Studies on the decomposition pathways of diastereoisomeric mixtures of aryl nucleoside α-hydroxyphosphonates under hydrolytic conditions. Synthesis of α-hydroxyphosphonate monoesters

Article abstract of DOI:10.1039/b9nj00717b

Decomposition pathways of nucleoside α-hydroxyphosphonates 1 (diastereomeric mixtures) bearing different aryl groups, both in the ester and the hydroxymethine fragment, were investigated under various hydrolytic conditions. We found that in aqueous basic media, the stability and decomposition pathways of these compounds were governed by the electronic features of the aryl group in the hydroxymethine moiety (hydroxyphosphonate ? H-phosphonate diester + aldehyde equilibria) and the nature of attacking nucleophiles (α-nucleophiles, e.g. hypoiodite or peroxide anions). The significant differences observed in the rates of hydrolysis of hydroxyphosphonates 1vs. their O-acylated derivatives point to the importance of an intramolecular acid catalysis exerted by the adjacent hydroxyl function. Based on these findings, an efficient synthetic protocol for otherwise difficult to access hydroxyphosphonate monoesters of type 7 has been developed.

Full text of DOI:10.1039/b9nj00717b

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Studies on the decomposition pathways of diastereoisomeric mixtures
of aryl nucleoside a-hydroxyphosphonates under hydrolytic conditions.
Synthesis of a-hydroxyphosphonate monoestersw
´ ´
Agnieszka Szymanska-Michalak, Jacek Stawinski and Adam Kraszewski*
Received (in Montpellier, France) 30th November 2009, Accepted 25th January 2010
First published as an Advance Article on the web 5th March 2010
DOI: 10.1039/b9nj00717b
Decomposition pathways of nucleoside a-hydroxyphosphonates 1 (diastereomeric mixtures)
bearing different aryl groups, both in the ester and the hydroxymethine fragment, were
investigated under various hydrolytic conditions. We found that in aqueous basic media, the
stability and decomposition pathways of these compounds were governed by the electronic
features of the aryl group in the hydroxymethine moiety (hydroxyphosphonate "
H-phosphonate diester + aldehyde equilibria) and the nature of attacking nucleophiles
(a-nucleophiles, e.g. hypoiodite or peroxide anions). The significant differences observed in
the rates of hydrolysis of hydroxyphosphonates 1 vs. their O-acylated derivatives point to
the importance of an intramolecular acid catalysis exerted by the adjacent hydroxyl function.
Based on these findings, an efficient synthetic protocol for otherwise difficult to access
hydroxyphosphonate monoesters of type 7 has been developed.
Introduction
Results and discussion
By exploring pronucleotides1 for AIDS therapy, we have
developed aryl nucleoside a-hydroxyphosphonates of type 1
(Scheme 1) as a new type of potential drug.2 These com-
pounds also appear to be very interesting from a chemical
point of view, since a change of aryl substituent in their
phosphoester and/or hydroxymethine parts, dramatically
affect their hydrolytic decomposition pathways. The decompo-
sition studies of these compounds were carried out in cell
culture media [RPMI/FBS, 9 : 1 (v/v), pH 7.4],3 which
resembles the physiological conditions that are routinely used
for the in vitro examination of the anti-HIV potency of new
compounds. Since RPMI media is a complex mixture of
chemical compounds, the decomposition pathways of investi-
gated compounds often differ significantly from those carried
out in simple buffered solutions of comparable pH.4 Due to
potential importance of the title compounds in the pro-drug
strategy for AIDS therapy, we undertook further studies to
gain a deeper insight into their hydrolytic properties, relevant
to their possible applications as anti-HIV agents. In particular,
we were interested in exploring possible new decompo-
sition pathways of these compounds, different from those
observed in RPMI media, e.g., a selective hydrolysis to the
corresponding a-hydroxyphosphonate monoester.
Searching for simple, well-defined hydrolytic conditions for
a-hydroxyphosphonate diesters 1,z we chose initially two
reaction media: (i) a pyridine : water mixture (1 : 3 v/v), and
(ii) acetonitrile : triethylamine : water (2 : 1 : 1 v/v/v). The
choice of the systems was dictated by the solubility of the
hydroxyphosphonates and their expected lability at pH 4 7.
Earlier, we postulated that in solution, aryl nucleoside
a-hydroxyphosphonates of type 1 exist in equilibrium with their
synthetic precursors, i.e. their respective H-phosphonate diesters
and aldehydes (of type 2 and 3, respectively; Scheme 1),2 and
we proved this by trapping these H-phosphonate diesters via
sulfurization with elemental sulfur. Unfortunately, since oxi-
dation by sulfur was significantly slower than the rates of the
equilibria, we could not estimate equilibrium constants for the
a-hydroxyphosphonate diesters investigated. To remedy this
problem, we attempted to use for the trapping experiments a
rapid oxidation of H-phosphonate diesters 2 with iodine,5,6
and for this reason we have also included in our investigation a
reagent system consisting of acetonitrile : triethylamine : water
(2 : 1 : 1 v/v/v) with iodine (3 equiv.) as an oxidant.
Hydrolysis of diastereoisomeric mixtures of
a-hydroxyphosphonates 1 in pyridine–water and
acetonitrile–triethylamine–water solvent systems
In a mixture of pyridine : water (1 : 3, v/v, pH 9.6), the investi-
gated a-hydroxyphosphonates 1 were rather stable, and after 4 h
Institute of Bioorganic Chemistry, Polish Academy of Sciences,
Noskowskiego 12/14, 61-704 Poznan´, Poland.
E-mail: adam.kraszewski@ibch.poznan.pl; Fax: +48 61 852 05 32;
Tel: +48 61 852 85 03
w This paper is dedicated to Prof. Dr Wojciech J. Stec, Center of
Molecular and Macromolecular Studies, Polish Academy of Sciences,
´
Łodz´ , Poland, on the occasion of his 70th birthday. This article is part
of a themed issue on Biophosphates.
z In the 31P NMR spectra of a-hydroxyphosphonates of type 1, four
signals of nearly the same intensity were observed due to the presence a
mixture of 4 diastereoisomers (see also ref. 2). Since in all the
experiments performed on these compounds the diastereoisomers
showed the same reactivity, no attempt was made to separate these
mixtures into individual compounds.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
976 | New J. Chem., 2010, 34, 976–983

View Article Online
Scheme 1 Decomposition paths of AZT aryl a-hydroxyphosphonates of type 1 in various hydrolytic media.
none or just a few percent of the decomposition products
could be observed by 31P NMR spectroscopy. These results
were somewhat surprising in light of the experiments carried
This seems to be consistent with the expected position of the
equilibria of hydroxyphosphonates, which are controlled by
stabilization of the carbocations formed as transient species
after scission of the P–C bond7 in 1, and is crucial for the
decomposition pathways observed for these compounds.
Because path A depends entirely on the equilibrium 1 " 2 + 3,
we attempted to eliminate this equilibrium by protecting the
hydroxyl function of 1, and in this way favour decompo-
sition pathway B. To this end, we chose a-hydroxyphosphonate
1c, which decomposed nearly exclusively via route A (Table 1),
and protected its a-hydroxyl function by an acetyl group
(Scheme 2). It was found that acetylation of the hydroxyl
function of hydroxyphosphonate 1c suppressed path A, and
the produced acetylated hydroxyphosphonate 9 mainly under-
went decomposition via pathway B. (Scheme 2). The observed
formation of H-phosphonate 4 (ca. 30%, 31P NMR) was
apparently due to the incomplete stability of the acetyl group
under the reaction conditions and the regeneration of the
starting hydroxyphosphonate 1c (Scheme 2).
1
out in RPMI solution, in which the decomposition t times of
2
a-hydroxyphosphonates 1a–g were between 10–30 min (Table 1).
In contrast to the pyridine : water mixture, hydrolysis in the
acetonitrile : triethylamine : water (2 : 1 : 1, v/v/v, pH 12.5)
system occurred much faster and was conveniently monitored
in 31P NMR time (few minutes). Despite the differences in
rates, for most of the investigated hydroxyphosphonates 1,
two decomposition pathways were observed: path A (the
dominant one), leading ultimately to the accumulation of
H-phosphonate monoester 4, and path B (the minor one),
affording a-hydroxyphosphonate monoesters 7 as products.
Along the series 1a–g, the advantage of pathway A over
pathway B was variable, but clearly pronounced in each case,
except for hydroxyphosphonate 1e, bearing a strong electron-
withdrawing Ar2 group (4-nitrophenyl), for which three decompo-
sition pathways were observed, with pathway A being the
minor one. From these data, it seemed apparent that the
participation of pathway A in the decomposition of hydroxy-
phosphonates 1 paralleled the increasing electron density of
Ar2, being the highest for hydroxyphosphonate 1c, with a
4-methoxy group as Ar2, and the lowest for the 4-nitrophenyl
derivative 1e. Interestingly, electron-donating or electron-
withdrawing substituents in the aromatic ring of Ar1 did not
affect significantly the hydrolytic decomposition pathways of
hydroxyphosphonates 1, as is apparent from a comparison
of 4-methoxyphenyl (1f) vs. p-chlorophenyl (1g) derivatives.
Apart from a completely different product distribution
pattern in the hydrolysis of hydroxyphosphonate 1c vs. its
acylated derivative 9 (pathway A vs. pathway B), significant
differences in rates of their hydrolysis (30 min vs. 424 h;
Table 1) were observed. These we attributed to the suppression
of path A, leaving path type B as the only one acting in
the decomposition of 9. At this point, we also noticed the
importance of the presence of a free hydroxyl function in an
a-position to the phosphorus centre in the hydrolysis of title
compounds 1 (see also later in the text for reactions in the
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 976–983 | 977

View Article Online
Table 1 The stability and decomposition pathways of diastereoisomeric mixtures of aryl nucleoside hydroxyphosphonates 1 under various
hydrolytic conditionsa
CH3CN : H2O : Et3N 2 : 1 : 1 (v/v);
pH 12.5
CH3CN : H2O : Et3N 2 : 1 : 1 (v/v) + I2;b
pH 10.4
RPMI : FBS 9 : 1 (v/v);2 pH 7.4
t /min A (%) B (%) C (%) tc/min
A (%) B (%) C (%) tc/min
Ad (%)
B (%)
C (%)
Cpd Ar1
Ar2
1
2
1a
27
19
100
100
0
0
0
0
30
45
65
85
35
15
0
0
o5
o5
10
90
0
1b
1c
1d
10
10
10
90
90
90
0
0
0
13
13
100
92
0
8
0
0
30
10
97
55
3
0
0
o5
o5
45
1e
1f
9
24
47
0
29
0
5
15
80
25
20
60
0
o5
10
15
90
85
0
0
26
100
40
10
1g
—
—
—
—
—
—
—
—
5
75
25
70
0
0
o5
2
98
97
0
0
9
424 h 30e
424 h
3f
a
b
c
The amount of a particular compound in the reaction mixture was estimated by the 31P NMR spectroscopy. 3.0 molar equivalents. Time for
d
the complete disappearance of substrate 1. Estimated on the basis of the amount of aryl nucleoside phosphate 6, a product of the oxidation of
e
f
aryl nucleoside H-phosphonate 2. The presence of H-phosphonate 4 was due to the deacetylation of 9 (for details, see text). Phosphodiester 6,
vide supra.
presence of iodine). Various mechanistic propositions can be
put forward to explain the kinetic enhancement of path B in
hydroxyphosphonates 1 when a free a-hydroxyl group is
present, and some of them are depicted in Scheme 3.
hydrolysis of hydroxyphosphonates 1 vs. their O-acylated
derivatives. However, an SN2(P) process, in which a nucleo-
philic attack by a hydroxide is reinforced by an intramolecular
acid catalysis exerted by the adjacent a-hydroxyl function
(species VI in Scheme 3), seems a viable mechanism for the
formation of phosphonate monoester 7. Although both mecha-
nisms in Scheme 3, i.e. the one proceeding via an oxaphos-
phirane III intermediate and that involving an intramolecular
acid catalysis, are possible, the high sensitivity to hydrolysis of
hydroxyphosphonates 1 to a-nucleophiles (vide infra) seems to
lend more support to the latter.
For hydroxyphosphonates 1 in basic media, it is likely that a
hydroxyl group can act as an intramolecular nucleophile towards
the phosphorus center (irrespective of its configuration) and
expel a leaving group OAr1 to form, via an intermediate penta-
valent phosphorane II, a cyclic oxaphosphirane of type III.8
The latter species is expected to be rather reactive, and upon
attack of a hydroxide ion, collapse to a hydroxyphosphonate
7, via a phosphorane IV intermediate.9
The decomposition pathways observed during the hydro-
lysis of aryl nucleoside a-hydroxyphosphonate 1e (Table 1) are
also understandable in light of the above discussion.
A simple SN2(P) mechanism involving a nucleophilic attack
of a hydroxide on the phosphorus center (Scheme 3; 1 - V - 7)
without participation of the a-hydroxyl function is rather
unlikely, since it does not explain the observed differences in
For hydroxyphosphonate 1e, due to the presence of a
strong electron-withdrawing 4-nitro substituent in Ar2, the
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
978 | New J. Chem., 2010, 34, 976–983

View Article Online
Scheme 2 Decomposition of aryl nucleoside 50-a-acetyloxyphosphonates 9.
decomposition path A is significantly suppressed. Since
the acidity of the a-hydroxyl group and the electrophilicity
of the phosphorus centre in 1e are simultaneously increased,
the reaction pathway via a phosphirane II (Scheme 3 and
Scheme 4) is favoured. This initially formed intermediate then
undergoes decomposition in two ways: by P–C bond scission
and formation of phosphotriester 8 (path C, phosphonate–
phosphate rearrangement7,10–12) or with the expulsion of leaving
group OAr1, followed by hydrolysis of an oxaphosphirane
intermediate III to form the corresponding hydroxyphospho-
nate 7 (path B). Since the former pathway proceeds apparently
via a carbanion, whose formation should be facilitated by
electron-withdrawing groups, i.e. 4-nitrophenyl, the decompo-
sition of 1e according to path C became the predominant one.
In a separate experiment, we observed that under anhydrous
reaction conditions (acetonitrile : triethylamine 2 : 1 v/v), aryl
nucleoside a-hydroxyphosphonate 1e underwent exclusively
the phosphonate–phosphate rearrangement to produce
phosphotriester 8.
instances, a-hydroxyphosphonates 7 were formed as the main
products (85–98%, Table 1). The reactions investigated were
insensitive to the electronic features of the OAr2 groups present
in hydroxyphosphonates 1, and the amounts of phospho-
diesters 6 formed correlated with the reaction times. This is
consistent with kinetic competition between path B and path A
(ca. 10% for phenyl hydroxyphosphonates 1a–e, 15% for
4-methoxyphenyl derivative 1f and 2% for 4-chlorophenyl
derivative 1g), indicating that the pKa value of the leaving
group OAr1 is the main factor governing the rate of hydrolysis
of hydroxyphosphonates 1 under these conditions.
As to a possible origin of the accelerating effect on hydroxy-
phosphonate 1 hydrolysis by using the acetonitrile–triethylamine–
water–iodine reagent system, some observations are pertinent.
Upon dissolving iodine in an acetonitrile : triethylamine : water
(2 : 1 : 1 v/v) mixture, a rapid change in colour from a deep-
violet to light-straw occurred, indicating apparently the
consumption of iodine. A plausible explanation for this pheno-
menon could be the known disproportionation of iodine under
basic aqueous conditions (I2 + 2HOꢁ " IOꢁ + Iꢁ + H2O)
to hypoiodites, which would then slowly disproportionate
Hydrolysis of a-hydroxyphosphonates 1 in
acetonitrile–triethylamine : water (2 : 1 : 1 v/v) containing
iodine (3 molar equiv.)
ꢁ
further to iodates and iodides (3IOꢁ " IO3 + 2Iꢁ).13 In
light of this, our reagent system might contain a signifi-
cant concentration of hypoiodite anions that are effective
a-nucleophiles towards the phosphorus centers. Thus, a com-
bination of efficient a-nucleophiles (hypoiodite) with an intra-
molecular acid catalysis provided by the adjacent hydroxyl
function, could constitute a plausible mechanistic basis for
the observed rapid hydrolysis of hydroxyphosphonates 1 using
the acetonitrile–triethylamine–water–iodine reagent system
(Scheme 5).14
As mentioned above, we attempted to use the reagent system,
acetonitrile : triethylamine : water (2 : 1 : 1 v/v) containing iodine,
to kinetically quench the equilibrium 1 " 2 + 3 (Scheme 1),
to obtain more quantitative data on this process. Somewhat
unexpectedly, although upon treatment of hydroxyphospho-
nates 1 with this reagent system we observed the formation
phosphodiesters 6 (due to oxidation of the corresponding
in situ-generated H-phosphonate diesters 2; Scheme 1), the
disappearance of the starting hydroxyphosphonates 1 was
significantly faster than that in the absence of iodine; in all
Similarly, as discussed above, the catalytic importance of a
free hydroxyl group in hydroxyphosphonate 1 hydrolysis was
Scheme 3 Possible mechanisms of nucleoside a-hydroxyphosphonate 7 formation (path B).
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 976–983 | 979

View Article Online
formed during the hydrolysis of hydroxyphosphonates 1
(Scheme 1). For this reason, we attempted to develop the
observed highly selective hydrolysis of 1 in the acetonitrile–
triethylamine–water–iodine reaction system via path B into a
synthetic protocol for the preparation of hydroxyphosphonate
monoesters 7. To this end, hydrolyses of hydroxyphosphonates
1a–g were undertaken on a preparative scale in the presence of
iodine, and in all instances, high isolated yields (75–85%) of
hydroxyphosphonates 7 were achieved.
Scheme 4 A phosphonate–phosphate rearrangement of aryl nucleo-
side a-hydroxyphosphonates.
demonstrated by carrying out the hydrolysis of O-acetylated
hydroxyphosphonate 9 using the acetonitrile–triethylamine–
water–iodine reagent system. As is apparent from the data in
Table 1, similar reaction times were observed for the hydro-
lysis of O-acylated hydroxyphosphonate 9 in both reac-
tion systems (with and without iodine added), and this
again (vide supra) emphasises the kinetic importance of a free
hydroxyl function in the substitution of OAr1 groups in
hydroxyphosphonates 1.
To find out if it was possible to further increase the selectivity
of formation of hydroxyphosphonates 7, we carried out the
hydrolysis of 1c in the presence of excess of p-anisaldehyde to
suppress the 1 " 2 + 3 equilibrium. Indeed, the selectivity of
hydrolysis of 1c via pathway B improved under such conditions,
and reached 98% when p-anisaldehyde was used in a large
excess (30 equiv.). It seems that such a modification of the
synthetic protocol for the preparation of 7 can be used in
instances when starting hydroxyphosphonates 1 are involved
in unfavourable equilibria.
To substantiate the assumption about a putative involvement
of powerful a-nucleophiles (hypoiodites) in hydroxyphosphonate
1 hydrolysis using the acetonitrile–triethylamine–water–iodine
reagent system, we carried out experiments in which another
a-nucleophile, a hydrogen peroxide ion, was used to effect the
hydrolysis of hydroxyphosphonate 1a. To this end, in the solvent
system acetonitrile–triethylamine–water (2 : 1 : 1, v/v/v), the
water was replaced with an equivalent volume of 30% hydrogen
peroxide, and this was used instead as the reaction medium for
the hydrolysis of 1a. As expected, the hydrolysis of 1a under
these conditions was very rapid, and the first 31P NMR
spectrum recorded (o5 min) showed the complete disappear-
ance of starting material 1a, together with the formation of the
corresponding hydroxyphosphonate monoester 7a (60%),
H-phosphonate monoester 4 (20%) and an unidentified side-
product (dP 4.35, m, no P–H bond; ca. 20%). We did not
observe any potential oxidation products derived from the
phosphorus compounds, which are known to be prone to
oxidation with hydrogen peroxide (e.g. aryl nucleoside
H-phosphonate of type 2). Thus, hydrogen peroxide in the
investigated reactions acted mainly as an a-nucleophile and
not as an oxidant. The above results strongly suggest that the
hydrolysis of hydroxyphosphonates 1 is significantly accele-
rated by a-nucleophiles, and thus lends support to a hypo-
thesis about the involvement of hypoiodites in the previously
investigated hydrolytic reactions.
Conclusions
In conclusion, we have investigated the stability and decompo-
sition pathways of diastereomeric mixtures of nucleoside
a-hydroxyphosphonates 1 bearing different aryl groups, both
in the ester and the hydroxymethine fragment. We have found
essential differences in products distributions during hydro-
lysis under previously described physiological-like conditions
(RPMI/FBS)2 and those well defined chemical conditions used
in the present studies. We have found that under aqueous
basic conditions, the stability and hydrolytic decomposition
pathways of these compounds were governed by the electronic
nature of the aryl group in the hydroxymethine moiety of
a-hydroxyphosphonates of type 1. The more electron-donating
is Ar2, the higher the participation of path A, due to equilibria
of a-hydroxyphosphonates with the corresponding aryl nucleo-
side H-phosphonate diesters and their respective aldehydes.
The hydrolysis of compounds of type 1 was found to be very
sensitive to the presence of a-nucleophiles, e.g. hypoiodite or
peroxide anions, and under such conditions, reactions towards
nucleoside a-hydroxyphosphonates monoesters of type 7 were
favoured. Significant differences in rates of hydrolysis of
hydroxyphosphonates
1 vs. their O-acylated derivatives
pointed to the importance of the intramolecular acid catalysis
exerted by the adjacent hydroxyl function. A preferential
hydrolysis of 1 via path B in the presence of a-nucleophiles
(hypoiodites) was exploited for the development of a simple
and efficient protocol for the preparation of otherwise difficult
to access hydroxyphosphonate monoesters of type 7, whose
biological properties are under study in our laboratory.
Synthesis of nucleoside a-hydroxyphosphonate monoesters 7
from the corresponding aryl nucleoside a-hydroxyphosphonate
diesters 1
Nucleoside a-hydroxyphosphonate monoesters 7 are often
difficult to access from their synthetic precursors 1 since, due
to the 1 " 2 + 3 equilibrium, other products are usually
Experimental
General methods and materials
1H, 13C and 31P NMR spectra were recorded on 300 or 400 MHz
machines. The 31P NMR experiments were carried out in
5 mm tubes using 0.1 M solutions of the phosphorus-containing
compounds. 31P NMR chemical shifts are reported in ppm
Scheme
5 Hypoiodite-promoted hydrolysis of aryl nucleoside
a-hydroxyphosphonates 1.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
980 | New J. Chem., 2010, 34, 976–983

View Article Online
relative to 85% H3PO4 in water (external standard). Mass
spectra were recorded by the liquid secondary ion mass
technique (LSIMS) using Cs+ (12 keV) for ionization. The
amount of water in solvents was measured by Karl Fisher
coulometric titration. Methylene chloride was dried over P2O5,
distilled and kept over molecular sieves 4 A until the amount
of water was less than 10 ppm. Pyridine was stored over
molecular sieves 4 A until the amount of water was below
20 ppm. For column chromatography, Kieselgel 60 Merck was
used. For TLC analysis, pre-coated plates (Merck silica gel
163.65 (C-4). HRMS [M]ꢁ m/z: 548.1089 calculated for
C23H24N5O7P 548.1102.
General procedure for synthesis of nucleoside a-
hydroxyphosphonates of type 7
Aryl nucleoside a-hydroxyphosphonate 1 (1 molar equiv.,
0.5 mmol) was dissolved in a freshly prepared solution con-
taining iodine (3 molar equiv.) in acetonitrile : triethylamine :
water (2 : 1 : 1 v/v) (5 mL) (straw-coloured). After the reaction
was complete (o5 min, 31P NMR), the solvents were evapo-
rated and the residue, dissolved in methylene chloride, applied
to a silica gel column pre-formed with iso-propanol. Products
7 were isolated using a stepwise gradient of water (0–10%) in
iso-propanol containing triethylamine (3% v/v). Fractions
containing pure products 7 were collected and evaporated
to dry foams. After freeze-drying from benzene–methanol, com-
pounds 7a–e were obtained as white amorphous solids (yellow
for the p-nitrophenyl derivative).
F
254) were used.
Aryl nucleoside a-hydroxyphosphonates of type 1 were
obtained as a mixture of four diastereoisomers by conden-
sation of AZT H-phosphonate 4 with the respective phenol 5,
followed by the addition of aldehyde 3 (the analytical data for
compounds 1a–e were reported previously2).
The reference compounds used for the identification of
certain reaction products or intermediates were obtained as
follows: AZT H-phosphonate 4, from the reaction of 30-azido-
30-deoxythymidyne with pyrophosphonic acid in pyridine;15
phosphotriester 8 from the condensation of phosphodiester 6
with 4-nitrobenzyl alcohol, aided by 2,4,6-triisopropylbenzene-
sulfonyl chloride in methylene chloride in the presence of
N-methylimidazole.16
30-Azido-30-deoxythymidin-50-yl a-hydroxy(phenyl)methyl-
phosphonate triethylammonium salt (7a). 0.21 g (yield 76%).
31P NMR (CDCl3) dP 16.84, 17.06. 1H NMR (CDCl3) dH 1.13,
1,16, 1.18 (t, J = 7.2 Hz, 9H, CH2CH3), 1.93, 1.94 (2s, 3H,
5-CH3), 2.22–2.34 (m, 2H, 20,20 0-H), 2.82, 2.84, 2.86, 2.89
(q, J = 7.2 Hz, 6H, CH2CH3), 3.98–4.40 (3m, 4H, 40, 30, 50,
50 0-H), 4.88, 4.91, 4.92, 4.95 (4s, 1H, a-H), 6.14, 6.16, 6.18,
6.20 (4s, 1H, 10-H), 7.17–7.35 (m, 3H, 6-H and ArH),
7.54–7.56 (m, 2H, ArH), 7.70–7.73 (m, 1H, ArH). 13C NMR
(DCl3) dC 8.44 (CH2CH3), 12.45 (5-CH3), 37.72, 37.80 (C-20),
45.27 (CH2CH3), 60.12–61.66 (C-30), 64.76, 64.83, 65.04, 65.12
(C-50), 70.56, 70.78, 72.48, 72.71 (a-C–OH), 83.88, 83.93,
83.97, 84.01 (C-40), 85.09, 85.22 (C-10), 110.10-111.11 (C-5),
114.54, 114.59, 114.69, 114.75 (C-Ar), 120.40 (C-Ar),
121.11–121.23 (C-Ar), 126.65–128.32 (5 ꢂ C-Ar), 135.82,
135.88 (C-Ar), 140.48, 140.72 (C-6), 150.68, 150.72 (C-2),
164.44–164.60 (C-4). HRMS [M–Et3NH+]ꢁ m/z: 436.1032
calculated for C17H219N5O7P 436.1022.
The purity of all compounds was 498%, as judged from
1H NMR spectra. The multiplicity of signals in 31P NMR
spectra were due to the presence of diastereoisomers.
Syntheses
30-Azido-30-deoxythymidin-50-yl 4-methoxyphenyl a-hydroxy-
(phenyl)methylphosphonate 1f. Obtained as previously described2 -
0.19 g (yield 72%). 31P NMR (CDCl3) dP 18.94, 18.97, 19.21,
19.31. 1H NMR (CDCl3) dH 1.76, 1.77, 1.79, 1.83 (4s, 3H,
5-CH3), 1.96–2.23 (2m, 2H, 20,200-H), 3.70, 3.71, 3.72 (3s, 3H,
OCH3), 3.88–3.93 (m, 1H, 40-H), 4.04–4.20 (m, 1H, 30-H),
4.21–4.26 (m, 2H, 50,50 0-H), 5.22–5.25 (m, 1H, a-H), 5.60–6.06
(m, 1H, 10-H), 6.74–7.51 (m, 10H, 6-H and ArH). 13C NMR
(CDCl3) dC 12.27 (5-CH3), 36.68 (C-20), 55.54, 55.57 (OCH3),
60.12–60.26 (C-30), 66.45 (C-50), 69.50, 69.68, 71.07, 71.23
(a-C–OH), 82.12, 82.18 (C-40), 85.49 (C-10), 111.41–111.52
(C-5), 114.54, 114.59, 114.69, 114.75 (C-Ar), 120.40 (C-Ar),
121.11–121.23 (C-Ar), 126.99–127.25 (C-Ar), 128.31–128.72
(2 ꢂ C-Ar), 135.40–135.86 (C-6), 143.34–143.51 (C-Ar), 150.22,
150.27 (C-2), 156.93 (C-Ar), 163.65–163.71 (C-4). HRMS [M]ꢁ
m/z: 544.1586 calculated for C24H27N5O8P 544.1597.
30-Azido-30-deoxythymidin-50-yl a-hydroxy(4-methylphenyl)-
methylphosphonate triethylammonium salt (7b). 0.24 g (yield
1
87%). 31P NMR (CDCl3) dP 17.04, 17.25. H NMR (CDCl3)
dH 1.14, 1.16, 1.18 (t, J = 7.2 Hz, 9H, CH2CH3), 1.93 (s, 3H,
5-CH3), 2.30 (s, 3H, CH3), 2.22–2.32 (m, 2H, 20,200-H), 2.84,
2.86, 2.87, 2.89 (q, J = 7.2 Hz, 6H, CH2CH3), 3.94–4.21 (3m,
3H, 30, 50, 50 0-H), 4.34–4.37 (m, 1H, 40-H), 4.84, 4.86, 4.87,
4.90 (4s, 1H, a-H), 6.16, 6.18, 6.19, 6.20 (4s, 1H, 10-H), 7.09,
7.11 (2s, 2H, ArH), 7.42, 7.44 (2s, 2H, ArH), 7.70, 7.72
(2s, 1H, 6-H). 13C NMR (CDCl3) dC 8.48 (CH2CH3), 12.44
(5-CH3), 21.03 (CH3), 37.68, 37.74 (C-20), 45.22 (CH2CH3),
61.65 (C-30), 64.82, 65.10 (C-50), 70.71, 70.90, 72.15, 72.35
(a-C–OH), 83.92 (C-40), 85.02, 85.11 (C-10), 110.01, 111.10
(C-5), 126.59, 126.64, 126.75, 126.80 (2 ꢂ C-Ar), 128.44
(C-Ar), 135.87 (2 ꢂ C-Ar), 136.22 (C-Ar), 137.29, 137.51
(C-6), 150.69 (C-2), 164.40, 164.52 (C-4). HRMS [M–Et3NH+]ꢁ
m/z: 450.1174 calculated for C18H21N5O7P 450.1179.
30-Azido-30-deoxythymidin-50-yl 4-chlorophenyl a-hydroxy-
(phenyl)methylphosphonate 1g. Obtained as previously described2-
0.22 g (yield 80%). 31P NMR (CDCl3) dP 18.71, 19.11. 1H NMR
(CDCl3) dH 1.81, 1.82, 1.83, 1.84 (4s, 3H, 5-CH3), 2.08–2.38
(m, 2H, 20,200-H), 3.90–3.98 (m, 1H, 40-H), 4.12–4.38 (m, 3H, 30,
50, 500-H), 5.20–5.27 (m, 1H, a-H), 5.99–6.10 (m, 1H, 10-H),
6.98–7.50 (m, 10H, 6-H and ArH). 13C NMR (CDCl3) dC 12.32
(5-CH3), 36.86, 36.96, 37.06 (C-20), 60.03 (C-30), 65.82–66.21
(C-50), 69.79,69.88 71.47 (a-C–OH), 82.12–82.33 (C-40), 85.48–
85.70 (C-10), 111.41–111.51 (C-5), 121.65, 121.69 (C-Ar),
30-Azido-30-deoxythymidin-50-yl a-hydroxy(4-methoxyphenyl)-
methylphosphonate ammonium salt (7c). 0.18 g (yield 75%).
31P NMR (CDCl3 + CD3OD) dP 17.80, 17.83. 1H NMR
(CDCl3 + CD3OD) dH 1.80 (s, 3H, 5-CH3), 2.08–2.23 (m, 2H,
126.97–127.82 (2
ꢂ
C-Ar), 128.31–128.94 (2
ꢂ
C-Ar),
129.64–130.11 (C-Ar), 135.10, 135.21 (C-6), 135.74, 135.84,
135.97 (C-Ar), 148.57–148.66 (C-Ar), 150.13 (C-2), 163.58,
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 976–983 | 981

View Article Online
20,20 0-H), 3.70 (s, 3H, OCH3), 3.77–3.96 (m, 3H, 30, 50, 50 0-H),
4.13–4.15 (m, 1H, 30-H), 4.72, 4.73, 4.75, 4.76 (4s, 1H, a-H),
6.01, 6.03, 6.05 (3s, 1H, 10-H), 6.77, 6.79 (2s, 2H, ArH), 7.31,
7.33 (2s, 2H, ArH), 7.40, 7.41, 7.42, 7.43 (4s, 1H, 6-H).
13C NMR (D2O) dC 11.55, 11.59 (5-CH3), 36.27, 36.36
(C-20), 55.26 (OCH3), 60.36, 60.45 (C-30), 64.02, 64.08,
64.26, 64.32 (C-50), 69.87, 70.44, 71.44, 72.01 (a-C–OH),
83.22, 83.29 (C-40), 84.90, 85.07 (C-10), 110.37, 111.45 (C-5),
128.49, 128.55, 128.71, 128.77 (2 ꢂ C-Ar), 131.28, 131.30,
131.34, 131.36 (2 ꢂ C-Ar), 137.18, 137.29 (C-6), 151.48, 151.55
for the parent aryl nucleoside a-hydroxyphosphonate 1c.
Obtained 0.13 g (yield 75%). 31P NMR (CDCl3) dP 15.07,
15.11, 15.44, 15.46. H NMR (CDCl3) dH 1.84, 1.86 (2s, 3H,
1
5-CH3), 2.01, 2.06, 2.15, 2.17 (4s, 3H, CH3COO), 1.94–2.37
(m, 2H, 20,20 0-H), 3.78, 3.79, 3.80 (3s, 3H, OCH3), 3.89–4.32
(4m, 4H, 40, 30, 50, 50 0-H), 6.06–6.21 (m, 1H, a-H), 6.27–6.35
(m, 1H, 10-H), 6.87–7.49 (m, 10H, 6-H and ArH). 13C NMR
(CDCl3) dC 12.34, 12.42 (5-CH3), 20.62, 20.67, 20.78
(CH3COO), 36.98, 37.06, 37.12, 37.19 (C-20), 55.32 (OCH3),
60.02, 60.08 (C-30), 65.77, 65.84, 65.92, 65.99, 66.32, 66.42,
66.60, 66.69 (C-50), 68.56, 68.67, 68.72, 68.89, 70.88, 71.01,
71.20 (a-C–OH), 81.91, 81.99, 82.06 (C-40), 84.51, 84.57, 84.76,
84.84 (C-10), 111.44, 111.46, 111.53, 111.56 (C-5), 120.02,
120.09, 120.14, 120.18, 120.24 (C-Ar), 129.45–130.08 (5 ꢂ
C-Ar), 134.94, 135.06, 135.17 (C-6), 149.77–150.26 (C-Ar,
C-2), 160.37–160.54 (C-4), 168.94–169.21 (CH3COO). HRMS
[M]ꢁ m/z 586.1711 calculated for C26H29N5O9P 586.1703.
(C-2), 158.28, 158.31, 158.34 (2 ꢂ C-Ar), 166.38, 166.39
+ ꢁ
m/z: 466.1144 calculated for
(C-4). HRMS [M–NH4
C18H21N5O8P 466.1128.
]
30-Azido-30-deoxythymidin-50-yl a-hydroxy(4-chlorophenyl)-
methylphosphonate triethylammonium salt (7d). 0.25 g (yield
1
88%). 31P NMR (CDCl3) dP 16.10, 16.30. H NMR (CDCl3)
dH 1.16, 1.18, 1.21 (t, J = 7.2 Hz, 9H, CH2CH3), 1.92, 1.93,
1.94 (3s, 3H, 5-CH3), 2.30 (s, 3H, CH3), 2.26–2.38 (m, 2H,
20,20 0-H), 2.88, 2.90, 2.93, 2.95 (q, J = 7.2 Hz, 6H, CH2CH3),
3.94–4.30 (3m, 3H, 30, 50, 50 0-H), 4.37–4.41 (m, 1H, 40-H), 4.85,
4.88, 4.90, 4.92 (4s, 1H, a-H), 6.12, 6.15, 6.17, 6.19 (4s, 1H,
10-H), 7.27, 7.28 (2s, 2H, ArH), 7.47–7.52 (m, 2H, ArH), 7.66,
7.69 (2s, 1H, 6-H). 13C NMR (CDCl3) dC 8.41 (CH2CH3),
12.42 (5-CH3), 37.77, 37.86 (C-20), 45.51 (CH2CH3), 61.55
(C-30), 64.87, 65.11, 65.19 (C-50), 69.59, 70.13, 71.86, 72.05
(a-C–OH), 83.89 (C-40), 85.22, 85.36 (C-10), 110.96, 111.05
(C-5), 127.80 (C-Ar), 128.02, 128.08 (C-Ar), 128.22, 128.28
(C-Ar), 132.36 (C-Ar), 135.71, 135.80 (C-Ar), 139.02, 139.32 (C-6),
150.64 (C-2), 164.44, 164.61 (C-4). HRMS [M–Et3NH+]ꢁ m/z:
470.0646 calculated for C17H18N5O7P 470.0632.
30-Azido-30-deoxythymidin-50-yl a-acetyloxy(4-methoxyphenyl)-
methylphosphonate triethylammonium salt (10). 0.14 g of 7c
dissolved in 3 mL of methylene chloride containing pyri-
dine (10% v/v) was treated with acetic anhydride (0.28 mL,
10 molar equiv.) at room temperature. After the reaction was
complete (0.5 h, 31P NMR), the solvent was evaporated and
compound 10, after dissolving in a small volume of methylene
chloride, precipitated from n-hexane. 0.16 g (yield 88%) of
white solid. 31P NMR (CDCl3) dP 13.08, 13.13. 1H NMR
(CDCl3) dH 1.13, 1.16, 1.18 (t, J = 7.2 Hz, 9H, CH2CH3),
1.90, 1.94, 2.03 (3s, 3H, CH3COO), 2.09, 2.10 (2s, 3H, 5-CH3),
2.23–2.31 (m, 2H, 20,200-H), 2.82, 2.85, 2.87, 2.91 (q, J = 7.2 Hz,
6H, CH2CH3), 3.76 (s, 3H, OCH3), 3.96–4.30 (3m, 4H, 40, 30,
50, 50 0-H), 6.02, 6.05, 6.06, 6.09 (4s, 1H, a-H), 6.20-6.24
(m, 1H, 10-H), 6.81, 6.84 (2s, 2H, ArH), 7.41, 7.44 (2s, 2H,
ArH), 7.56, 7.59 (2s, 1H, 6-H). 13C NMR (CDCl3) dC 8.08,
8.34 (CH2CH3), 12.30, 12.36 (5-CH3), 20.98, 21.03, 21.05
(CH3COO), 37.24 (C-20), 45.39 (CH2CH3), 55.25, 55.28
(OCH3), 60.59, 60.66 (C-30), 64.77, 64.83, 65.10, 65.16
(C-50), 70.45, 70.61, 72.08, 72.24 (a-C–OH), 83.02, 83.09
(C-40), 84.53 (C-10), 111.18, 111.21 (C-5), 113.59 (2 ꢂ C-Ar),
126.07 (C-Ar), 127.54, 127.61 (C-Ar), 128.29 (C-Ar), 129.25,
129.29 (C-6), 150.31 (C-2), 159.47, 159.48 (C-Ar), 163.89
(C-4), 169.68, 169.76 (CH3COO). HRMS [M–Et3NH+]ꢁ m/z
508.1229 calculated for C20H23N5O9P 508.1234.
30-Azido-30-deoxythymidin-50-yl a-hydroxy(4-nitrophenyl)-
methylphosphonate triethylammonium salt (7e). 0.25 g (yield
1
85%). 31P NMR (CDCl3) dP 15.09, 15.31. H NMR (CDCl3)
d
H 1.15, 1.17, 1.20 (t, J = 7.2 Hz, 9H, CH2CH3), 1.93 (3s, 3H,
5-CH3), 2.28–2.41 (m, 2H, 20,200-H), 2.90, 2.93, 2.95, 2.98
(q, J = 7.2 Hz, 6H, CH2CH3), 3.91–4.30 (3m, 3H, 30, 50,
50 0-H), 4.37–4.40 (m, 1H, 40-H), 5.00, 5.03, 5.06, 5.08 (4s, 1H,
a-H), 6.08, 6.10, 6.13, 6.15 (4s, 1H, 10-H), 7.62, 7.63, 7.65, 7.66
(2d, J = 0.9, 1.2 Hz, 1H, 6-H), 7.73–7.78 (m, 2H, ArH),
8.15–8.18 (m, 2H, ArH). 13C NMR (CDCl3) dC 8.37
(CH2CH3), 12.40 (5-CH3), 37.78, 37.87 (C-20), 45.43
(CH2CH3), 61.39 (C-30), 64.92, 65.00, 65.05, 65.27, 65.30
(C-50), 70.33, 70.41, 72.21, 72.29 (a-C–OH), 83.72, 83.75,
83.81, 83.83 (C-40), 85.32, 85.44 (C-10), 110.96, 111.03
(C-5), 120.08, 120.16 (C-Ar), 127.25, 127.30, 127.43, 127.49
(2 ꢂ C-Ar), 135.65, 135.75 (C-6), 146.72, 146.76 (C-Ar),
148.38, 148.40 (C-Ar), 148.71 (C-Ar), 150.65, 150.67 (C-2),
164.58, 164.73 (C-4). HRMS [M–Et3NH+]ꢁ m/z 481.0888
calculated for C17H18N6O9P 481.0873.
Acknowledgements
Financial support from the government of the Polish Republic
(partial project no. PBZ-MNiSW-07/I/2007; 2008–2010) is
gratefully acknowledged.
References
30-Azido-30-deoxythymidin-50-yl phenyl a-acetyloxy(4-methoxy-
phenyl)methylphosphonate (9). 0.16 g of 1c dissolved in 3 mL of
methylene chloride containing pyridine (10% v/v) was treated
with acetic anhydride (0.28 mL, 10 molar equiv.) at room
temperature. After the reaction was complete (2.5 h, 31P
NMR), the resulting mixture was washed with water, the
organic layer separated after drying over anhydrous Na2SO4
and evaporated. The further isolation procedure for 9 was as
1 For the most recent review, see: S. J. Hecker and M. D. Erion,
J. Med. Chem., 2008, 51, 2328–2345.
2 A. Szymanska, M. Szymczak, J. Boryski, J. Stawinski,
´
A. Kraszewski, G. Collu, G. Sanna, G. Giliberti, R. Loddo and
P. La Colla, Bioorg. Med. Chem., 2006, 14, 1924–1934.
3 RPMI multicomponent aqueous solution: Sigma cat. no. R8758;
foetal bovine serum (FBS): Sigma cat. no. F6178.
4 C. Meier, L. W. Habel, J. Balzarini and E. De Clercq, Liebigs Ann.
Chem., 1995, 2195–2202.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
982 | New J. Chem., 2010, 34, 976–983

View Article Online
5 H. McCombie and G. J. Stacey, J. Chem. Soc., 1945, 921–922.
6 P. J. Garegg, T. Regberg, J. Stawinski and R. Stromberg,
Nucleosides, Nucleotides Nucleic Acids, 1987, 6, 429–432.
7 C. Meier, Angew. Chem., Int. Ed. Engl., 1993, 32, 1704–1706.
8 F. Hammerschmidt, E. Schneyder and E. Zbiral, Chem. Ber., 1980,
113, 3891–3897.
9 Due to the presence of a highly strained 3-membered ring, the
formation of oxaphosphirane III proceeds most likely via the initial
formation of phosphorane II, bearing an oxaphosphirane ring in an
apical–equatorial position, followed by a pseudorotation with expul-
sion of the leaving group OAr1. Compound III may then collapse to
hydroxyphosphonate 7 via an intermediate phosphorane, IV, which
may be considered as a transition state or a transient compound.
10 W. F. Barthel, B. H. Alexander, P. A. Giang and S. A. Hall, J. Am.
Chem. Soc., 1955, 77, 2424–2427.
12 F. Hammerschmidt and S. Schmidt, Eur. J. Org. Chem., 2000,
2239–2245.
13 N. N. Greenwood and A. Earnshaw, Chemistry of the Elements;
Pergamon Press, Oxford, 1984, p. 999.
14 Since the pH of the reagent system solution was lower than that of
the system without iodine (ca. 10.4 vs. 12.5), there was a remote
possibility that this might affect the observed rates of hydrolysis.
To address this issue, the hydrolysis of reference a-hydroxy-
phosphonate 1a was carried out in an acetonitrile: water (2: 1 v/v)
solvent system adjusted to pH 10.4 by the addition of triethyl-
amine. It was found that under such reaction conditions, the
hydrolysis of 1a was very slow. After 24 h, 31P NMR spectroscopy
still revealed the presence of the unreacted starting material.
15 J. Stawinski and M. Thelin, Nucleosides, Nucleotides Nucleic Acids,
1990, 9, 129–135.
11 A. F. Janzen and O. C. Vaidya, Can. J. Chem., 1973, 51,
1136–1138.
16 V. A. Efimov, S. V. Reverdatto and O. G. Chakhmakhcheva,
Nucleic Acids Res., 1982, 10, 6675–6694.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 976–983 | 983