Stereoselective P-Cyclisation and Diastereoisomeric Purification of 5-Phenyl-3-(pyridin-2-yl)-1,3,2-oxazaphospholidine Formed from a Thermolabile Protecting Group

Article abstract of DOI:10.1002/ejoc.201501596

A one-pot, two-step synthesis of 5-phenyl-3-(pyridin-2-yl)-1,3,2-oxazaphospholidine from linear precursor bis(diisopropylamino){2-[(pyridin-2-yl)amino]-1-phenylethoxy}phosphine is achieved using a stereoselective intramolecular cyclisation. Application of a pure enantiomer {1-phenyl-2-[(pyridin-2-yl)amino]ethanol} enabled partial diastereopurification by crystallisation. For all four diastereoisomers, the absolute configuration of the P-centre was determined using X-ray crystallography and correlative ³¹P NMR data. Stereochemically pure 5a was then used in nucleoside phosphitylation reactions with partial loss of stereopurity by retention of configuration on the phosphorus centre. A one-pot, two-step synthesis of 5-phenyl-3-(pyridin-2-yl)-1,3,2-oxazaphospholidine from linear bis(diisopropylamino){2-[(pyridin-2-yl)amino]-1-phenylethoxy}phosphine by stereoselective intramolecular P-cyclisation is described. For all four diastereoisomers produced, the absolute configuration of the P-centre is determined by X-ray crystallography and correlated with ³¹P NMR data.

Full text of DOI:10.1002/ejoc.201501596

DOI: 10.1002/ejoc.201501596
Full Paper
P-Cyclisations
Stereoselective P-Cyclisation and Diastereoisomeric Purification
of 5-Phenyl-3-(pyridin-2-yl)-1,3,2-oxazaphospholidine Formed
from a Thermolabile Protecting Group
Tomasz P. Kaczyński,[a] Tomasz Manszewski,[a] and Marcin K. Chmielewski*[a]
Abstract: A one-pot, two-step synthesis of 5-phenyl-3-(pyridin- tallisation. For all four diastereoisomers, the absolute configura-
2-yl)-1,3,2-oxazaphospholidine from linear precursor bis(diiso- tion of the P-centre was determined using X-ray crystallography
propylamino){2-[(pyridin-2-yl)amino]-1-phenylethoxy}phos- and correlative 31P NMR data. Stereochemically pure 5a was
phine is achieved using a stereoselective intramolecular cyclisa- then used in nucleoside phosphitylation reactions with partial
tion. Application of a pure enantiomer {1-phenyl-2-[(pyridin-2- loss of stereopurity by retention of configuration on the phos-
yl)amino]ethanol} enabled partial diastereopurification by crys- phorus centre.
Introduction
stall the internucleotidic diester phosphate bond in the dimer
and 17-mer combination with complete removal of the base
Five-membered heterocycles[1] with a stereogenic phosphorus protecting groups under standard deprotection conditions
atom are broadly applicable in stereoselective syntheses of (28 % NH4OH, 55 °C, 12 h).[15]
compounds involved in biopharmaceutical drug design ef-
forts.[2] As potential therapeutic agents, oligodeoxyribonucleo-
tide phosphorothioates (PS-ODNs)[3] and PS-nucleotides[4] re-
quire improved synthetic availability, hopefully by methods that
provide defined chirality on the P-centre. Oxathiaphosphol-
anes,[5,6] indoloxazaphosphorines,[7] xylose-oxazaphosphorin-
anes,[8] N-acylphosphoramidites,[9] and oxazaphospholidines[10]
have all been used to form P-stereospecific backbones in nu-
cleic acids. Application of stereocontrol in synthetic oligoribo-
nucleotide phosphorothioates (PS-ORNs) has been achieved by
the oxazaphospholidine approach[11] using highly nucleophilic
azole activators.[12] PS-ORNs have been successfully synthesized
on solid support with both sufficient coupling efficiencies (94–
99 %) and acceptable diastereoselectivities (≥ 98:2).[13]
Compounds bearing the 1,3,2-oxazaphospholidine frame-
work are promising agents in P-stereospecific synthesis. Such
agents have been obtained by the reaction of trivalent phos-
phorus with substituted amino alcohols or other amine deriva-
tives. 2-Octadecyl-1,3,2-oxazaphospholidine has been used in
oxidative decyclisations with bromine to produce cationic
amido diester phospholipids.[14] It is also well established that
2-chloro-3-methyl-1,3,2-oxazaphospholidine can serve as an al-
ternative to synthons commonly employed in oligonucleotide
synthesis. Moreover, 3-methyl-1,3,2-oxazaphopspholidine nu-
cleoside phosphoramidites have been successfully used to in-
However, 3-(pyridin-2-yl)-1,3,2-oxazaphospholidine is formed
as a result of intramolecular cyclisation of 2-PyTPG (2-Pyridinyl
Thermolabile Protecting Groups).[16,17] Consequently, 2-PyTPGs
temporarily “lose” their thermolabile properties, but gain stabil-
ity under ambient conditions. The presence of a pyridinyl moi-
ety changes bond polarisation and facilitates its reopening and
restoration of thermolability. This process is known as the “click-
clack“ approach, to denote the facile switching between two
states: labile and stable.[18] It has been shown that 2-PyTPGs
can support oxidation of H-phosphonates to phosphate diesters
when using 1,3,2-oxazaphospholidine oxide.[19] On the basis of
these realisations we reasoned that 2-(diisopropylamino)-5-
phenyl-3-(pyridin-2-yl)-1,3,2-oxazaphospholidine might be a
potentially convenient phosphitylating agent. The prospect that
this reagent might offer some level of stereocontrol during the
phosphitylation process enhanced our enthusiasm, especially in
light of the importance of stereopure nucleosides and oligo-
nucleotide analogues in both therapeutic and biological con-
texts.
Results and Discussion
Here, we report the stereoselective formation of 2-(diisopropyl-
amino)-5-phenyl-3-(pyridin-2-yl)-1,3,2oxazaphospholidines (5a–
d) resulting from a one-pot, two-step reaction of 1-phenyl-2-
[(pyridin-2-yl)amino]ethanol (3) (Scheme 1) and their potential
as phosphorylating agents. Also, studies of the crystal structure
and partial diastereopurification of 5a–d are presented. The
presence of a phenyl ring had two significant effects: (i) it gen-
erated a stereogenic centre, which, after P-cyclisation, formed
the diastereoisomers that are easily identifiable by 31P NMR
[a] Institute of Bioorganic Chemistry, Polish Academy of Sciences
Z. Noskowskiego str. 12/14, 61-704 Poznań, Poland
E-mail: maro@ibch.poznan.pl
www.ibch.poznan.pl
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/ejoc.201501596.
Eur. J. Org. Chem. 2016, 2522–2527
2522
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
spectroscopy, and (ii) it introduced steric considerations that without final purification (Scheme 2). Stable oxazaphospholi-
influence the stereochemistry at the phosphorus centre. P-Cycli- dine 5, the result of P-cyclisation, was purified by chromatogra-
sation of phosphorus-containing thermolabile protecting phy and its structure rigorously elucidated.
groups based on intramolecular attack of a nitrogen atom upon
the phosphorus atom in bis(diisopropylamino){1-phenyl-2-[(pyr-
idin-2-yl)amino]ethoxy}phosphine (4) afforded oxazaphosphol-
idines 5a–d. Consequently, the thermolabile properties of the
protecting group were “frozen”, and their recovery was not pos-
sible until the oxazaphospholidine ring was opened. P-Cyclisa-
tion of TPGs is possible in precursors bearing an H atom at the
pyridine-conjugated exocyclic amino moiety. 1-phenyl-2-[(pyr-
idin-2-yl)amino]ethanol (3) is such a precursor of TPGs and is
readily synthesised in a microwave reactor with an open vessel
using a waveguide-assisted platform to increase the yield. Puri-
fication of 3 was possible by crystallisation; its crystal structure
has been previously published.[20] This preliminary study
showed that the electron-withdrawing effect of the phen-
yl group depletes electron density of the asymmetric center
(Scheme 1) and thereby influences N–P bond generation in-
volved in cyclisation (Scheme 2).[20] The presented P-cyclisation
process was induced by activating one nitrogen atom in the
diisopropylamine substituent in 4, either spontaneously in the
presence of diisopropylamine hydrochloride, or by using a weak
acid in the reaction mixture. Our preliminary observation based
on 31P NMR analysis showed that P-cyclisation of 4 gave une-
qual quantities of two products. However, we observed that
both products were formed simultaneously. The one-pot, two-
step synthesis started from 1-phenyl-2-[(pyridyn-2-yl)amino]eth-
anol (3) and chloro(diisopropylamino)phosphine to obtain 4
P-Cyclisation of a racemic mixture of 4a proceeded with a
high degree of selectivity leading to a diastereomeric product
ratio of 87:13 (as determined on the basis of integration;
Scheme 2). Subsequently, a mixture of all four possible stereo-
isomers (5a–d), grouped in two pairs of enantiomers, were ob-
tained. In this arrangement, 31P NMR analysis in an achiral sol-
vent (benzene) showed two signals with different intensities
and a chemical shift of δ = 128.4 ppm for one pair of enantio-
mers and δ = 116.7 ppm for the other enantiomer pair (Fig-
ure 2A). Correlating the specific and relevant diasteroisomers
to concrete spectroscopic signals was not possible with any
certainty. However, after purification, compounds 5a–d were
easily crystallised. The 31P NMR analysis of redissolved crystals
showed the presence of only one peak (δ = 128.4 ppm; Fig-
ure 2B), and crystallographic analysis confirmed the presence
of both enantiomers 5a and 5b (Figure 1). This enabled us to
assign absolute configurations of (RC,SP) and (SC,RP) to 5a and
5b, respectively.
Scheme 1. Microwave-assisted 1-phenyl-2-[(pyridin-2-yl)amino]ethanol syn-
thesis.
Figure 1. Crystal structure of enantiomers 5a and 5b. The molecule in the
background is generated by a mirror plane. Ellipsoids are drawn at the 50 %
probability level. Hydrogen atoms are omitted for clarity.
In order to simplify the analysis, the P-centre was protected
by a TPG using precursor 3b with a defined (R) configuration
at the lone asymmetric center. 31P NMR studies of 4b generated
in this way gave the same result as had the racemic mixture of
4a; two peaks in a ratio of ca. 85:15 (Figure 3A) were clearly
apparent. Using the (R)-configured substrate 3b allowed us to
eliminate enantiomer 5b (SCRP) paired previously with 5a (RCSP)
in the crystal and observed as the same 31P NMR signal (δ =
128.4 ppm; Figure 2B).
Scheme 2. One-pot, two-step synthesis of 5 with mechanism of stereoselect-
ive intramolecular cyclisation and corresponding 31P NMR chemical shifts in-
dicated. (i) chloro(diisopropylamino)phosphine, benzene; (ii) benzene, spon-
taneously. Compound 4 was characterised by 31P NMR spectroscopy but not
isolated.
The products of P-cyclisation of 4b were also obtained as
crystals, but their crystallographic analysis[21] this time revealed
a different composition of diastereoisomers: 5a (RCSP) and 5c
Eur. J. Org. Chem. 2016, 2522–2527
www.eurjoc.org
2523
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Thus, crystallisation is possible only between two matching
diastereoisomeric forms (Figure 3B). Therefore, if 5a and 5c are
initially present in the solution in an unequal ratio (Figure 3A),
the solution after combining 5c with the corresponding
amount of 5a during crystallisation contains only the remaining
diastereoisomer 5a (Figure 3C); 5a is stable during weeks of
storage, and no transformation to 5b or other forms was ob-
served. The discussed process, i.e. diastereoselective purification
by crystallisation, may be referred to as “crystallisation by exclu-
sion”. This term denotes the fact that one diastereoisomer,
which is present in smaller quantities, is excluded from solution
by a crystallisation process that precedes dissolution in contrast
to simple purification by crystallisation.
The experiments described above enabled us to determine
the correlation between absolute configurations of 5a–d and
chemical shifts in their 31P NMR spectra. As a result, we could
confidently assign the 31P resonance at δ = 128 ppm to the
(RCSP) configuration and the resonance at δ = 116 ppm to the
(RCRP) configuration. The obtained data reveal that the pair of
enantiomers 5a and 5b correlate to one peak at δ = 128 ppm,
and the second pair of enantiomers 5c and 5d gives rise to the
NMR signal at δ = 116 ppm.
Figure 2. 31P NMR analysis of 5a–d: (A) reaction mixture; (B) redissolved crys-
tals (presence of a peak at δ = 116 ppm is probably caused by the solvent
residue, which contains some amount of 5c and 5d) at the surface of crystals.
Crystallographic data and 31P NMR studies also showed that
the syn configuration of the phenyl group and phosphorus sub-
stituent (in relation to the oxazaphospholidine ring) is preferred
among the P-cyclisation products identified. This stereoselectiv-
ity is probably dictated by steric interactions between all three
spatially large substituents of the oxazaphospholidine ring.
(RCRP) in the crystals and the redissolved crystals showed two
signals in an equal ratio upon 31P NMR analysis (Figure 3B).
Because 5c was obtained in smaller quantity and crystallised
completely with 5a, it became clear that the obtained solution
contained pure diastereoisomer 5a.
The obtained fraction with a pure diastereoisomer allowed
us to study the stereopurity of subsequent phosphitylation re-
actions and to exploit thermolabile properties in the formation
of phosphate and thiophosphate[22] backbones in nucleic acid
chemistry. To demonstrate this, a solution containing only one
diastereoisomer (RCSP) (5a) was used in phosphitylation reac-
tions with (i) a free primary hydroxy group of 3′-O-acetylthym-
idine, and (ii) a secondary hydroxy group of 5′-O-DMT-thym-
idine (Scheme 3). Unfortunately, the stereopurity was lost, and
two kinetic products were formed in unequal proportions. It is
known that phosphitylation reactions catalysed by weak acids,
for example 1H-tetrazole, are performed with complete inver-
sion of configuration. Based on studies of the kinetic and ther-
modynamic stability of the product formed,[23] it was concluded
that phosphitylation by 5a leads to the formation of two epi-
mers in a ratio of 75:25 in the case of 6 and a ratio of 85:15 in
the case of 8; both ratios result from inversion of the P-atom
configuration. Similar stereochemical outcomes were observed
when the mixture of all four diastereoiseomers (5a–d) was ap-
plied to similar phosphitylation reactions.
Finally, we found that both 6 and 8 were amenable to sulfur-
isation and that this chemistry did not alter the diastereomeric
ratio for each set of products. The ratio of diastereoisomers for
both 6 and 7 was 75:25 and the ratio for both 8 and 9 was
85:15; in neither case did P–S bond formation impact the
stereopurity of substrates 6 and 8. Notably, 1,3,2-oxazaphos-
pholidine 2-sulfides 10 (Scheme 4) are stable to acid hydrolysis
and do not generally suffer ring-opening reactions.
Figure 3. 31P NMR analysis of 5a and 5c: (A) reaction mixture; (B) redissolved
crystals; (C) solution after crystallisation.
Eur. J. Org. Chem. 2016, 2522–2527
www.eurjoc.org
2524
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 3. Reaction of (RCSP)-oxazaphospholidine (5a) with 3′-O-acetylthymidine and 5′-O-DMT-thymidine affording 6 and 8. Both 6 and 8 display partial loss
of stereopurity on the phosphorus centre. Phosphitylated 6 and 8 were then sulfurized to 7 and 9, respectively.
Experimental Section
Materials and General Methods: Common chemicals and solvents
were purchased from various commercial sources and used without
further purification. Benzene was distilled from P2O5 and CaH2, and
stored over 3 Å molecular sieves prior to use. Anhydrous acetonitrile
– commercially available (max. 30 ppm H2O) – was stored over 3 Å
molecular sieves to decrease the water level to ca. 5 ppm. Molecular
sieves were also used for anhydrous amines. 3′-O-Acetylthymidine
and 5-benzylmercaptotetrazole (BMT) were dried by lyophilisation
from benzene. The progress of the reactions was monitored by thin
layer chromatography (TLC) conducted on 2.5 cm × 7.0 cm glass
plates coated with a 0.25 mm thick layer of silica gel 60 F254. Chro-
matography on silica gel columns was performed using silica gel
60 (70–230 mesh).
Scheme 4. Compound 10 is useless as a phosphitylating agent (for more
details see the Supporting Information).
Conclusions
Microwave Reactor: An ERTEC® open-vessel microwave reactor was
used, power range 0–750 W with stepless adjustment, field fre-
quency 2.45 GHz and reflected power monitoring. The measure-
ment was performed using a pyrometer, and the reaction tempera-
ture was monitored in the range of 0–500 °C.
In the case of the presented cyclisation of 4a to 5, diastereo-
selectivity was observed. Using substrate 4b with a known con-
figuration as well as correlating the results of 31P NMR and crys-
tallographic analyses allowed us to ascribe specific resonances
to individual stereoisomers. Hence, we suggest that the syn
configuration of substituents on the asymmetric carbon and
phosphorus atoms is preferred. It was also possible to isolate,
by crystallisation, a pure product 5a with known absolute con-
figuration. We attempted to use 5a as a phosphitylating agent
to stereoselectively obtain a thymidine oxazaphospholidine an-
alogue that is stable during storage, yet easily convertible to 2-
PyTPG. Because nucleoside phosphitylation by 5a resulted in
partial loss of P-atom stereopurity, we plan to continue studies
1H, 13C, 31P NMR and Mass Spectrometer Parameters: The NMR
spectra were obtained with a Bruker spectrometer (400 or 500 MHz
for 1H and 125 MHz for 13C). For 31P NMR analysis, a Varian 300 MHz
and a Bruker 400 MHz spectrometer were used. Chemical shifts δ
are reported in parts per million (ppm). The residual protonated
solvent was used as internal standard ([D6]DMSO at δ = 2.50 ppm).
J values are given in Hz. Abbreviations used in the description of
resonances are: s (singlet), d (doublet), t (triplet), br. (broad), dd
(double doublet), ddd (double double doublet) and m (multiplet).
to find an effective method for stereopure phosphitylation as The mass spectrometer was equipped with an electrospray ionisa-
tion source (ESI) and a q-TOF analyser. Source parameters are as
follows: ESI source voltage of 3.2 kV, nebulisation with nitrogen at
0.4 bar, dry gas flow of 4.0 L/min at 220 °C or 50 °C.
a stereocontrolled method in nucleic acid chemistry. Further
research will be focused on oxazaphospholidine ring-opening
reactions using nucleophilic agents such as hydroxy, amino or
phosphate groups as well as possible ways of recovering the
thermolabile properties of TPGs.
X-ray Crystallography Studies: All crystals were obtained from
hexane by slow concentration at room temperature until the first
Eur. J. Org. Chem. 2016, 2522–2527
www.eurjoc.org
2525
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
crystals could be observed in the solution. Then, the solution was
stored at 5 °C overnight. Data collections were performed at 130 K
with a SuperNova diffractometer[24] using a mirror monochromator
7.8 Hz, 1 H), 6.74 (d, J = 8.3 Hz, 1 H), 6.72 (dd, J = 6.9, 5.1 Hz, 1 H),
5.19 (dd, J = 9.9, 6.3 Hz, 1 H), 4.04 (ddd, J = 12.0, 10.1, 6.3 Hz, 1 H),
3.24 (td, J = 10.1, 1.5 Hz, 1 H) ppm. 13C NMR (125 MHz, DMSO): δ =
for Cu-Kα radiation (λ = 1.5418 Å). Corrections were made for the 156.7, 156.6, 148.1, 140.2, 137.7, 136.9, 126.7, 126.1, 114.4, 109.1,
Lorentz-polarisation effect and for absorption. Unit-cell parameters
were determined by a least-squares fit of 5558 (A = 5a+5b) and
22781 (B0 = 5a+5c) reflections of highest intensity, selected from
the whole experiment. SIR92[25] was used for structure solution. Re-
finement with the full-matrix procedure on F2 was carried out in
SHELXL97.[26] The function Σw(|Fo|2 – |Fc|2)2, where w–1 = [σ2(Fo)2 +
77.2, 50.5, 44.9, 44.8, 25.1, 23.8 ppm. 31P NMR (300 MHz, benzene):
δ = 128.5, 116.9 ppm. HRMS (ESI-q-TOF): calcd. for C19H27N3OP [M
+ H]+ 344.1886, found 344.1891.
3′-O-Acetyl-5′-O-[5-phenyl-3-(pyridin-2-yl)-1,3,2-oxazaphos-
pholidin-2-yl]thymidine (6): 2-(Diisopropylamino)-5-phenyl-3-(pyr-
idin-2-yl)-1,3,2-oxazaphospholidine (5) (0.19 g, 0.55 mmol) and 3′-O-
acetylthymidine (0.156 g, 0.55 mmol) were dissolved in anhydrous
acetonitrile (2.75 mL). Then, 5-benzylmercaptotetrazole (1.206 mL
2
A·P2 + B·P] and P = [max(Fo2,0) + 2Fc ]/3, was minimised. All non-
hydrogen atoms were refined anisotropically, while the positions of
hydrogen atoms were calculated and refined as a riding model.
CCDC 1435750 [for A (5a+5b)], and 1435740 [for B0 (5a+5c)] con-
tain the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge Crystallo-
graphic Data Centre.
of 0.25
M solution in anhydrous acetonitrile, 0.35 mmol) was added.
The reaction was monitored by 31P NMR spectroscopy and finished
after 24 h. The resulting mixture was concentrated, and the product
was isolated on a silica gel column, eluting with dichloromethane/
methanol (methanol gradient 0→5 % in dichloromethane). The
fractions were concentrated and dried. The product was obtained
1-Phenyl-2-[(pyridin-2-yl)amino]ethanol (3): 2-Amino-1-phenyl-
ethanol (2 g, 14.6 mmol) was dissolved in 2-bromopyridine (1.152 g,
7.3 mmol). Triethylamine (0.96 g, 9.5 mmol) and a few drops of N-
ethyldiisopropylamine were added to the mixture. The reaction was
carried out in a microwave reactor (microwave power: 185 W), at a
temperature range of 158–165 °C using a magnetic stirrer. Progress
of the reaction was monitored using TLC (dichloromethane/meth-
anol, 95:5). The reaction was terminated after 8 h. The resulting
mixture was concentrated, and the product was isolated on a silica
gel column, eluting with dichloromethane/methanol (methanol gra-
dient 0→5 % in dichloromethane). The fractions were concentrated
and dried to give the product as white powder (0.8 g, 51 % yield).
1H NMR (500 MHz, DMSO): δ = 7.97 (dd, J = 5.0, 1.2 Hz, 1 H), 7.38–
7.31 (m, 5 H), 7.25–7.22 (m, 1 H), 6.52 (dd, J = 11.4, 7.2 Hz, 2 H),
6.48–6.46 (m, 1 H), 5.67 (s, 1 H), 4.75 (dd, J = 7.7, 4.3 Hz, 1 H), 3.51
(ddd, J = 13.4, 6.7, 4.3 Hz, 1 H), 3.32–3.24 (m, 1 H) ppm. 13C NMR
(125 MHz, DMSO): δ = 158.8, 147.3, 144.3, 136.6, 127.9, 126.8, 126.0,
111.6, 108.6, 71.6, 49.3 ppm. HRMS (ESI-q-TOF): calcd. for C13H15N2O
[M + H]+ 215.1179, found 215.1178.
1
with 25 % yield (0.08 g). H NMR (500 MHz, DMSO): δ = 11.33 (s, 1
H), (t, 1 H), 7.6 (d, J = 11.4 Hz, 1 H), 7.49 (dd, J = 11.1, 8.1 Hz, 2 H),
7.45–7.35 (m, 4 H), 6.87–6.84 (m, 1 H), 6.73 (dd, J = 8.3, 4.3 Hz, 1 H),
5.84 (ddd, J = 13.1, 10.1, 6.4 Hz, 1 H), 5.20 (dd, J = 12.4, 5.6 Hz, 2
H), 4.20–3.99 (m, 4 H), 3.97 (d, J = 1.9 Hz, 2 H), 2.27 (dd, J = 8.5,
6.1 Hz, 2 H), 2.06 (s, 3 H), 1.78 (s, 3 H) ppm. 13C NMR (125 MHz,
DMSO): δ = 170.0, 163.7, 163.6, 150.5, 147.8, 138.5, 128.5, 128.3,
126.5, 126.4, 115.5, 109.9, 109.7, 107.8, 107.7, 84.6, 83.7, 74.7, 61.3,
36.5, 20.9, 12.3 ppm. 31P NMR (300 MHz, acetonitrile): δ = 127.6,
132.6, 134, 137.8 ppm. HRMS (ESI-q-TOF): calcd. for C25H28N4O7P [M
+ H]+ 527.1690, found 527.1692.
3′-O-Acetyl-5′-O-[5-phenyl-3-(pyridin-2-yl)-2-sulfido-1,3,2-ox-
azaphosphosphlidin-2-yl]thymidine (7): 3′-O-Acetyl-5′-O-[5-
phenyl-3-(pyridin-2-yl)-1,3,2-oxazaphospholidin-2-yl]thymidine (6)
(50 mg, 0.95 mmol) was dissolved in anhydrous acetonitrile
(4.75 mL), then sulfur (3 equiv., 9 mg, 0.28 mmol) was added. The
reaction mixture was stirred under argon and monitored by 31P
NMR spectroscopy. After 15 min, the reaction was completed with
a yield of ca. 90 %, estimated from 31P NMR peak areas. 31P NMR
(300 MHz, acetonitrile): δ = 74.4, 75.2 ppm. HRMS (ESI-q-TOF): calcd.
for C25H28N4O7PS [M + H]+ 559.1411, found 559.1439.
2-(Diisopropylamino)-5-phenyl-3-(pyridin-2-yl)-1,3,2-oxaza-
phospholidine (5a–d): PCl3 (0.089 g, 0.65 mmol) was dissolved un-
der an inert gas in anhydrous benzene (3.25 mL) in a flask with a
magnetic stirrer, at 0 °C (ice cooling). Then, a mixture of anhydrous
diisopropylamine (0.066 g, 0.65 mmol) and N-ethyldiisopropylamine 5′-O-(Dimethyltrityl)-3′-O-[5-phenyl-3-(pyridin-2-yl)-1,3,2-ox-
(0.295 g, 2.27 mmol) was added in portions. The mixture was
warmed to room temperature and left for 24 h. The level of conver-
sion was monitored by 31P NMR spectroscopy. After full conversion
azaphospholidin-2-yl]thymidine (8): 2-(Diisopropylamino)-5-
phenyl-3-(pyridin-2-yl)-1,3,2-oxazaphospholidine (5) (150 mg,
0.43 mmol) and 5′-O-(dimethyltrityl)thymidine (240 mg, 0.55 mmol)
of diisopropylphosphoramidous dichloride (loss of signal δ = were dissolved in anhydrous acetonitrile (5.5 mL). Then, 5-benzylm-
168.8 ppm) to chlorobis(diisopropylamino)phosphine (δ =
134.5 ppm), the mixture was cooled to 0 °C in an ice bath, and then
ercaptotetrazole (0.5 equiv., 41 mg, 0.22 mmol) was added. The
reaction was monitored by 31P NMR spectroscopy and finished after
dry 1-phenyl-2-[(pyridin-2-yl)amino]ethanol (0.14 g, 0.65 mmol) was 24 h. The resulting mixture was concentrated, and the product was
added in portions using a flow of inert gas for 3 h. After all alcohol
had been added, the reaction was carried out at room temperature
for 12 h, and the progress was monitored by 31P NMR spectroscopy
to obtain bis(diisopropylamino){1-phenyl-2-[(pyridin-2-yl)-
amino]ethoxy}phosphine (4) with a 31P NMR shift at δ = 113.1 ppm;
after another 12 h without other operations, products of cyclisation
(5a–d) were observed at δ = 115.9 and 128.5 ppm. Excess benzene
isolated on a silica gel column, eluting with dichloromethane/meth-
anol (methanol gradient 0→5 % in dichloromethane). The fractions
were concentrated and dried. The product was obtained with 52 %
yield (180 mg). 1H NMR (500 MHz, [D6]DMSO): δ = 11.32 (s, 1 H),
8.06 (dd, J = 4.9, 1.8 Hz, 1 H), 7.63 (td, J = 7.8, 1.9 Hz, 1 H), 7.51 (s,
1 H), 7.47 (d, J = 7.1 Hz, 3 H), 7.42 (t, J = 7.3 Hz, 2 H), 7.40–7.31 (m,
4 H), 7.28 (t, J = 7.5 Hz, 2 H), 7.21 (d, J = 8.9 Hz, 4 H), 6.84 (pd, J =
was evaporated, and the product was isolated on a silica gel column 8.8, 3.3 Hz, 6 H), 5.78 (dd, J = 10.0, 6.3 Hz, 1 H), 4.98 (ddt, J = 10.1,
(benzene/triethylamine, 95:5); the fractions were analysed by TLC. 6.8, 3.4 Hz, 1 H), 4.03 (d, J = 3.9 Hz, 1 H), 3.73 (d, J = 1.6 Hz, 2 H),
The mixture was dissolved in n-hexane (5 mL) and allowed to crys- 3.70 (d, J = 3.3 Hz, 6 H), 3.27–3.15 (m, 2 H), 2.44 (q, J = 7.1 Hz, 2 H),
tallise. After a few hours, the crystals were dried. The product was
obtained with 65 % yield (0.144 g). The expected structure of the
product was also confirmed by crystallographic analysis. 1H NMR
(500 MHz, DMSO): δ = 8.11 (dd, J = 4.9, 1.2 Hz, 1 H), 7.58–7.55 (m,
1 H), 7.49–7.48 (m, 2 H), 7.42 (dd, J = 14.3, 6.6 Hz, 2 H), 7.34 (t, J =
1.49 (s, 3 H) ppm. 13C NMR (500 MHz, DMSO): δ = 163.6, 158. 1,
158.1, 155.5, 155.4, 150.3, 150.3, 147.8, 144.6, 138.7, 138.7, 138.5,
138.4, 135.7, 135.3, 135.2, 135.2, 135.1, 129.7, 129.62, 128.5, 127.8,
127.6, 126.3, 126.3, 115.4, 115.2, 113.2, 109.7, 107.7, 107.6, 85.9, 84.0,
84.0, 83.7, 81.3, 81.2, 74.0, 73.9, 63.1, 55.0, 55.0, 55.0, 45.7, 11.7,
Eur. J. Org. Chem. 2016, 2522–2527
www.eurjoc.org
2526
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
11.6 ppm. 31P NMR (300 MHz, acetonitrile): δ = 137.36, 128.79 ppm.
HRMS (ESI-q-TOF): calcd. for C44H43N4NaO8P [M + Na]+ 809.2711,
found 809.2714.
[2] D. Krasowska, J. Chrzanowski, J. Drabowicz, Adv. Heterocycl. Chem. 2015,
117, 179–259.
[3] N. Oka, T. Wada, Chem. Soc. Rev. 2011, 40, 5829–5843.
[4] P. Guga, M. Koziolkiewicz, Chem. Biodiversity 2011, 8, 1642–1681.
[5] P. Guga, A. Okruszek, W. J. Stec, Top. Curr. Chem. 2002, 220, 169–200.
[6] Nawrot, B. Rebowska, O. Michalak, M. Bulkowski, D. Baziak, P. Guga, W. J.
Stec, Pure Appl. Chem. 2008, 80, 1859–1871.
[7] J. C. Wang, G. Just, Tetrahedron Lett. 1997, 38, 3797–3800.
[8] Y. Jin, G. Just, J. Org. Chem. 1998, 63, 3647–3654.
[9] A. Wilk, A. Grajkowski, L. R. Phillips, S. L. Beaucage, J. Am. Chem. Soc.
2000, 122, 2149–2156.
[10] T. Wada, Front. Org. Chem. 2005, 1, 41–61.
[11] N. Oka, T. Kondo, S. Fujiwara, Y. Maizuru, T. Wada, Org. Lett. 2009, 11,
967–970.
[12] N. Oka, T. Wada, K. Saigo, J. Am. Chem. Soc. 2002, 124, 4962–4963.
[13] Y. Nukaga, K. Yamada, T. Ogata, N. Oka, T. Wada, J. Org. Chem. 2012, 77,
7913–7922.
[14] D. R. Cooper, C. R. Hall, J. M. Harrison, T. D. Inch, J. Chem. Soc. Perkin
Trans. 1 1977, 1969–1980.
5′-O-(Dimethyltrityl)-3′-O-[5-phenyl-3-(pyridin-2-yl)-2-sulfido-
1,3,2-oxazaphospholidin-2-yl]thymidine (9): 5′-O-(Dimethyl-
trityl)-3′-O-[5-phenyl-3-(pyridin-2-yl)-1,3,2-oxazaphospholidin-2-yl]-
thymidine (8) (50 mg, 0.64 mmol) was dissolved in anhydrous aceto-
nitrile (3.25 mL), then sulfur (3 equiv., 6 mg, 0.28 mmol) was added.
The reaction mixture was stirred under argon and monitored by 31
P
NMR spectroscopy. After 30 min, the reaction was completed with
a yield of ca. 95 %, estimated from 31P NMR peak areas. 31P NMR
(300 MHz, acetonitrile): δ = 73.7, 73.9 ppm. HRMS (ESI-q-TOF): calcd.
for C44H43N4NaO7PS [M + Na]+ 841.2431, found 841.2398.
2-(Diisopropylamino)-3-(pyridin-2-yl)-5-phenyl-1,3,2-oxaza-
phospholidine 2-Sulfide (10): 2-(Diisopropylamino)-5-phenyl-3-
(pyridin-2-yl)-1,3,2-oxazaphospholidine (5a–d) (50 mg, 0.145 mmol)
was dissolved in anhydrous acetonitrile (2.25 mL), then sulfur
(3 equiv. 14 mg, 0.43 mmol) was added. The reaction mixture was
stirred under argon and monitored by 31P NMR spectroscopy. After
30 min, the reaction was completed with a yield of ca. 95, estimated
from 31P NMR peak areas. 31P NMR (300 MHz, acetonitrile): δ = 68.9,
71.4 ppm. HRMS (ESI-q-TOF): calcd. for C19H26N3NaOPS [M + Na]+
398.1426, found 398.1429.
[15] D. A. Predvoditelev, M. A. Malenkovskaya, E. E. Nifant'ev, Russ. J. Gen.
Chem. 2005, 75, 53–57.
[16] M. K. Chmielewski, V. Marchan, J. Cieślak, A. Grajkowski, V. Livengood, U.
Munch, A. Wilk, S. L. Beaucage, J. Org. Chem. 2003, 68, 10003–10012.
[17] M. K. Chmielewski, Tetrahedron Lett. 2012, 53, 666–669.
[18] M. K. Chmielewski, Org. Lett. 2009, 11, 3742–3745.
[19] T. Ratajczak, M. K. Chmielewski, J. Org. Chem. 2012, 77, 7866–7872.
[20] M. K. Chmielewski, E. Tykarska, W. T. Markiewicz, W. R. Rypniewski, New
J. Chem. 2012, 36, 603–612.
[21] Extended crystallographic analyses can be found in the Supporting In-
Acknowledgments
formation (Figure S4, B0, page 26).
[22] Compound 5 can be easily transformed into the corresponding sulfide.
However, based on the results of the performed experiments it can be
concluded that: (i) it cannot be used as a phosphitylating agent in a
reaction with nucleotides, and (ii) it is resistant to hydrolysis under acidic
conditions, which makes ring-opening difficult. More information can be
found in the Supporting Information (Figure S3, page 4).
[23] The kinetic studies with 5 can be found in the Supporting Information
(Scheme S1).
This research was supported by a grant from the Polish National
Centre of Sciences granted under decision No. DEC-2012/07/B/
ST5/03035. This publication was supported by the Polish Minis-
try of Science and Higher Education under the KNOW pro-
gramme.
[24] CrysAlis PRO, version 171.35.4, Agilent Technologies, 2010, http://
www.agilent.com/.
[25] A. Altomare, G. Cascarano, C. Giacovazzo, A. Gualardi, J. Appl. Crystallogr.
1993, 26, 343–350.
Keywords: Thermolability · Protecting groups · Cyclization ·
1,3,2-Oxazaphospholidine · Stereoselectivity ·
Diastereoselectivity · Phosphorus heterocycles · Nucleosides
[26] G. M. Sheldrick, Acta Crystallogr., Sect. A 2008, 64, 112–122.
[1] P. Lemmen, W. Richter, B. Werner, R. Karl, R. Stumpf, I. Ugi, Synthesis 1993,
Received: December 28, 2015
Published Online: April 20, 2016
1–10.
Eur. J. Org. Chem. 2016, 2522–2527
www.eurjoc.org
2527
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim