (to M.J.P.) and 4 P05F
006 17 (to W.J.S.).
[1]
K. Wüthrich, NMR of Proteins and Nucleic Acids, Wiley, New
York, 1986.
[2]
S. S. Wijmega, M. M. W. Mooren, C. W. Hilbers, chapter NMR
of Nucleic Acids; From Spectrum to Structure in NMR of Mac-
romolecules, A Practical Approach, (Ed.: G. C. K. Roberts),
IRL Press, Oxford, 1993, pp. 217Ϫ289.
[3]
H. Shindo in Phosphorus-31 NMR: Principles and Applications,
Ed. D. G. Gorenstein, Academic Press, New York, 1984, pp.
401Ϫ422.
[4]
C. W. Hilbers, S. S. Wijmega, in Encyclopedia of Nuclear Mag-
netic Resonance, (Eds.: D. M. Grant, R. K. Harris), Wiley,
Chichester, 1996, pp. 3346Ϫ3359.
[5]
P. M. Macdonald, M. J. Damba, K. Ganeshan, R. Braich, S.
V. Zabarylo, Nucleic Acid Res. 1996, 24, 2868Ϫ2876 and refer-
NMR Measurements in the Solid State: Cross-polarization magic
angle spinning solid-state 31P NMR spectra were recorded on a
Bruker 300 MSL instrument with high-power proton decoupling at
121.46 MHz for 31P. The 31P CP/MAS NMR spectra were recorded
in the presence of high-power proton decoupling. Powder samples
of SP-1 and RP-1 were placed in a cylindrical rotor and spun at
ences cited therein.
[6]
X.-B. Yang, K. Misiura, M. Sochacki, W. J. Stec, Bioorg. Med.
Chem. Lett. 1997, 7, 2651Ϫ2656.
J. Zemlicka, J. Smrt, Tetrahedron Lett. 1964, 5, 2081Ϫ2086;
Y. Mizuno, T. Sasaki, Tetrahedron Lett. 1965, 6, 4579Ϫ4584.
[7]
[8]
[9]
H. Mitsuya, K. J. Weinhold, P. A. Furman, M.H. StClair, S.N.
1
Lehrman, R.C. Gallo, D. Bolognesi, D. W. Barry, S. Broder,
2.0Ϫ4.5 kHz. The field strength for H decoupling was 1.05 mT, a
Proc. Natl. Acad. Sci. USA 1985, 82, 7096Ϫ7010.
contact time of 5ms, a repetition of 6s and a spectral width of
50 kHz were used, and 8 K data points represented the FID. Spec-
tra were accumulated 500 times, and gave a reasonable signal-to-
noise ratio. 31P chemical shifts were calibrated indirectly through
[10]
X.-B. Yang, K. Misiura, W. J. Stec, Heteroatom Chem. 1999,
10, 91Ϫ104.
N. I. Sokolova, N. G. Dolinnaya, N. F. Krynetskaya, Z. A.
Shabarova, Nucleosides Nucleotides 1990, 9, 515Ϫ531.
[11]
1500
Eur. J. Org. Chem. 2001, 1491Ϫ1501

Diastereomers of Deoxyxylothymidyl-3Ј-O-acetylthymidyl (3Ј,5Ј)-O-(2-Cyanoethyl)phosphorothioate
FULL PAPER
[12]
H. Rosemeyer, M. Krecmerova, F. Seela, Helv. Chim. Acta
1991, 74, 2054Ϫ2067.
ski, T. Lis, J. Pluskowski, W. Ciesielski, J. Org. Chem.1998,
63, 4209Ϫ4217
[13]
[22]
[23]
H. Rosemeyer, F. Seela, Nucleosides Nucleotides 1995, 14,
H. Borrmann, I. Persson, M. Sandström, C. M. V. Stälhandske,
J. Chem. Soc., Perkin Trans. 2 2000, 393Ϫ402.
C. Altona, M. Sundaralingam, J. Am. Chem. Soc. 1972, 94,
8205Ϫ8212.
J. J. Fox, N. C. Miller, J. Org. Chem. 1963, 28, 936Ϫ941.
H. Rosenmeyer, F. Seela, Helv. Chim. Acta 1991, 74, 749Ϫ760.
B. A. Frenz, EnrafϪNonius Structure Determination Package;
SDP User’s Guide. Version of 17 December 1986.
EnrafϪNonius, Delft, The Netherlands 1986.
H. D. Flack, Acta Cryst. 1983, A39, 876Ϫ881.
A. C. T. North, D. C. Phillips, F. S. Mathews, Acta Cryst. 1968,
A24, 351Ϫ359.
1041Ϫ1045.
[14]
[15]
M. H. Caruthers, Science 1985, 230, 281Ϫ285.
WINDAISY program, version 940108, BrukerϪFranzen Ana-
lytik, Bremen, 1994.
F. A. A. M. de Leeuw, C. Altona, J. Chem. Soc., Perkin Trans.
1I 1982, 375Ϫ384.
F. A. A. M. de Leeuw, C. Altona, J. Comput. Chem. 1983,
4, 438Ϫ441.
C. A. G. Haasnoot, F. A. A. M. de Leeuw, H. P. M. de Leeuw,
C. Altona, Org. Magn. Reson. 1981, 15, 43Ϫ52.
G. Jeschke, G. Grossmann, J. Magn. Reson. 1993, A103,
323Ϫ328.
[24]
[25]
[26]
[16]
[17]
[18]
[19]
[20]
[21]
[27]
[28]
[29]
[30]
G. M. Sheldrick, Acta. Cryst. 1990, A46, 467Ϫ473.
G. M. Sheldrick, J. Appl. Cryst. 1993.
WIN-MAS program (1994) version 940108, BrukerϪFranzen
Analytik, Bremen
Received July 31, 2000
M. J. Potrzebowski, M. Michalska, A. E. Koziol, S. Kazmier-
[O00403]
Eur. J. Org. Chem. 2001, 1491Ϫ1501
1501