Novel approach to nucleoside-5′-(1-hydroxymethylene-1, 1-bisphosphonates): Synthesis of new AZT analogues

Article abstract of DOI:10.1081/NCN-200051890

An efficient synthetic method of nucleoside-5′-(1-hydroxymethylene-1, 1-bisphosphonates) is reported here. The procedure was optimized with 3′-protected thymidine and then applied to synthesis of new AZT analogues. Copyright

Full text of DOI:10.1081/NCN-200051890

Nucleosides, Nucleotides, and Nucleic Acids, 24 (2):121–133, (2005)
Copyright D 2005 Taylor & Francis, Inc.
ISSN: 1525-7770 print/ 1532-2335 online
DOI: 10.1081/NCN-200051890
NOVEL APPROACH TO NUCLEOSIDE-5’-(1-HYDROXYMETHYLENE-1,
1-BISPHOSPHONATES): SYNTHESIS OF NEW AZT ANALOGUES#
Evelyne Migianu, Ma¨elle Monteil, Pascale Even, and Marc Lecouvey
5
Laboratoire de Chimie Structurale Biomoleꢀculaire (UMR 7033-CNRS), UFR S.M.B.H.,
UniversiteꢀParis 13, Bobigny Cedex, France
5
An efficient synthetic method of nucleoside-5’-(1-hydroxymethylene-1,1-bisphosphonates) is reported
here. The procedure was optimized with 3’-protected thymidine and then applied to synthesis of new
AZT analogues.
Keywords Bisphosphonate, AZT-H-phosphonate, Thymidine-H-phosphonate
INTRODUCTION
Human immunodeficiency virus type 1 (HIV-1) is the causative retrovirus of
acquired immunodeficiency syndrome (AIDS)[1] and requires reverse transcriptase
(RT) enzyme to copy its single-stranded RNA genome into a double-stranded DNA
copy for integration into the host cell genome. Several members of HIV inhibitors
are currently approved for the treatment of HIV-infected patients.[1,2] 3’-Azido-3’-
deoxythymidine (AZT), a nucleoside reverse transcriptase inhibitor (NRTI), is one
of them. It is the first clinically successful drug used for AIDS. It continues to play
an important role in chemotherapy,[3,4] particularly in combination with other
NRTIs, non-NRTIs, and HIV protease inhibitors, despite its toxicity to the host and
the ability of the virus to mutate and gain resistance. Indeed, many undesirable side
effects such as bone marrow suppression, myopathy, hepatic abnormalities,[5]
neutropenia, and anemia[6–10] are frequently observed clinically. Thus, great efforts
have been continuously made to develop and optimize the structure of AZT.
#We thank Pr. J. Liquier from the Laboratoire de Chimie Structurale Biomol´eculaire (UMR 7033-CNRS,
Bobigny, France) for IR spectra, the Laboratoire de Biochimie des Prot´eines et Prot´eomique (UPRES EA 3408,
Bobigny, France) for mass spectra, and Dr. P. Nicolas from the Service de Pharmacologie de l’Hoˆpital Avicenne
(Bobigny, France) for his kind gift of AZT.
Received 21 September 2004, accepted 17 December 2004.
Address correspondence to Marc Lecouvey, Laboratoire de Chimie Structurale Biomol´eculaire (UMR 7033-
CNRS), UFR S.M.B.H., Universit´e Paris 13, 74 Rue Marcel Cachin, Bobigny Cedex, F-93017 France; Fax: 00-33-1-
48-38-76-25; E-mail: m.lecouvey@smbh.univ-paris13.fr
121
Order reprints of this article at www.copyright.rightslink.com

122
E. Migianu et al.
As other nucleoside analogues, AZT does not exert antiviral activity directly,
but it acts rather like a prodrug of active phosphorylated metabolites. It must be
successively phosphorylated in position 5’ to the corresponding mono-, di-, and
triphosphates by thymidine kinases[11–14] to be recognized by HIV-1 RT. Thus,
it terminates the viral DNA chain replication, since modified nucleosides do not
possess the 3’-hydroxyl group necessary for chain elongation.[14–16] Unfortunately,
some 2’,3’-dideoxynucleoside-5’-monophosphates are known to have a low anti-
HIV activity[17] as some AZT analogues.[18] Moreover, AZT-5’-triphosphate (TP) is
poorly absorbed when administered externally and readily dephosphorylated in the
extracellular environment.[19] However, these drawbacks can be avoided by using
an analogue of AZT that could serve as an intracellular slow-release depot to the
active AZT-5’-TP. Hence, we propose to synthesize AZT-bisphosphonates (BPs) as
potential prodrugs of AZT-5’-TP. Previously, it was been shown that bisphos-
phonates were metabolized intracellularly by dictyostelium amoebae into methylene-
containing adenine nucleotides.[20] These AZT analogues could act as competitive
inhibitors, preventing the incorporation of natural RT substrates. It would consist of
AZT-5’-hydroxymethylene bisphosphonates (HMBPs) substituted with alkyl or aryl
groups on the central carbon. This structural modification replaces a diphosphate
moiety by a dephosphorylation-resistant methylenebisphosphonate group. It could
increase lipophilicity of AZT analogues due to the presence of an aliphatic or an
aromatic group. Moreover, these compounds that might still be substrates for HIV-
1 RT could be encapsulated in liposomes as described previously with BPs[21] in
order to cross cellular membranes more efficiently and reach intracellular targets.
To reduce toxicity of AZT, increase its anti-HIV activity, and cause fewer
nucleoside-resistant mutants of HIV, a lot of work has been reported on devel-
opment of 5’-O-ester prodrugs of AZT.[3] The 5’-H-phosphate of AZT (PZT)
especially is one of the most significant compounds.[22] Indeed, PZT, which is less
active as an anti-HIV agent than AZT itself, is nevertheless far less toxic (CC50
values of 2.5 mM and 210 mM for PZT and AZT, respectively). The overall
selectivity for PZT (CC50/IC50) is superior to that of AZT,[23] and now PZT is
currently in clinical trials. PZT, unlike highly acidic nucleoside 5’-phosphates, which
cannot enter the cell, may penetrate the cell membrane due to its weakly acidic
SCHEME 1 General procedure for synthesis of HMBP monoalkyl esters.

Novel Approach to Nucleoside-5’-(1-Hydroxymethylene-1,1-Bisphosphonates)
123
FIGURE 1 Structures of 3’-O-benzoyl thymidine-5’-HMBP 1 and AZT-5’-HMBPs 2.
undissociated nature. These good results encouraged us to develop the synthesis
of AZT-5’-HMBPs as new phosphorylated anti-HIV molecules in order to over-
come AZT’s drawbacks and improve its efficacy. These derivatives may be more
stable in vivo and so have a longer potential antiviral activity due to the resistance
of the P–C–P bond to chemical and enzymatic cleavage.
We report herein a new efficient method for preparation of these AZT-5’-
HMBPs analogues. Contrary to previous works,[24] our strategy does not consist of
reaction between HMBP and AZT but in mono- and then diphosphorylation of the
nucleoside at the 5’-position. Our method is based on recent results obtained in our
laboratory from synthesis of HMBP monoalkyl esters[25] (Scheme 1). We first
describe the preliminary study with thymidine to prepare 3’-O-benzoyl thymidine-
HMBP 1, then synthesis of AZT analogues 2 (Figure 1). Stereochemistry will not
be discussed here.
RESULTS AND DISCUSSION
Synthesis of 3’-O-Benzoyl Thymidine-5’-HMBP 1
The synthetic method of nucleoside 5’-HMBPs was first tested with thymidine.
Before 5’-phosphorylation, thymidine was protected at the 3’-position with a benzoyl
group (Bz) using intermediate dimethoxytrityl protection of 5’-OH, as described by
Shott et al.[26]
The first step consisted in the preparation of 3’-O-benzoyl thymidine-H-
phosphonate 5 (Scheme 2). This compound was obtained by a reaction reported
by Dreef et al.[27] between 3’-O-benzoyl thymidine 3 and 1.1 equivalents of 2-chloro-
benzodioxaphosphorin-4-one 4 in a mixture acetonitrile/pyridine (75:25) under
argon. The reaction mixture was then stirred one hour at room temperature before
hydrolysis for two hours with distilled water. After evaporation under vacuum and
washes of the crude product with diethyl ether in order to remove the formed
salicylic acid, the pyridinium salt of thymidine-H-phosphonate was precipitated in
cold methanol and filtered. It was then converted to the corresponding sodium salt

124
E. Migianu et al.
SCHEME 2 Preparation of 3’-O-benzoyl thymidine-H-phosphonate 5.
by stirring in distilled water with ion-exchange resin (Dowex Na+ 50 W Â 2) to give
a colorless solid with 60% yield. This monophosphate derivative 5 was identified by
1
means of NMR spectroscopies. The H NMR spectrum (in CD3OD) showed a
characteristic doublet at 6.92 ppm corresponding to the signal of the proton on the
1
phosphorus atom, with a JP-H coupling constant of 683.6 Hz. Moreover, the 31P
{1H} NMR spectrum presented a singlet for the signal of the phosphorus atom
about 8.1 ppm.
Next, thymidine-H-phosphonate 5 was silylated as described by Hata and
Sekine[28] using four equivalents of trimethylsilyl chloride and triethylamine in
pyridine under inert atmosphere (Scheme 3). The evolution of the reaction was
monitored by 31P {1H} NMR. After 1 h at room temperature, the 31P {1H} NMR
spectrum showed the appearance of a single peak around 117 ppm correspond-
ing to compound 6 and characteristic of a bis(silylated) phosphite. This inter-
mediate was not isolated but directly used in the last reaction. Indeed, silylated
a-ketophosphonate 7a was dissolved in chloroform and added to bis(silylated)
phosphite 6 at 0°C under argon. The reaction mixture was then stirred at room
temperature for 2 h. After a few hours of methanolysis and evaporation under
vacuum, 3’-O-benzoyl thymidine-5’-HMBP 1 was obtained. The crude product was
washed with diethyl ether and converted to the corresponding sodium salt by
stirring in distilled water with ion-exchange resin (Dowex Na+ 50 W Â 2) to give a
SCHEME 3 Synthesis of 3’-O-benzoyl thymidine-5’-HMBP 1.

Novel Approach to Nucleoside-5’-(1-Hydroxymethylene-1,1-Bisphosphonates)
125
colorless solid with 55% yield. One must note that silylated a-ketophosphonate 7a
was prepared according to the procedure described in Scheme 1. The purified
product 1 was characterized by NMR spectroscopies. As expected, because of the
presence of two new asymmetrical centers and superposition of signals of
enantiomers, the 31P {1H} NMR spectrum showed two separate doublets at 22.5
and 21.0 ppm, respectively, with equal 2JP-P coupling constants corresponding to the
first diastereoisomers. Two other different doublets were distinguished respectively
at 22.1 and 21.5 ppm corresponding to the second diastereoisomers, formed in the
same proportions. Signals about 22 ppm were assigned to phosphorus atom of
phosphonic acid moiety, whereas signals about 21 ppm corresponded to
phosphorus atom of phosphonate monoester fragment. On the other hand, the
13C {1H} NMR spectrum showed a characteristic signal at 74.3 ppm (dd,
1
1JC-P = 145.0 Hz and JC-P = 145.0 Hz) for the carbon atom of the bisphosphonate
moiety substituted by hydroxyl and a methyl group. Further characterizations were
made by running COSY and HSQC 2D experiments.
Synthesis of AZT-5’-HMBPs 2
AZT-5’-HMBPs analogues were prepared according the procedure described
above and perfected with thymidine. Indeed, AZT-H-phosphonate 9 was first
synthesized by reaction under argon between AZT 8 and 1.1 equivalents of
2-chloro-benzodioxaphosphorin-4-one 4 in a mixture THF/pyridine (75:25)
(Scheme 4). THF was preferred to acetonitrile for solubility reasons. The reaction
mixture was then stirred one hour at room temperature before hydrolysis for two
hours with distilled water. After evaporation under vacuum, the crude product was
washed with diethyl ether in order to remove the resulting salicylic acid. Next, the
pyridinium salt of AZT-H-phosphonate was converted to the corresponding sodium
salt by stirring in distilled water with ion-exchange resin (Dowex Na+ 50 W Â 2).
Then, compound 9 was purified by reverse-phase column chromatography using a
C-18 resin and distilled water as eluent. A colorless solid was obtained after
1
lyophilization with 90% yield. As observed with thymidine analogue 5, the H
and 31P {1H} NMR spectra in D2O showed the expected characteristic signals
of monophosphate 9 (dH = 6.81 ppm, doublet for H-phosphonate proton and
1JP-H = 638.8 Hz; dP = 7.2 ppm, singlet).
SCHEME 4 Preparation of AZT-H-phosphonate 9.

126
E. Migianu et al.
Attempts were undertaken in order to prepare the bis(silylated) analogue of
monophosphate 9 with trimethylsilyl chloride and triethylamine as previously done
with the thymidine derivative. Unfortunately, the expected intermediate was never
obtained. However, this reaction was possible using five equivalents of N,O-
bis(trimethylsilyl)acetamide (BSA) as reported in the literature by Sekine et al.[29]
Distilled pyridine was used as reaction solvent instead of acetonitrile as
recommended. Thus, phosphonate monoester 9 was dissolved in dry pyridine,
five equivalents of BSA were added at 0°C under argon and the reaction mix-
ture was then stirred at room temperature (Scheme 5). After 1 h, reaction com-
pletion was controlled by 31P {1H} NMR with the appearance of a single peak
(dP = 117.3 ppm) different from the starting product (dP = 7.2 ppm), characteristic
of the bis(silylated) phosphite 10. The last reaction between compound 10 and
silylated a-ketophosphonate 7a or 7b (1.1 equivalents) was carried out in the
same conditions as thymidine analogue 1, except chloroform was replaced by
pyridine. The reaction mixture was stirred at room temperature under argon
overnight. After a few hours of methanolysis, a precipitate appeared and was
filtered off. This product was identified by NMR spectroscopies as the expected
AZT-5’-HMBP analogue 2a or 2b and converted to the corresponding sodium salt
by stirring in distilled water with ion-exchange resin (Dowex Na+ 50 W Â 2) and
lyophilized. AZT-5’-HMBP analogues 2a and 2b were obtained as colorless
powders with 53% and 59% yields, respectively. The 31P {1H} NMR spectrum of
compound 2a (R = CH3) showed only two different doublets at 20.4 and 18.6 ppm
(2JP-P = 34.8 Hz), signals that could be attributed to the two superposed dia-
stereoisomers. The 31P {1H} NMR spectrum of product 2b (R = Ph) distinguished
the different diastereoisomers in 60:40 proportions. Indeed, the spectrum indi-
cated two separate doublets at 18.1 and 14.9 ppm (2JP-P = 27.7 Hz) for the first
diastereoisomers and two other ones at 17.5 and 15.0 ppm (2JP-P = 26.3 Hz) for the
1
second diastereoisomers. On the H NMR spectrum, diastereoisomers could not
SCHEME 5 Synthesis of AZT-5’-HMBPs 2.

Novel Approach to Nucleoside-5’-(1-Hydroxymethylene-1,1-Bisphosphonates)
127
be distinguished. Only one signal was observed for each proton. However, the
protons on carbon 2’ appeared as two different signals for analogue 2b.
Compounds were fully characterized by doing COSY experiments.
In conclusion, we have developed a new efficient synthetic method for
preparation of AZT-5’-HMBPs analogues from AZT via AZT-H-phosphonate.
Thus, two different aliphatic and aromatic HMBP derivatives have been
synthesized in good yields. These compounds will be studied in biological tests
in order to evaluate their anti-HIV activity and their toxicity.
EXPERIMENTAL SECTION
Unless otherwise noted, starting materials were purchased from commercial
suppliers. 3’-O-benzoyl thymidine 3 was synthesized according to the procedure
described by Shott et al.[26] Acetyl and benzoyl phosphonic acid dimethyl esters
were prepared from the corresponding acyl chlorides and trimethyl phosphite.[25]
Me3SiCl and Me3SiBr were distilled prior use. Commercial BSA was used
without previous purification. Pyridine and triethylamine were distilled over KOH.
CH3CN and CHCl3 were distilled over P2O5; CH2Cl2 and Et2O were distilled
over sodium and THF was distilled from benzophenone sodium. AZT was
obtained from Retrovir capsules (250 mg) and purified by solubilization in
EtOH/CHCl3 (75:25), filtration to remove excipients and evaporation of the
filtrate. Melting points are uncorrected. NMR spectra were recorded with a
VARIAN Unity Inova 500 MHz (1H: 500.6 MHz, 31P: 200.7 MHz, 13C: 125.9
MHz) or a VARIAN Gemini 200 MHz (1H: 200 MHz, 31P: 80.9 MHz, 13C:
50.3 MHz) spectrometer in CD3OD or D2O. Chemical shifts (d) were given in
1
ppm. H NMR spectra were recorded using trimethylsilane or HOD as internal
standards in CD3OD or D2O. 31P and 13C NMR spectra were recorded using,
respectively, phosphoric acid and methanol as external references. 1D experiments
were performed at 20°C in HDO or in CDCl3. To suppress water signal, pre-
saturation in proton spectrum was made with a duration of 1.5 s after relaxation
delay. On 6000 Hz, data were collected with a hard pulse proton of 8.2 ms and
a power of 58 dB. 1D 13C spectrum was recorded on spectral width (25,000 Hz)
with a 90° nonselective carbon pulse of 31 ms at 50 dB. During acquisition, pro-
ton decoupling (Waltz 16) was used with a power of 35 dB on a duration of 0.4 s.
For the 2D experiments, we collected data using the hypercomplex method.
A recycle delay of 5 s was used to allow complete relaxation of all proton and
carbon magnetizations.
A COSY experiment in absolute value was recorded with 512 t1 increments of
8 transients each. The total experiment time was 1.3 h. The 90° nonselective proton
pulses were 8.2 ms at 58 dB and presaturation on water signal was used with a power
of 2 dB during 1.5 s before the first 90° nonselective proton pulse. Data were
acquired with 2048 points. Zero filling was applied in both dimensions before
Fourier transformation for a final 4K*2K matrix and sine-bell apodization were

128
E. Migianu et al.
applied in both dimensions before Fourier transformation. 1H13C gHSQC
experiment[30] was used with a 13C natural abundance sample. The gradient
pulses were in the ratio 4:1 at a power level of 72 gauss Á cmÀ 1. The 90° non-
selective proton pulses were 8.2 ms at 58 dB while the 90° nonselective carbon
1 1
pulses were 31 ms at 50 dB. The /4 JC-H delay was set to a value of 1.8 ms as
1
optimized for a corresponding JC-H of 140 Hz. GARP decoupling was applied
during acquisition to gain sensitivity. Spectral widths were 4000 Hz and 8000 Hz
1
for H and 13C nucleus, respectively. Zero filling up to 2K*1K and squared sine-
bell apodization were applied in both dimensions before Fourier transformation.
A total of 256 experiments of 8 transients each were collected. The total experi-
ment time was 1.5 h. IR spectra were recorded with a Perkin-Elmer FT-IR model
2000 spectrophotometer in the 4000–500 cmÀ 1 spectral domain. Spectral re-
solution was 2 cmÀ 1 and usually 5 scans were accumulated. Multiple point
baseline correction was performed using the PE Spectrum software. Samples were
studied in H2O and D2O solutions placed in cells closed by ZnSe windows. Mass
spectra were recorded in positive reflectron mode with DHB as a matrix on a
MALDI-TOF-MS (Bruker). Microanalyses were not performed on account of
hygroscopic compounds.
Synthesis of 3’-O-Benzoyl Thymidine-H-Phosphonate (5) and
AZT-H-Phosphonate (9)
To a suspension of 3’-O-benzoyl thymidine 3[26] or AZT 8 (2 mmol) in 7.5 mL
of freshly distilled acetonitrile (for thymidine) or THF (for AZT) and 2.5 mL of dry
pyridine was added 2-chloro-1,3,2-benzodioxaphosphorin-4-one 4 (2.2 mmol) in
one portion under argon (exothermic reaction from 25°C to 35°C). The reaction
mixture was stirred 1 h at room temperature and was hydrolyzed for 2 h with 2 mL
of distilled water. After solvent removal in vacuum, the crude product was then
washed with diethyl ether.
For 3’-O-benzoyl thymidine-H-phosphonate (5). The obtained
pyridinium salt was precipitated in cold methanol and filtered. It was then
converted to the corresponding sodium salt by stirring in distilled water with ion-
exchange resin Dowex 50 W Â 2 (Na+ form). Evaporation of the filtrate and
lyophilization gave monophosphonate 5 as a colorless solid with 60% yield.
3’-O-benzoyl-3’-deoxythymidine-5’-H-phosphonate monoso-
dium salt (5). mp 216°C; IR (H2O): n = 1728 (C = O benzoyl), 1704
(C = O thymine), 1682 (C = O thymine), 1670 (C = C), 1277 (P-O-thymidine), 1205
1
(P = O), 1115 (P-O); H NMR (CD3OD): d 1.93 (s, 3H, CH3), 2.47–2.60 (m, 2H,
3
H2’), 4.31–4.43 (m, 3H, H5’ +H4’), 5.35–5.44 (m, 1H, H3’), 6.42 (t, 1H, JH-H
=
5.8 Hz, H1’), 6.92 (d, 1H, 1JP-H = 683.6 Hz, PH), 7.51 (dd, 2H, 3JH-H = 7.3 Hz and
3JH-H = 7.3 Hz, m-C6H5), 7.63 (d, 2H, 3JH-H = 7.3 Hz, o-C6H5), 7.71 (s, 1H, H6), 8.08
(d, 1H, JH-H = 7.3 Hz, p-C6H5); 31P NMR {1H} (CD3OD): d 8.1 (s); 13C NMR
3

Novel Approach to Nucleoside-5’-(1-Hydroxymethylene-1,1-Bisphosphonates)
129
{1H} (CD3OD): d 13.4 (CH3), 38.8 (C2’), 66.0 (C3’), 78.0 (C5’), 85.5 (C4’), 87.0 (C1’),
113.4 (C5), 130.6 (o-C6H5), 131.1 ( p-C6H5), 131.4 (m-C6H5), 135.7 (OCO-C6H5),
138.5 (C6), 153.3 (C2), 167.4 (C4), 168.7 (OCO).
For AZT-H-phosphonate (9). The obtained pyridinium salt was
converted to the corresponding sodium salt by stirring in distilled water with ion-
exchange resin Dowex 50 W Â 2 (Na+ form). The filtrate was concentrated and the
resulting salt was then purified by reverse-phase column chromatography using C-
18 resin (Polygoprep 60–130, Macherey-Nagel) and distilled water as eluent. The
UV-absorbing fractions (l = 254 nm) were combined and concentrated under
reduced pressure to give monophosphonate 9 as a colorless solid after lyo-
philization with 90% yield.
3’-Azido-3’-deoxythymidine-5’-H-phosphonate monosodium
salt (9). mp 155°C; IR (H2O): n = 2114, (azido), 1695 (C = O thymine),
1663 (C = O thymine), 1639 (C = C), 1274 (P-O-AZT), 1206 (P = O), 1058 (P-O);
3
3
1H NMR (D2O): d 1.94 (s, 3H, CH3), 2.53 (dd, 2H, JH-H = 6.0 Hz and JH-H
=
6.0 Hz, H2’), 4.08–4.17 (m, 2H, H5’), 4.18–4.24 (m, 1H, H4’), 4.51–4.54 (m, 1H,
3
1
H3’), 6.28 (t, 1H, JH-H = 6.0 Hz, H1’), 6.81 (d, 1H, JP-H = 638.8 Hz, PH), 7.71 (s,
1H, H6); 31P NMR {1H} (D2O): d 7.2 (s); 13C NMR {1H} (D2O): d 14.9 (CH3),
39.5 (C2’), 63.7 (C3’), 66.3 (C5’), 86.2 (C4’), 88.2 (C1’), 114.9 (C5), 140.5 (C6), 154.9
(C2), 169.7 (C4).
Synthesis of 3’-O-Benzoyl Thymidine Bis(trimethylsilyl)
Phosphite (6) Using Me3SiCl and Et3N
3’-O-benzoyl thymidine-H-phosphonate 5 (205 mg, 0.5 mmol) were dissolved
in 5 mL of dry pyridine. Distilled triethylamine (0.29 mL, 2 mmol) and distilled
trimethylsilyl chloride (0.26 mL, 2 mmol) were successively added under argon.
The reaction mixture was then stirred at room temperature for 1 h (the end of the
reaction was controlled by 31P {1H} NMR) to furnish the expected bis(silylated)
phosphite 6.
3’-O-benzoyl thymidine bis(trimethylsilyl) phosphite (6). 31P NMR {1H}: d
116.9 (s).
Synthesis of AZT Bis(trimethylsilyl) Phosphite (10) Using BSA
A suspension of AZT-H-phosphonate 9 (178 mg, 0.5 mmol) in 5 mL of dry
pyridine was stirred under argon at room temperature for 1 h to improve the
solubilization of starting material. BSA (0.62 mL, 2.5 mmol) was added drop-wise at
0°C and the reaction mixture was stirred at room temperature for 1 h (the end of the
reaction was controlled by 31P {1H} NMR) to give the expected bis(silylated)
phosphite 10.
AZT bis(trimethylsilyl) phosphite (10). 31P NMR {1H}: d 117.3 (s).

130
E. Migianu et al.
Synthesis of Bis(silylated) a-Ketophosphonates (7)
To acetyl phosphonic acid dimethyl ester (84 mg, 0.55 mmol) or benzoyl
phosphonic acid dimethyl ester (118 mg, 0.55 mmol) in 2 mL of distilled THF at
0°C under argon, trimethylsilyl bromide (0.18 mL, 1.38 mmol) was added drop-
wise. The reaction mixture was stirred at room temperature for 5 h (the end of the
reaction was controlled by 31P {1H} NMR) and evaporation of volatile fractions
(0.01 Torr) at 50°C gave bis(silylated) a-ketophosphonate 7a or 7b, respectively.
Acetyl phosphonic acid di(trimethylsilyl) ester (7a). 31P NMR {1H}: d À 17.4 (s).
Benzoyl phosphonic acid di(trimethylsilyl) ester (7b). 31P NMR {1H}: d
À 17.3 (s).
Synthesis of 3’-O-Benzoyl Thymidine-5’-Hydroxymethylene
Bisphosphonate (1)
Bis(silylated) a-ketophosphonate 7a freshly prepared was dissolved in 2 mL of
distilled chloroform and added drop-wise to thymidine bis(silylated) phosphite 6 at
0°C under argon. The reaction mixture was stirred at room temperature for 2 h.
Methanol was then added to the residue and the solution was stirred for 2 h. After
solvent removal in vacuum, the crude product was washed with 3 Â 10 mL of
diethyl ether. The obtained salt was converted to the corresponding sodium salt by
stirring in distilled water with ion-exchange resin Dowex 50 W Â 2 (Na+ form).
Evaporation of the filtrate and lyophilization gave 3’-O-benzoyl thymidine-5’-HMBP
1 as a colorless solid with 55% yield.
3’-O-benzoyl-3’-deoxythymidine-5’-[hydroxy(methyl)methy-
lene] bisphosphonate (1). mp >260°C; IR (H2O): n = 1727 (C = O
benzoyl), 1703 (C = O thymine), 1664 (C = O thymine), 1643 (C = C), 1275 (P-
1
O-thymidine), 1205 (P = O), 1102, 1074, 1062 (P-O); H NMR (D2O): d 1.61 (dd,
3H, 3JP-H = 16.0 Hz and 3JP-H = 16.0 Hz, CH3COH), 1.96 (s, 3H, CH3), 2.60–2.69
(m, 2H, H2’), 3.65–3.72 (m, 2H, H5’), 4.31–4.42 (m, 1H, H4’), 5.69–5.74 (m, 1H,
3
3
3
H3’), 6.49 (t, 1H, JH-H = 7.0 Hz, H1’), 7.58 (dd, 2H, JH-H = 7.3 Hz and JH-H
=
7.3 Hz, m-C6H5), 7.73 (d, 1H, 3JH-H = 7.3 Hz, p-C6H5), 7.87 (s, 1H, H6), 8.12 (d, 2H,
3JH-H = 7.3 Hz, o-C6H5); 31P NMR {1H} (D2O): d 21.0 [d, JP-P = 32.4 Hz, 5’-O-
2
2
2
P(O)(OH)], 21.5 (d, JP-P = 30.9 Hz, 5’-O-P(O)(OH)), 22.1 [d, JP-P = 30.9 Hz,
P(O)(OH)2], 22.5 (d, JP-P = 32.4 Hz, P(O)(OH)2); 13C NMR {1H} (D2O): d 14.9
2
1
(CH3), 23.4 (CH3COH), 39.5 (C2’), 69.0 (C3’), 74.3 (dd, JC-P = 145.0 Hz and
1JC-P = 145.0 Hz, COH), 79.4 (C5’), 87.0 (C4’), 88.2 (C1’), 115.1 (C5), 132.0 (o-C6H5),
132.1 ( p-C6H5), 132.8 (m-C6H5), 137.3 (OCO-C6H5), 140.5 (C6), 154.9 (C2), 169.7
(C4), 171.1 (OCO); MS: m/z 579.0 [M + 1]+; 601.0 [M + Na]+.
Synthesis of AZT-5’-Hydroxymethylene Bisphosphonates (2)
Bis(silylated) a-ketophosphonate 7a or 7b freshly prepared was dissolved in
2 mL of dry pyridine and added drop-wise to AZT bis(silylated) phosphite 10 at

Novel Approach to Nucleoside-5’-(1-Hydroxymethylene-1,1-Bisphosphonates)
131
0°C under argon. The reaction mixture was stirred at room temperature overnight.
Methanol was then added to the residue and the solution was stirred for 2 h. A
precipitate appeared and was filtered. Three successive precipitations were real-
ized in order to recover all the expected product. The obtained pyridinium salt
was converted to the corresponding sodium salt by stirring in distilled water with
ion-exchange resin Dowex 50 W Â 2 (Na+ form). The filtrate was concentrated and
the resulting salt was then purified by reverse-phase column chromatography using
C-18 resin (Polygoprep 60–130, Macherey-Nagel) and distilled water as eluent. The
UV-absorbing fractions (l = 254 nm) were combined and concentrated under
reduced pressure to give AZT-5’-HMBP 2a or 2b as a colorless solid after
lyophilization with 53% and 59% yields, respectively.
3’-Azido-3’-deoxythymidine-5’-[hydroxy(methyl)methylene]
bisphosphonate (2a). mp >260°C; IR (H2O): n = 2116, (azido), 1689
(C = O thymine), 1661 (C = O thymine), 1632 (C = C), 1276 (P-O-AZT), 1197
1
3
(P = O), 1106, 1072, 1061 (P-O); H NMR (D2O): d 1.57 (dd, 3H, JP-H = 15.0 Hz
and 3JP-H = 15.0 Hz, CH3COH), 1.94 (s, 3H, CH3), 2.46–2.58 (m, 2H, H2’), 4.16–
4.22 (m, 1H, H4’), 4.23–4.30 (m, 2H, H5’), 4.51–4.59 (m, 1H, H3’), 6.27 (t, 1H,
3JH-H = 7.0 Hz, H1’), 7.74 (s, 1H, H6); 31P NMR {1H} (D2O): d 18.6 [d, 2JP-P = 34.8
2
Hz, 5’-O-P(O)(OH)], 20.4 (d, JP-P = 34.8 Hz, P(O)(OH)2); 13C NMR {1H} (D2O):
d 14.8 (CH3), 23.3 (CH3COH), 39.2 (C2’), 63.9 (C3’), 68.6 (C5’), 75.3 (dd,
1
1JC-P = 141.9 Hz and JC-P = 141.9 Hz, COH), 86.6 (C4’), 88.1 (C1’), 115.1 (C5),
140.7 (C6), 155.0 (C2), 169.8 (C4); MS: m/z 500.0 [M + 1]+; 522.0 [M + Na]+.
3’-Azido-3’-deoxythymidine-5’-[hydroxy(phenyl)methylene]
bisphosphonate (2b). mp >260°C; IR (H2O): n = 2110 (azido), 1691
(C = O thymine), 1661 (C = O thymine), 1632 (C = C), 1276 (P-O-AZT), 1202
1
(P = O), 1093, 1079, 1067 (P-O); H NMR (D2O): d 1.94 (s, 3H, CH3), 1.97–2.08
(m, 1H, H2’), 2.24–2.34 (m, 1H, H2’), 3.81–3.93 (m, 1H, H4’), 3.94–4.03 (m, 2H,
3
H5’), 4.10–4.17 (m, 1H, H3’), 6.12 (t, 1H, JH-H = 6.5 Hz, H1’), 7.24–7.32 (m, 1H,
p-C6H5), 7.33–7.40 (m, 2H, m-C6H5), 7.49–7.55 (m, 1H, H6), 7.74–7.84 (m, 2H,
o-C6H5); 31P NMR {1H} (D2O): d 14.9 (d, 2JP-P = 27.7 Hz, 5’-O-P(O)(OH)), 15.0 (d,
2
2JP-P = 26.3 Hz, 5’-O-P(O)(OH)), 17.5 (d, JP-P = 26.3 Hz, P(O)(OH)2), 18.1 (d,
2JP-P = 27.7 Hz, P(O)(OH)2); 13C NMR {1H} (D2O): d 15.0 (CH3), 39.3 (C2’), 63.9
(C3’), 67.9 (C5’), 73.3 (dd, 1JC-P = 139.2 Hz and 1JC-P = 139.2 Hz, COH), 86.6 (C4’),
88.1(C1’),114.9(C5),129.3(o-C6H5),129.7(p-C6H5),130.8(m-C6H5),140.7(OCO-C6H5),
C6H5), 142.4 (C6), 154.9 (C2), 169.9 (C4); MS: m/z 562.0 [M + 1]+; 584.0 [M + Na]+.
REFERENCES
1. De Clercq, E. Toward improved anti-HIV chemotherapy: therapeutic strategies for intervention with HIV
infections. J. Med. Chem. 1995, 38, 2491–2517.
2. Schinazi, R.F. Competitive inhibitors of human immunodeficiency virus reverse transcriptase. Perspect. Drug
Discov. Des. 1993, 1, 151–180.

132
E. Migianu et al.
3. Parang, K.; Wiebe, L.I.; Knaus, E.E. Novel approaches for designing 5’-O-ester prodrugs of 3’-azido-2’,3’-
dideoxythymidine (AZT). Curr. Med. Chem. 2000, 7, 995–1039.
4. Ohrui, H.; Mitsuya, H. 4’-C-substituted-2’-deoxynucleosides: a family of antiretroviral agents which are potent
against drug-resistant HIV variants. Curr. Drug Targets Infect. Disord. 2001, 1, 1–10.
5. Moore, K.H.; Raasch, R.H.; Brouwer, K.L.; Opheim, K.; Cheeseman, S.H.; Eyster, E.; Lemon, S.M.; van der
Horst, C.M. Pharmacokinetics and bioavailability of zidovudine and its glucuronidated metabolite in patients
with human immunodeficiency virus infection and hepatic disease. Antimicrob. Agents Chemother. 1995,
39, 2732–2737.
6. Inoue, T.; Tsushita, K.; Itoh, T.; Ogura, M.; Hotta, T.; Saneyoshi, M.; Yoshida, S.; Saitoh, H.; Tomoda, Y. In
vitro bone marrow toxicity of nucleoside analogs against human immunodeficiency virus. Antimicrob.
Agents Chemother. 1989, 33, 576–579.
7. Dornsife, R.E.; Averett, D.R. In vitro potency of inhibition by antiviral drugs of hematopoietic progenitor
colony formation correlates with exposure at hemotoxic levels in human immunodeficiency virus-positive
humans. Antimicrob. Agents Chemother. 1996, 40, 514–519.
8. Chen, C.H.; Vazquez-Padua, M.; Cheng, Y.C. Effect of anti-human immunodeficiency virus nucleoside
analogs on mitochondrial DNA and its implication for delayed toxicity. Mol. Pharmacol. 1991, 39, 625–
628.
9. Faraj, A.; Fowler, D.A.; Bridges, E.G.; Sommadossi, J.P. Effects of 2’,3’-dideoxynucleosides on proliferation
and differentiation of human pluripotent progenitors in liquid culture and their effects on mitochondrial DNA
synthesis. Antimicrob. Agents Chemother. 1994, 38, 924–930.
10. Larder, B.A.; Darby, G.; Richman, D.D. HIV with reduced sensitivity to zidovudine (AZT) isolated during
prolonged therapy. Science 1989, 243, 1731–1734.
11. Balzarini, J.; Herdewijn, P.; De Clercq, E. Differential patterns of intracellular metabolism of 2’,3’-didehydro-
2’,3’-dideoxythymidine and 3’-azido-2’,3’-dideoxythymidine, two potent anti-human immunodeficiency virus
compounds. J. Biol. Chem. 1989, 264, 6127–6133.
12. Furman, P.A.; Fyfe, J.A.; St. Clair, M.H.; Weinhold, K.; Rideout, J.L.; Freeman, G.A.; Lehrman, S.N.;
Bolognesi, D.P.; Broder, S.; Mitsuya, H. Phosphorylation of 3’-azido-3’-deoxythymidine and selective
interaction of the 5’-triphosphate with human immunodeficiency virus reverse transcriptase. Proc. Natl. Acad.
Sci. U. S. A. 1986, 83, 8333–8337.
13. St. Clair, M.H.; Richards, C.A.; Spector, T.; Weinhold, K.J.; Miller, W.H.; Langlois, A.J.; Furman, P.A. 3’-
Azido-3’-deoxythymidine triphosphate as an inhibitor and substrate of purified human immunodeficiency
virus reverse transcriptase. Antimicrob. Agents Chemother. 1987, 31, 1972–1977.
14. Kedar, P.S.; Abbotts, J.; Kovacs, T.; Lesiak, K.; Torrence, P.; Wilson, S.H. Mechanism of HIV reverse
transcriptase: enzyme-primer interaction as revealed through studies of a dNTP analogue, 3’-azido-dTTP.
Biochemistry 1990, 29, 3603–3611.
15. Vrang, L.; Bazin, H.; Remaud, G.; Chattopadhyaya, J.; Oberg, B. Inhibition of the reverse transcriptase from
HIV by 3’-azido-3’-deoxythymidine triphosphate and its threo analogue. Antiviral Res. 1987, 7, 139–149.
16. Schneider, B.; Sarfati, R.; Deville-Bonne, D.; Veron, M. Role of nucleoside diphosphate kinase in the
activation of anti-HIV nucleoside analogs. J. Bioenerg. Biomembr. 2000, 32, 317–324.
17. Hitchcock, M.J.M. 2’,3’-Didehydro-2’,3’-dideoxythymidine (d4T), an anti-HIV agent. Antiviral Chem.
Chemother. 1991, 2, 125–132.
18. Cardona, V.M.; Ayi, A.I.; Aubertin, A.M.; Guedj, R. Synthesis and anti-HIV activity of some novel
arylphosphate and H-phosphonate derivatives of 3’-azido-2’,3’-dideoxythymidine and 2’,3’-didehydro-2’,3’-
dideoxythymidine. Antiviral Res. 1999, 42, 189–196.
19. Boal, J.H.; Iyer, R.P.; Egan, W. The 5’-hydrogen phosphonate analog as a potential prodrug of 3’-azido-2’,3’-
dideoxythymidine (AZT). Nucleosides Nucleotides 1993, 12, 1075–1084.
20. Rogers, M.J.; Ji, X.; Russell, R.G.G.; Blackburn, G.M.; Williamson, M.P.; Bayless, A.V.; Ebetino, F.H.; Watts,
D.J. Incorporation of bisphosphonates into adenine nucleotides by amoebae of the cellular slime mould
dictyostelium discoideum. Biochem. J. 1994, 303, 303–311.
21. Danenberg, H.D.; Golomb, G.; Groothuis, A.; Gao, J.; Epstein, H.; Swaminathan, R.V.; Seifert, P.; Edelman,
E.R. Liposomal alendronate inhibits systemic innate immunity and reduces in-stent neointimal hyperplasia in
rabbits. Circulation 2003, 108, 2798–2804.
22. Xiao, Q.; Sun, J.; Ju, Y.; Zhao, Y.F.; Cui, Y.X. Novel approach to the synthesis of AZT 5’-O-hydrogen
phospholipids. Tetrahedron Lett. 2002, 43, 5281–5283.
23. Machado, J.; Salomon, H.; Oliveira, M.; Tsoukas, C.; Krayevsky, A.A.; Wainberg, M.A. Antiviral activity and
resistance profile of phosphazid, a novel prodrug of AZT. Nucleosides Nucleotides 1999, 18, 901–906.

Novel Approach to Nucleoside-5’-(1-Hydroxymethylene-1,1-Bisphosphonates)
133
24. McKenna, C.E.; Kashemirov, B.A.; Roze, C.N. Carbonylbisphosphonate and (diazomethylene) bisphosph-
onate analogues of AZT 5’-diphosphate. Bioorg. Chem. 2002, 30, 383–395.
25. Lecouvey, M.; Leroux, Y.; Kraemer, M.; Crepin, M.; El Manouni, D.; Louriki, M. New bisphosphonates
derivatives, their preparation and use. Chem. Abstr. 2003, 138, 122736.
26. Shott, H.; Semmler, R.; Closs, K.; Eckstein, H. Preparative synthesis of guanylate-rich fragments of the
terminal sequence of macronuclear DNA from hypotrichous ciliates using the phosphotriester method in
solution. Makromol. Chem. 1987, 188, 1313–1346.
27. Dreef, C.E.; Valentijn, A.R.P.M.; de Vroom, E.; van der Marel, G.A.; van Boom, J.H. An approach towards
the formation of an ester bond between the primary hydroxyl of a b-D-galactopyranoside and 2-
aminoethylphosphonic acid and N-methyl substituted derivatives. Tetrahedron Lett. 1988, 29, 1199–1202.
28. Hata, T.; Sekine, M. Oxidation of nucleoside phosphites by means of 2,2’-dipyridyl disulfide via nucleoside
silylphosphite intermediates. Tetrahedron Lett. 1974, 15, 3943–3946.
29. Wada, T.; Mochizuki, A.; Sato, Y.; Sekine, M. A convenient method for phosphorylation involving a facile
oxidation of H-phosphonate monoesters via bis(trimethylsilyl) phosphites. Tetrahedron Lett. 1998, 39,
7123–7126.
30. Wilker, W.; Leibfritz, D.; Kerssebaum, R.; Bermel, W. Gradient selection in inverse heteronuclear correlation
spectroscopy. Magn. Reson. Chem. 1993, 31, 287–292.