Synthesis of AZT 5′-triphosphate mimics and their inhibitory effects on HIV-1 reverse transcriptase

Article abstract of DOI:10.1021/jm040116w

In search of active nucleoside 5′-triphosphate mimics, we have synthesized a series of AZT triphosphate mimics (AZT P3Ms) and evaluated their inhibitory effects on HIV-1 reverse transcriptase as well as their stability in fetal calf serum and in CEM cell

Full text of DOI:10.1021/jm040116w

6902
J. Med. Chem. 2004, 47, 6902-6913
Synthesis of AZT 5′-Triphosphate Mimics and Their Inhibitory Effects on HIV-1
Reverse Transcriptase
Guangyi Wang,* Nicholas Boyle, Fu Chen, Vasanthakumar Rajappan, Patrick Fagan, Jennifer L. Brooks,
Tiffany Hurd, Janet M. Leeds, Vivek K. Rajwanshi, Yi Jin, Marija Prhavc, Thomas W. Bruice, and P. Dan Cook
Research Laboratories, Biota, Inc., 2232 Rutherford Road, Carlsbad, California 92008
Received June 18, 2004
In search of active nucleoside 5′-triphosphate mimics, we have synthesized a series of AZT
triphosphate mimics (AZT P3Ms) and evaluated their inhibitory effects on HIV-1 reverse
transcriptase as well as their stability in fetal calf serum and in CEM cell extracts. Reaction
of AZT with 2-chloro-4H-1,3,2-benzodioxaphosphorin-4-one, followed by treatment of the
phosphite intermediate 2 with pyrophosphate analogues, yielded the cyclic triphosphate
intermediates 4b-4f, which were subjected to boronation and subsequent hydrolysis to give
AZT 5′-R-borano-â,γ-bridge-modified triphosphates 6b-6f in moderate to good yields. Reaction
of the cyclic intermediate 4d with iodine, followed by treatment with a series of nucleophiles,
afforded the AZT 5′-â,γ-difluoromethylene-γ-substituted triphosphates (7b-7i). Several different
types of AZT P3Ms containing R-P-thio (or dithio) and â,γ-difluoromethylene (13,14), R,â-
difluoromethylene and γ-P-methyl(or phenyl) (15,16), and R-borano-â,γ-difluoromethylene and
γ-O-methyl/phenyl (11,12) were also synthesized. The effectiveness of the compounds as
inhibitors of HIV-1 reverse transcriptase was determined using a fluorometric assay and a
poly(A) homopolymer as a template. A number of AZT P3Ms exhibited very potent inhibition
of HIV-1 reverse transcriptase. Modifications at the â,γ-bridge of triphosphate rendered the
AZT P3Ms 6b-6f with varied activities (Ki from 9.5 to .500 nM) while modification at the
R,â-bridge of triphosphate led to weak AZT P3M inhibitors. The results imply that the AZT
P3Ms were substrate inhibitors, as is AZT triphosphate. The most active compound, AZT 5′-
R-Rp-borano-â,γ-(difluoromethylene)triphosphate (AZT 5′-RB-âγCF2TP) (6d-I), is as potent as
AZT triphosphate with a Ki value of 9.5 nM and at least 20-fold more stable than AZT
triphosphate in the serum and cell extracts. Therefore, for the first time, a highly active and
stable nucleoside triphosphate mimic has been identified, which is potentially useful as a new
type of antiviral drug. The promising triphosphate mimic, 5′-R-borano-â,γ-(difluoromethylene)-
triphosphate, is expected to be valuable to the discovery of nucleotide mimic antiviral drugs.
Introduction
necessarily a prerequisite for antiviral activities. There-
fore, any alternative route that can lead to sufficient
cellular concentration of a desired NTP may be highly
useful. A conceivable alternative is to directly use a
nucleotide as a drug entity, which can bypass cellular
phosphorylation. To bypass the first cellular phospho-
rylation, NMP prodrugs have been explored as an
alternative form of nucleoside antiviral drugs. Several
types of NMP prodrugs have been intensively studied
and have shown promising in vitro activities.6-10 How-
ever, a major obstacle to success is the inherent insta-
bility of NMPs’ natural 5′-monophosphate to cellular
dephosphorylating enzymes such as 5′-nucleotidases
and phosphodiesterases. In the case of a starting
nucleoside that cannot be phosphorylated by cellular
enzymes, the NMP prodrug, most likely, will be de-
graded to the inactive, parent nucleoside. In most cases,
a mixture of nucleotides at all possible phosphorylation
stages is likely formed after the NMP is released from
an NMP prodrug. As a result, administration of NMP
prodrugs may not offer many advantages over current
nucleoside drugs themselves. It seems that the NMP
prodrug approach is more suitable to stable NMPs or
NMP mimics. In fact, the NMP prodrug approach has
been successfully applied to nonhydrolyzable acyclic
Dideoxynucleoside (ddN) antiviral drugs are succes-
sively phosphorylated in cells to ddN 5′-monophosphates
(ddNMPs), ddN 5′-diphosphates (ddNDPs), and ddN 5′-
triphosphates (ddNTPs).1 The triphosphates of the
antiviral nucleosides are the active chemical species that
inhibit viral DNA synthesis. Although ddNTP binding
in the catalytic site of a viral polymerase exerts a
competitive inhibitory effect, overall viral DNA synthe-
sis is inhibited predominantly through incorporation of
the ddNMP into the elongating DNA chain and subse-
quent chain termination.1-4 Therefore, a prospective
antiviral nucleoside should meet two basic require-
ments: (1) ready phosphorylation to NTP by cellular
enzymes and (2) ready incorporation into a viral DNA
sequence and subsequent termination of the viral DNA
chain. Many nucleosides that could not be phosphory-
lated by cellular enzymes failed to show antiviral
activities and were eliminated in the initial drug
screening, although the triphosphates of these nucleo-
sides might be potent inhibitors of viral polymerases.5
Cellular phosphorylation of an antiviral nucleoside by
kinases is a process leading to an active NTP, but not
* Corresponding author. Phone. 760 496 0425. Fax: 760 804 2740.
E-mail: g.wang@biota-inc.com.
10.1021/jm040116w CCC: $27.50 © 2004 American Chemical Society
Published on Web 12/03/2004

Synthesis of AZT 5′-Triphosphate Mimics
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6903
nucleoside phosphonates, a type of NMP mimic. Cur-
rently, tenofovir disoproxil fumarate (a bis-POC prodrug
of PMPA)11 and adefovir dipivoxil (a bis-POM prodrug
of PMEA)12 are used clinically as anti-HIV and anti-
HBV drugs, respectively. After it is liberated from an
NMP mimic prodrug, the NMP mimic still needs further
cellular phosphorylation to become an active species.
Since NTPs are labile to cellular dephosphorylating
enzymes, it is hardly imaginable to use NTPs as
potential drugs. However, a NTP mimic (NP3M) that
can entirely bypass cellular phosphorylation and pos-
sesses satisfactory cellular stability may be useful as
an antiviral agent. If such an NP3M is identified, it is
expected to offer certain advantages over nucleoside
drugs concerning toxicity and drug resistance. Use of
stable NP3Ms can avoid cellular buildup of undesired
NMPs and NDPs, thus preventing unintended cellular
interactions involving NMPs and NDPs. Drug resistance
toward some antiviral nucleosides results from viral
DNA repair arising from the deletion of incorporated
ddNMP.13 When an NP3M is employed, the viral repair
activity is expected to be greatly reduced, or even
diminished. Furthermore, an active NP3M may be
obtained simply by attaching a favorable triphosphate
mimic (P3M) to a variety of nucleoside analogues,
regardless of their properties toward cellular phospho-
rylations.
Now the task is to identify active and stable P3Ms.
There are previously known NP3Ms, in which a bridging
oxygen of triphosphate is replaced by methylene,14
halomethylene,15,16 or imido17,18 or in which a nonbridg-
ing oxygen is replaced by other functional groups.19-23
A few NP3Ms containing a single â,γ-bridge modifica-
tion have shown potent inhibition of HIV reverse
transcriptase.18,24 As expected, a single bridge modifica-
tion on a triphosphate moiety can enhance the stability
of a P3M to dephosphorylating enzymes, but it does not
render the P3M sufficient stability for human thera-
peutics.25 A couple of NP3Ms comprising two bridge
modifications such as dimethylene and diimido were
also reported, but unfortunately, no significant biological
activities were observed.18,26 Research in the field of
NP3Ms has spanned more than 2 decades, but explora-
tion of NP3Ms as human therapeutics remains in its
infancy owing to challenges in chemical synthesis,
purification, and drug delivery. To address the chal-
lenges, our first task was to identify useful P3Ms that
can lead to very potent NP3M antiviral agents with
satisfactory stability. In this paper, we report the
synthesis of a new generation of AZT P3Ms as well as
their inhibitory properties against HIV-1 reverse tran-
scriptase and their stability in serum and cell extracts.
Figure 1. AZT P3Ms.
ported to be very potent inhibitors of both wild type and
mutant (R72A, K65A, K65R) HIV-1 reverse transcriptas-
es.24 For the wild type, both compounds were more
potent than AZT triphosphate (AZTTP), with Ki being
2.6- and 2.4-fold lower. For all three mutant enzymes,
both compounds exhibited comparable potency to that
of the wild-type enzyme, while AZTTP and d4TTP were
less effective inhibitors of the mutant enzymes. Activity
enhancement against mutant enzymes by the R-RP-
borano analogues was attributed to a decreased excision
rate of R-RP-borano-incorporated DNA chain by the
mutant reverse transcriptases.24 The work also dem-
onstrated that the nucleoside 5′-R-RP-boranotriphos-
phates had comparable binding affinity to HIV-1 RT as
the NTPs, as measured by their KD values. It appears
that this P3M possesses desirable properties and can
serve as a starting point in our search for superior
P3Ms, although its stability to dephosphorylating en-
zymes is still of concern. Since our initial goal was to
identify those P3Ms that are chemically and biologically
stable enough to achieve their anticipated biological
activities, stabilized analogues of active NP3Ms such as
AZT 5′-R-RP-boranotriphosphate naturally became our
first priority. An easily conceivable approach to enhance
the stability of 5′-R-RP-boranotriphosphate without
deteriorating its ability to be incorporated into DNA
may be to modify the â,γ-bridge of the P3M by replace-
ment of oxygen with methylene, halomethylene, or
imido. In addition, the γ-P position of the P3M may be
attached to a relatively small substituent in order to
further enhance enzyme stability.20 Besides 5′-R-RP-
borano, 5′-R-P-thiotriphosphate was also expected to
provide adequate binding and incorporative properties.19
For comparison purposes, it was appropriate to synthe-
size a few R,â-bridge-modified NP3Ms. Our first group
of AZT P3Ms selected from variations as shown in
Figure 1 has been synthesized and evaluated.
Chemistry
In this work, a very efficient synthetic procedure via
an activated phosphite and a cyclized triphosphate
intermediate21 was adopted for preparation of AZT
P3Ms 6b-6f as shown in Scheme 1. The difference from
the literature21 was that methylenediphosphonate, ha-
lomethylenediphosphonates, and imidodiphosphate, in-
stead of pyrophosphate, were employed as a coupling
reagent. It was quickly concluded that the pyrophos-
phate analogues behaved similarly as pyrophosphate in
this one-pot, multi-intermediate reaction after a few
desired AZT P3Ms were successfully prepared. In the
same manner as commercial pyrophosphate tributyl-
ammonium reagent, methylenediphosphonic acid (form-
ing tributylammonium salt 3b in the reaction mixture)
Considerations in Design of NP3Ms
Previously reported data have shown that AZT 5′-â,γ-
imidotriphosphate was a much more potent inhibitor
of HIV-1 reverse transcriptase than AZT 5′-R,â-imido-
triphosphate, with Ki values of 0.087 vs 22 µM.18 As the
data have revealed, the mechanism of action mostly
likely involves incorporation of the NMP mimic and
subsequent DNA chain termination. Therefore, our
initial work on NP3M antiviral agents was focused on
incorporable NP3Ms having sufficient cellular stability.
AZT and D4T 5′-R-RP-boranotriphosphates were re-

6904 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Wang et al.
Scheme 1. Synthesis of AZT 5′-R-P-Borano-â,γ-bridge-modified Triphosphates
was used directly without further treatment. Com-
mercially available dichloromethylenediphosphonic acid
disodium salt was converted to its bis(tributylammo-
nium) salt (3e) before the coupling reaction. Owing to
its sensitivity to acid, imidodiphosphate sodium salt was
converted to its tetrakis(tributylammonium) salt (3f).
Fluoromethylenediphosphonic acid and difluorometh-
ylenediphosphonic acid were prepared, respectively, by
treatment with trimethylbromosilane of tetraisopropyl
fluoromethylenediphosphonate and tetraisopropyl dif-
luoromethylenediphosphonate, which were obtained by
reaction of tetraisopropyl methylenediphosphonate with
N-fluorobenzenesulfonimide in the presence of sodium
bis(trimethylsilyl)amide. Initial preparations of tetrai-
sopropyl difluoromethylenediphosphonate were con-
ducted at low temperature and achieved only moderate
yields. An improved procedure, including multiple,
alternate addition of sodium bis(trimethylsilyl)amide
and N-fluorobenzenesulfonimide at ambient tempera-
ture, gave tetraisopropyl difluoromethylenediphospho-
nate in more than 80% yield. Both fluoromethylene-
diphosphonic acid and difluoromethylenediphosphonic
acid were converted to their bis(tributylammonium)
salts (3c and 3d), respectively.
Synthesis of the AZT P3Ms 6a-6f is depicted in
Scheme 1. Reaction of AZT (1) with 2-chloro-4H-1,3,2-
benzodioxaphosphorin-4-one yielded the activated phos-
phite 2, which was condensed with pyrophosphate
analogues 3 to form the cyclic triphosphate intermedi-
ates 4. It was critical to use the tributylammonium salt
of pyrophosphate analogues free of water, since the
intermediates 4 were very labile to moisture, leading
to formation of undesired AZT 5′-H-phosphonate (not
shown). Treatment of 4 with borane-diisopropylethyl-
amine complex yielded, after hydrolysis, the AZT 5′-R-
P-borano-â,γ-bridged-modified triphosphates 6b-6f.
Compound 6a was prepared according to a known
procedure.21 We observed that in most cases a slight
excess of 2-chloro-4H-1,3,2-benzodioxaphosphorin-4-one
(1.2 equiv) and pyrophosphate analogue tributylammo-
nium salts (1.3 equiv) would ensure the complete
formation of the intermediates 2 and possibly 4. The
AZT P3Ms 6a-6e were formed, respectively, in good
yields (LCMS ∼70-90% before purification). However,
6f containing the â,γ-imido bridge was obtained in a
lower yield, presumably owing to its lability to hydroly-
sis during purification.
Compounds 6a, 6b, 6d, 6e, and 6f all consist of two
diastereomers in approximately 1:1 ratio, which were
purified first by anion-exchange HPLC and then sepa-
rated on a reverse-phase HPLC, except for 6b and 6f,
which were not separable on the reverse-phase HPLC.
The separated diastereomers were designated as dias-
tereomer I and diastereomer II, according to whether
they were eluted earlier (I) or later (II) from the C18
reverse-phase HPLC. Determination of the stereochem-
istry of the diastereomers was not attempted. The
diastereomers I tend to have much stronger inhibition
of HIV-1 reverse transcriptase (Table 1) than the
diastereomers II. Since the diastereomer I of 6a was
previously assigned as the R-RP-borano diastereomer24
and was a much more potent inhibitor than the other
diastereomer, we tentatively designate the diastereo-
mers I of 6a, 6c, 6d and 6e as the R-RP-borano
diastereomers. Unless specified, the diastereomer I
means the Rp-isomer in this context. Compound 6c
consisted of four stereoisomers, and an HPLC separa-
tion yielded four fractions: the pure diastereomer I, the
diastereomer II contaminated by other diastereomers,
the diastereomer III contaminated by other diastereo-
mers, and the pure diastereomer IV. The diastereomer
I was first eluted and the diastereomer IV was last
eluted.
Scheme 2 shows synthesis of AZT 5′-γ-P-substituted
â,γ-(difluoromethylene)triphosphates 7a-i. As in oli-
gonucleotide synthesis,27 the phosphite intermediate 4d
could be oxidized with iodine, and a subsequent treat-
ment with a nucleophile yielded 7a-7i. In the case of
water as a nucleophile, the nucleophilic attack on any
of the three phosphorus atoms of 4d would result in the
same product 7a. When the nucleophile was not a
hydroxyl anion or water, its attack on the different
phosphorus atoms would result in the formation of two
different products, R- or γ-substituted. Initially, we
expected 7b-7i were either the R-P-substituted AZT
P3Ms 10 (shown in Scheme 3) or a mixture of 10 and 7.
Surprisingly, only one diastereoisomer was obtained for
1
each reaction, as indicated by H, 19F, and 31P NMR.

Synthesis of AZT 5′-Triphosphate Mimics
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6905
been simple for the γ-P, most likely a sextet, while the
R-P would be a double-double splitting. In fact, the R-P
of 7g displays a doublet coupling pattern, identical to
that of 7a, while the γ-P has a complicated, multiple
splitting pattern (double-double-triple), instead of a
sextet, indicating that fluorine is attached to the γ-P.
From the above discussion, 7b-7i can be clearly as-
signed as the γ-P-substituted AZT P3Ms.
To our knowledge, the formation of the γ-P-substitu-
tion described above has not been reported in the
literature. After the structural determination of 7b-
7i, a possible reaction path was postulated as shown in
Scheme 3. A nucleophile attacks on either the â or â′ of
8, an oxidized intermediate of 4d, and the resulting
intermediate 9a was subsequently hydrolyzed to give
7. If a nucleophile attacked the R-P of 8, an intermediate
9b would be formed, which might be hydrolyzed to 10.
However, the absence of a detectable amount of 10 in
the products suggests that a barrier to the formation of
either 9b or 10 exist. The observed regional selectivity
can be ascribed to a favorable formation of 9a over 9b.
The stronger electronegativity of difluoromethylene over
oxygen may also contribute to the attack on the â-P or
â′-P of 8. The reaction mechanism remains to be further
investigated.
Several other AZT P3Ms having diverse modifications
on the triphosphate moiety were also synthesized, as
shown in Figure 2. Compound 11 was prepared by
reaction of 6d-I/II with the corresponding methyl trif-
luoromethanesulfonate. For this reaction, only the
monomethylated 11 was isolated, although dimethyl
and trimethyl derivatives were also observed on LCMS.
The results revealed that possible substitution at the
R-P or â-P position of 6d might have resulted in
degradation of the triphosphate mimic moiety upon
quenching reactions with water. This postulation was
supported by fragments observed on LCMS. Thus,
significant amounts of fragments 225, 239, and 345,
corresponding to methyl (difluoromethylene)diphospho-
nate, dimethyl (difluoromethylene)diphosphonate, and
AZT 5′-P-(OMe)-P-boranophosphinic acid (or N3-methyl-
AZT 5′-P-boranophosphinic acid), were observed. Com-
pound 12 was synthesized in good yield from 6d-I/II
through a metaphosphate intermediate (most likely 5d),
which was reacted with phenol in the presence of
triethylamine. Reaction of 4d (X ) CF2) with elemental
sulfur and subsequent treatment with water28 yielded
AZT 5′-R-P-thiotriphosphate (13). Compound 14 was
obtained by treatment of 4d first with sulfur and then
with lithium sulfide.28 Compounds 15 and 16 were
prepared, respectively, by reactions of the tris(tributy-
lammonium) salt of AZT 5′-(difluoromethylene)diphos-
phonate with methylphosphonic dichloride and phe-
nylphosphonic dichloride by a similar procedure as that
described for the preparation of nucleoside γ-substituted
triphosphates.29 AZT 5′-(difluoromethylene)diphospho-
nate was obtained by reaction of AZT 5′-O-tosylate30
with the tetrabutylammonium salt of (difluoromethyl-
ene)diphosphonic acid. An excess of the diphosphonic
acid salt (3-5 equiv) was necessary to suppress forma-
tion of a di-AZT diphosphonate (not shown).
Figure 2. AZT P3Ms having R- and/or γ-P-modifications.
Scheme 2. Synthesis of AZT 5′-γ-P-Substituted
â,γ-(Difluoromethylene)triphosphates
This largely excluded the possibility of the R-P-substitu-
tion that would yield two R-P-substituted diastereomers.
A careful examination of 31P NMR spectra revealed that
compared to unsubstituted 7a, compounds 7b-7i main-
tained similar chemical shifts and the identical splitting
pattern at the R-P, whereas their chemical shifts and/
or splitting patterns of the γ-P are significantly different
from that of 7a. The results clearly support a γ-P
substitution. Further evidence was obtained from 31P
NMR data of 7g, a P-fluoro AZT P3M. If fluorine was
attached to the R-P, the coupling pattern would have
Biology
Inhibition of HIV-1 Reverse Transcriptase. The
effectiveness of the compounds as inhibitors of HIV-1

6906 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Wang et al.
Scheme 3. Nucleophilic Reaction Pathways of the Oxidized Cyclic Triphosphate Intermediate 8
Table 1. Inhibition of HIV-1 Reverse Transcriptase by AZT
fication on the triphosphate moiety. Activities of the
diastereomers II (containing an R-Sp-borano) including
6d-II and 6c-IV were varied but mostly decreased
compared to those of the diastereomers I (containing an
R-Rp-borano). An exception is 6e-II, which had a similar
activity as 6e-I. Even with two modifications on the
triphosphate moiety, a few AZT P3Ms still maintained
the same level of inhibition as the triphosphates of
clinically used nucleoside reverse transcriptase inhibi-
tors (NRTIs). However, compound 6b-I/II, a mixture of
diastereomers I and II, was among the least active AZT
P3Ms of this work, with 17% inhibition at 0.5 µM,
indicating a striking effect of â,γ-bridge modifications.
Α·Τ 5′-â,γ-(difluoromethylene)triphosphate (7a), the
5′-γ-P-substituted AZT P3Ms 7b-7i, 11, and 12 exhib-
ited varied inhibitory effects. Except for 7f, which was
a weak inhibitor, 7a and the other γ-substituted AZT
P3Ms 7b-e, 7g-I containing γ-P-F, γ-P-NH2, γ-P-OMe,
γ-P-OPh, γ-P-NHMe, γ-P-NHEt, and γ-P-NHPh all
showed significant activities with Ki values in a range
of 0.04-0.85 µM. Positive effects of the γ-P-substituents
on the inhibition were not observed, but certain toler-
ance of the reverse transcriptase to the γ-P-substituents
was clearly demonstrated. γ-P-Substituted compounds
11 and 12 containing R-Rp/Sp-borano exhibited signifi-
cantly stronger inhibitory effects than 7h and 7i, which
clearly indicates a positive effect of the R-P-borano
group. Compounds 13, containing R-Rp/Sp-thio showed
slightly higher Ki than 7a (0.090 vs 0.041 µM), was still
a potent inhibitor, whereas 14, containing the R-P-
dithio, was a much weaker inhibitor. Compounds 15 and
16, which contain an R,â-difluoromethylene bridge in
the triphosphate moiety, were the least active inhibitors
of this work.
P3Ms
rel
Ki (µM) inhibnc
compda
R1
R2
X
X1
X2
AZT TP
6a-I
6d-I
6d-II
6e-I
6e-II
6c-I
6c-IV
6f-I/II
6b-I/II
7a
OH OH
BH3 OH
BH3 OH
BH3 OH
BH3 OH
BH3 OH
BH3 OH
BH3 OH
BH3 OH
BH3 OH
OH OH
OH NH2
OH NHMe
OH NHEt
OH NHPh
OH N3
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
S
O
O
0.0084
0.0096
0.0095
4.56
0.093
0.074
0.027
1
1.1
1.1
543
11
8.8
3.2
O
O
O
CF2
CF2
CCl2
CCl2
CHF
O
O
O
O
O
CHF 49%b
O
NH
CH2
CF2
CF2
CF2
CF2
CF2
CF2
CF2
CF2
CF2
CF2
CF2
CF2
CF2
O
0.105
12.5
4.94
4.88
5.56
14.8
49.4
30.2
O
17%b
0.041
0.234
0.124
0.415
0.254
24%b
0.852
0.597
0.303
0.069
0.113
0.090
19%b
14.2%b
11.5%b
O
7b
O
7c
O
7d
O
7e
O
7f
O
7g
OH
F
O
18.8
71.1
36.1
8.2
13.5
10.7
7h
OH OMe
OH OPh
BH3 OMe
BH3 OPh
SH OH
SH OH
OH Me
O
7i
O
11-I/II
12-I/II
13-I/II
14
O
O
O
O
15
16
O
O
CF2
CF2
OH Ph
O
a Diastereomer I assigned to the AZT P3M having shorter HPLC
retention time; Diastereomer II assigned to the AZT P3M having
longer HPLC retention time; I/II refers to a mixture b Percentage
inhibition at 0.5 µM of AZT P3M. c The inhibitory effect of AZTTP
is designated as 1. Relative inhibition indicates the inhibitory effect
of a compound relative to AZTTP. More potent, smaller the
numbers are.
Serum and Cell Extract Stability. Serum stability
of selected AZT P3Ms was assessed in fetal calf serum
at 37 °C following a published procedure.20 The most
active AZT P3M 6d-I of this work also demonstrated
satisfactory serum stability (Table 2), having a half-life
of more than 48 h relative to a half-life of 2 h that
AZTTP and TTP had under the assay conditions used.
Several other AZT P3Ms containing R-P-borano and â,γ-
bridge modifications all showed satisfactory stability in
serum, with half-lives in the range 36 h to more than
48 h. As expected, AZT 5′-R-P-boranotriphosphate (6a-I
and 6a-II) had only moderately higher half-life (6 h)
reverse transcriptase (RT) was determined using a
fluorometric assay31 and a poly(A) homopolymer as a
template. Several compounds exhibited very potent
inhibition, as shown in Table 1. Compound 6d-I was
equipotent to two of the most potent inhibitors of HIV-1
RT, AZTTP, and 6a-I, with Ki values of 0.0095, 0.0084,
and 0.0096 µM, respectively. Compounds 6c-I, 6e-I, and
6f-I/II were also very active, with Ki values of 0.027,
0.093, and 0.105 µM, respectively. These AZT P3Ms
have or contain an R-Rp-borano and a â,γ-bridge modi-

Synthesis of AZT 5′-Triphosphate Mimics
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6907
Table 2. Serum and Cell Extract Stability of AZT P3Ms
RT and Q151M HIV RT gives useful insight into the
binding and incorporation of ddN 5′-R-P-boranotriph-
osphates.33,34 Introduction of the second sulfur at the
R-P position led to a weakly active compound 14, which
may be explained by a weakened magnesium-triphos-
phate complex. A few nucleoside R-P-alkyl triphosphates
have been synthesized previously in other laboratories
and found to be only moderately active.35 It seems that
for potent inhibition of HIV-1 RT, a preferred substitute
for the R-P-hydroxyl should be ionizable to an anion
under physiological conditions as is R-P-borane or R-P-
thio.
half-life (h)
compd
serum
cell extract
6d-I
6d-II
6e-I
6e-II
6f-I/II
6a-I
>48
>48
45
>48
36
6
>48
>48
>48
>48
>48
NT
6a-II
AZTTP
TTP
6
NT
2
NT
NT
0.25
2
ATP
2
As shown in Table 1, compounds 6a-6f exhibited
varied activities from very potent to almost inactive. The
only structural difference among them is the â,γ-bridge
modification. Several factors including size, polarity,
and electronegativity may affect the activity. Since all
these bridge modifications are relatively small, size does
not appear to be a predominant factor. In fact, the
bulkier bridge CCl2 led to a much more potent inhibition
than the smaller CH2. It seems that the inhibitory
effects correlate well with the electronegativity of the
bridge modifications. It appears that the more electro-
negative atom at the â,γ-bridge position leads to the
lower Ki with an inhibitory order O ∼ CF2 > CHF >
CCl2 ∼ NH . CH2. This order is roughly the order of
the last ionizable proton’s pKa values of adenosine 5′-
â,γ-bridge-modified triphosphates: CF2 (6.7) < CCl2
(7.0) ≈ O (7.1) < CHF (7.4) < NH (7.7) < CH2 (8.4).14 It
seems that size of the bridge modifications becomes
more important when the bridge modifications have
similar electronegativity. Dichloromethylene and oxygen
have similar electronegativities, as reflected by the pKa
values shown above, but the larger size of dichlorom-
ethylene may have made 6e less active than 6a-I. Less
electronegative fluoromethylene led to a more active
inhibitor 6c-I than 6e containing the more electroneg-
ative dichloromethylene. A likely postulation is that the
electronegativity and size of the bridge modifications
may affect the ability of a P3M to form a complex with
a magnesium ion, thereby influencing the binding
affinity of the P3M in the catalytic site and subsequently
the substrate property of the corresponding AZT P3Ms.
In addition, more electronegative bridge modifications
may also facilitate the departure of the bridge-modified
diphosphate.
than the NTPs. Compounds 6d, 6e, and 6f were also
subjected to a cell extract stability assay using CEM
cells. The results show that all the AZT P3Ms had half-
lives of more than 48 h relative to a half-life of 0.25 h
for ATP.
Discussion
Substitution of borano for hydroxyl at the R-P position
of triphosphate is crucial in maintaining and enhancing
HIV-1 RT inhibition, leading to the AZT P3M 6d-I,
which is more potent than its R-P-hydroxyl analogue
7a and equipotent to AZTTP. A borane attached to the
R-phosphorus of a triphosphate maintains a negative
charge but is not able to complex with metal ions, since
it lacks a lone pair of electrons. A recent publication24
suggested that only one coordinating oxygen at the R-P
position be required for formation of a magnesium-
triphosphate complex in the catalytic site of HIV-1 RT,
and the other nonbridging oxygen may be replaced by
a suitable substituent. Consistent with this, the sub-
stitution of a thiol for hydroxyl at the R-P position
yielded the active AZT P3M 13. Sulfur has properties
more similar to oxygen than does borane, and conse-
quently, 13 has a similar activity as 7a. It is very
interesting to note that 6d-I is several fold more active
than 7a and 13, suggesting an unusual role for the R-Rp-
borano in the inhibition of the viral DNA synthesis
mediated by HIV-1 RT. Under physiological conditions,
P-OH, P-SH, and P-BH3 all are negatively charged,
and the negative charge is likely polarized toward sulfur
in the case of R-P(dO)S-, whereas the charge is polar-
- 32
ized to oxygen in the case of R-P(dO)BH3
.
It seems
that the electron distribution is one of factors governing
the activity, since the least electronegative borano
effects the strongest inhibition, while more electroneg-
ative oxygen and sulfur led to less active 7a and 13. A
previous work24 has shown that AZT 5′-R-Rp-boranot-
riphosphate had a similar binding affinity as AZTTP
in the HIV-1 RT catalytic site, but its incorporation
efficiency increased 3-9-fold. According to the results
from the previous work24 and this work, we surmise that
a complex involving the R-nonbridging oxygen of
Conclusion
A series of novel AZT P3Ms have been synthesized,
and practical synthetic and purification procedures have
been developed. A number of AZT P3Ms exhibited very
potent inhibition of HIV-1 RT and satisfactory stability
to serum and cell extracts. The SAR presented here
provides the guidance for future NP3M design. As can
be seen, the most promising P3M of this work is 5′-R-
Rp-borano-â,γ-(difluoromethylene)triphosphate (RB-â,γ-
CF2TP). The importance of this P3M is not only the high
potency it leads to when attached to AZT, but also its
favorable biological stability. This P3M is also antici-
pated to be useful when attached to other prospective
nucleosides toward antiviral drugs. Compared to all the
P3Ms known to date, RB-â,γCF2TP is the only one
having two modifications that can render NP3Ms with
both high potency of inhibition and dramatically en-
hanced biological stability relative to NTPs and NP3Ms
-
P(dO)BH3 and a magnesium ion is stronger than the
complex involving the R-nonbridging oxygen of P(dO)O-
and a magnesium ion, which facilitates the breakage
of the PR-O-Pâ bond in the R-P-BH3 analogue, result-
ing in more efficient incorporation for 6d-I compared
to 7a. The effects of the R-P-borano on the binding in
the catalytic site and on the PR-Ã-Pâ bond cleavage
remain to be further explained. Recent work on sup-
pression of dideoxynucleotide resistance by K65R HIV

6908 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Wang et al.
aqueous sodium bicarbonate and then with brine, dried over
sodium sulfate, and concentrated. Chromatography on silica
gel with 5-30% ethyl acetate in hexanes gave 2.18 g of
tetraisopropyl (difluoromethylene)diphosphonate and 0.32 g
of tetraisopropyl (fluoromethylene)diphosphonate. CF2-prod-
uct: 1H NMR (CDCl3) δ 1.39 (2d, J ) 3.3 Hz, 4Me, 12H), 1.41
(2d, J ) 3.3 Hz, 4Me, 12H), 4.92 (m, Me2CH, 4H); 31P NMR δ
3.41 (t, J ) 86.9 Hz); 19F NMR δ -122.10 (t, J ) 87.2 Hz).
CF-product: 1H NMR (CDCl3) δ 1.37 (m, 8Me, 24H), 4.85 (m,
Me2CH, 4H); 31P NMR δ 10.72 (t, J ) 63.6 Hz). 19F NMR
(CDCl3) δ -225.9 (dt, 2JF-H ) 43.6 Hz, 2JF-P ) 63.0 Hz, CHF).
Improved Procedure for Preparation of Tetraisopro-
pyl (Difluoromethylene)diphosphonate. NaHMDS (1.0 M,
235 mL) and anhydrous THF (250 mL solution) in an addition
funnel and NFSi (82 g, 260 mmol) in THF (250 mL solution)
in another addition funnel were added alternately and in 10
portions each, starting with NaHMDS, to stirred tetraisopropyl
methylenediphosphonate (27 g, 76 mmol) in a three necked, 1
L flask under argon. After each addition a 1-min interval was
allowed. A heavy precipitate occurred after the second and
latter round of additions. A pale-amber color developed and
substantial heat was given off, so the reaction mixture was
near reflux during the additions. After the additions, the
mixture was stirred until it reached room temperature, cooled
to -78 °C, quenched with saturated aqueous ammonium
chloride (100 mL), and diluted with 200 mL of diethyl ether.
The mixture was warmed to room temperature and the organic
layer was separated. The aqueous layer was extracted with
ether (2 × 100 mL). Combined organics were washed with 5%
citric acid/brine, bicarbonate, and brine; dried over magnesium
sulfate; and concentrated. The resulting syrup was loaded on
a silica gel column and eluted with 10% ether in hexanes (1
L), 20% ether in hexanes (1 L), and finally 50% ether/hexanes.
Fractions containing pure product (TLC: eluted with ether/
hexanes 3:1, stained with KMnO4) were collected and concen-
trated to give 24.0 g (81%) of (difluoromethylene)diphospho-
nate as a colorless liquid. 1H, 31P, and 19F NMR data are
identical to those shown above.
containing one modification. The most potent compound,
6d-I, of this work is equipotent to AZTTP in the HIV-1
RT assay and at least 20-fold more stable than AZTTP
in serum and cell extract assays. Therefore, for the first
time, a highly active and stable NP3M has been identi-
fied. Further biological evaluation of 6d-I clearly is
justified and is underway. Considering the physical
properties related to polarity and negative charges, an
effective drug delivery approach may be required for the
AZT P3Ms to be useful anti-HIV agents. Significant
progress toward AZT P3M prodrugs has been made in
these laboratories and will be reported in due time.
Experimental Section
1H NMR spectra were recorded on a Varian Mercury 300
NMR spectrometer. Tetramethylsilane was used as internal
reference for 1H NMR, 85% phosphoric acid as external
reference for 31P NMR, and CF3Cl as external reference for
19F NMR. Pyrophosphate tributylammonium salt, dichlorom-
ethylenediphosphonate disodium salt, imidodiphosphate so-
dium salt, and methylenediphosphonic acid were purchased
from Sigma and used with or without further treatment as
indicated in the corresponding experiments. Anhydrous sol-
vents purchased from Aldrich were used directly in the
reactions without further treatment.
Purification of AZT P3Ms. AZT P3Ms were purified by
anion exchange (AX) chromatography using a 10 × 160 mm
Mono Q column (Pharmacia). Initial conditions were typically
0-35 mM NaCl. A linear elution gradient was typically
initiated at 0-35 mM NaCl and terminated at 350 mM to 1
M NaCl in two to three column volumes at 6.5 mL/min. A
constant concentration of 50 mM Tris, pH 8, was maintained
throughout the purification. Fractions containing the target
compounds were collected and desalted by reversed-phase
HPLC (RP-HPLC) using a Luna C18 250 × 21 mm column
(Phenomenex) with a flow rate of 10 mL/min. Elution gradients
were generally from 0-20% to 95% methanol in 20-60 min
at a constant concentration of triethylammonium acetate (50
mM). The AZT P3Ms of this work, unless specified, were
purified using both the anion-exchange HPLC and then RP-
HPLC. Fractions containing the desired AZT P3Ms were
collected and lyophilized to give AZT P3M as mixed sodium
and triethylammonium salts. Compounds that did not require
anion exchange-HPLC purification were purified by RP-HPLC
only, using the same conditions as described above. In these
cases, the final AZT P3M products were triethylammonium
salts. Yields of all AZT P3M products of this work were
calculated on the basis of UV absorbance.
LCMS and HPLC Analysis of AZT P3Ms. Mass spectra
and purity of the AZT P3Ms were obtained using on-line
HPLC-mass spectrometry on a ThermoFinnigan (San Jose,
CA) Deca XP plus. A Phenomenex Luna C18(2) or C5, 75 × 2
mm, 3-µm particle size was used for RP-HPLC. A 0-50% linear
gradient (15 min) of acetonitrile in 10 mM N,N′-dimethyl-n-
hexylammonium acetate, pH 7, was performed in series with
mass spectra detection in the negative ionization mode.
Nitrogen gas and a pneumatic nebulizer were used to generate
the electrospray. The mass range of 150-1500 was typically
sampled. All the purified AZT P3Ms of this work were
subjected to the LCMS analysis as described above.
Preparation of Tetraisopropyl (Difluoromethylene)-
diphosphonate and Tetraisopropyl (Fluoromethylene)-
diphosphonate. Sodium bis(trimethylsilyl)amide (NaHMDS,
1.0 M in THF, 28.7 mL) was added to a stirred solution of
tetraisopropyl methylenediphosphonate (4.5 g, 13.07 mmol)
and N-fluorobenzenesulfonimide (NFSi, 9.89 g, 31.36 mmol)
in anhydrous THF (20 mL) at -78 °C under argon. The
reaction mixture was stirred at -78 °C for 1 h, quenched with
saturated aqueous ammonium chloride (20 mL), warmed to
room temperature, and diluted with ether (60 mL). The organic
layer was separated and the aqueous layer was extracted with
ether (60 mL). Combined organic layer was washed with 10%
Preparation of Difluoromethylenediphosphonic Acid
Bis(tributylammonium) Salt (3d). Trimethylbromosilane
(4.17 mL, 31.58 mmol) was added dropwise to a stirred solution
of tetraisopropyl (difluoromethylene)diphosphonate (2.0 g, 5.26
mmol) in anhydrous acetonitrile (30 mL). The resulting
solution was stirred at 40-42 °C for 24 h under argon,
concentrated to dryness, and coevaporated with anhydrous
acetonitrile. The residue was redissolved in an acetonitrile/
water mixture and then coevaporated with DMF. The residue
was dissolved in a solution of tributylamine (1.93 g, 2.48 mL,
10.43 mmol) in DMF (20 mL), concentrated to dryness, and
coevaporated with anhydrous DMF three times. The residue
was dried at 40 °C under vacuum for 4 h to give a slightly
amber-colored residue (3.35 g): 1H NMR (D2O) δ 0.779 (t, J )
7.5 Hz, Me), 1.23 (m, -CH2Me), 1.52 (m, C-CH2-CMe), 2.97
(t, J ) 8.1 Hz, NCH2-)); 31P NMR δ 3.41 (t, J ) 83.5 Hz); 19
NMR δ -121.47 (t, J ) 83.3).
F
3′-Azido-3′-deoxythymidine 5′-r-PI-Borano-â,γ-(difluo-
romethylene)triphosphates (6d). 2-Chloro-4H-1,3,2-ben-
zodioxaphosphorin-4-one (2.18 g, 10.8 mmol) in DMF (9 mL)
was added to a stirred solution of 3′-azido-3′-deoxythymidine
(2.41 g, 9.0 mmol) in anhydrous DMF (18 mL) and pyridine
(4.5 mL) at 0 °C under argon. The reaction mixture was stirred
at room temperature for 2 h and an additional amount of
2-chloro-4H-1,3,2-benzodioxaphosphorin-4-one (0.37 g, 1.8 mmol)
was added. The reaction mixture was stirred for 5 h at room
temperature and then cooled with ice. Tributylamine (5.4 mL,
49.5 mmol) was added, followed by addition of (difluorometh-
ylene)diphosphonic acid bis(tributylammonium) salt (6.5 g,
11.1 mmol) in DMF (20 mL). The reaction mixture was stirred
at room temperature for 3 h and cooled with ice. Borane-
diisopropylethylamine complex (24 mL) was added, and the
resulting mixture was stirred at room temperature overnight,
cooled with ice, and quenched by slow addition of water (90
mL). The mixture was stirred at room temperature for 2 h,
concentrated at room temperature under reduced pressure to

Synthesis of AZT 5′-Triphosphate Mimics
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6909
a small volume, diluted with water (300 mL), and extracted
with ethyl acetate three times. HPLC purification as described
in the general section yielded 4.82 mmol (53.6%) of 6d-I/II (two
diastereomers) as a colorless foam: 1H NMR (D2O) δ -0.3 to
0.9 (br, BH3, 3H), 1.13 (t, J ) 7.2 Hz, CH3 of TEA, 6H), 1.81
(s, 5-Me, 3H), 2.33 (t, J ) 5.7 Hz, H-2′, 2H), 3.04 (q, J ) 7.2
Hz, CH2 of TEA, 4H), 4.08 (m, H-4′, H-5′, 3H), 4.40 (m, H-3′,
0.5H), 4.48 (m, H-3′, 0.5H), 6.12, 6.13 (2t, J ) 6.9 Hz, H-1′,
1H), 7.59 (s, H-6, 1H); 31P NMR δ -4.09 to -1.86 (m, Pâ), 4.20
(t, J ) 74.2 Hz, 1/2Pγ), 4.68 (t, J ) 74.2 Hz, 1/2Pγ), 83.3 (br,
PR).
To a stirred solution of 3′-azido-3′-deoxythymidine (78 mg.
0.29 mmol) in anhydrous DMF (1 mL) and pyridine (0.2 mL)
at 0 °C under argon was added a solution of 2-chloro-4H-1,3,2-
benzodioxaphosphorin-4-one (77 mg, 0.38 mmol). The reaction
mixture was stirred at room temperature for 1 h and cooled
with ice. Tributylamine (0.18 mL) was added, followed by
addition of (fluoromethylene)diphosphonic acid bis(tributy-
lammonium) salts (250 mg, 0.43 mmol) in DMF (0.6 mL). The
reaction mixture was stirred at room temperature for 1 h and
cooled with ice. Borane-diisopropylethylamine complex (1.20
mL) was added, and the resulting mixture was stirred at room
temperature for 6 h, cooled with ice, and quenched by slow
addition of water (3 mL). The mixture was stirred at room-
temperature overnight, diluted with water (10 mL), and
extracted with chloroform three times. HPLC purification gave
6c in four portions: (a) the pure isomer 6c-I, 7.20 µmol; (b)
the isomer 6c-II (containing other isomers), 15.82 µmol; (c) the
isomer 6c-III (containing other isomers), 11.06 µmol; (d) the
pure isomer 6c-IV, 4.89 µmol. The total yield was 13.4%. 6c-
I: MS m/z 520 (M - H)-; HPLC analysis, 99.4% purity. 6c-
IV: MS m/z 520 (M - H)-; HPLC analysis, 99.6% purity. 6c-
II: 1H NMR (D2O) δ 0.1-0.6 (br, BH3, 3H), 1.78 (s, 5-CH3, 3H),
2.30-2.38 (m, H-2′, 2H), 3.98-4.17 (m, H-4′, H-5′, 3H), 4.36-
4.45 (m, H-3′, 1H), 6.12 (apparent q, J ) 7.0 Hz, H-1′, 1H),
7.60 (s, H-6, 1H); 31P NMR (D2O) δ 1.79 (m, Pâ), 9.65 (dq, 2JP-P
) 14.3 Hz, 2JP-F ) 61.2 Hz, Pγ), 83.80 (m, PR); 19F NMR (D2O)
δ -211.3 (m, CHF).
3′-Azido-3′-deoxythymidine 5′-(r-P-Borano-â,γ-dichlo-
romethylene)triphosphate (6e). An aqueous solution of
(dichloromethylene)diphosphonic acid disodium salt (1.0 g, 3.46
mmol) was loaded on a column of DOWEX 50WX8-100 ion-
exchange resin and eluted with water. Tributylamine (1.65
mL, 6.92 mmol) was added and the mixture was shaken
vigorously. The resulting solution was concentrated to dryness
and coevaporated with anhydrous DMF three times. The
residue was dried under vacuum overnight to give (dichlo-
romethylene)diphosphonic acid bis(tributylammonium) salt as
a colorless semisolid.
The two diastereoisomers (R-PI and R-PII) were separated
on reverse-phase HPLC under the conditions as described in
the general section to give 3′-azido-3′-deoxythymidine R-PI-
borano-â,γ-(difluoromethylene)triphosphate (6d-I, the isomer
that has shorter retention time is designated as diastereomer
I) and 3′-azido-3′-deoxythymidine R-PII-borano-â,γ-(difluorom-
ethylene)triphosphate (6d-II, the isomer that has longer
retention time is designated as diastereomer II). 6d-I: 1H NMR
(D2O) δ 0.14-0.56 (br, BH3, 3H), 1.81 (d, J ) 0.9 Hz, 5-CH3,
3H), 2.32 (t, J ) 5.4 Hz, H-2′, 2H), 4.02-4.11(m, H-4′, H-5′,
3H), 4.48-4.49 (m, H-3′, 1H), 6.13 (t, J ) 7.2 Hz, H-1′, 1H),
7.60 (d, J ) 1.2 Hz, H-6, 1H); 31P NMR (D2O) δ - 4.47 to -
2
2
3.99 (m, Pâ), 4.37 (ddt, JP-P ) 59.1 Hz, JP-F ) 81.6 Hz, Pγ),
84.02 (br, PR); 19F NMR (D2O) δ -119.86 (m, CF2); MS m/z
538 (M - H)-; HPLC analysis, 97.8% purity. 6d-II: 1H NMR
(D2O) δ 0.14-0.48 (m, BH3, 3H), 1.81 (s, 5-CH3, 3H), 2.31-
2.35 (t, J ) 4.8 Hz, H-2′, 2H), 4.03-4.12 (m, H-4′, H-5′, 3H),
4.40-4.41 (m, H-3′, 1H), 6.12 (t, J ) 6.9 Hz, H-1′, 1H), 7.61
(d, J ) 1.2 Hz, H-6, 1H); 31P NMR (D2O) δ - 5.52 to - 3.31
(m, Pâ), 4.40 (ddt, 2JP-P ) 59.6 Hz, 2JP-F ) 81.6 Hz, Pγ), 85.73
(br, PR); 19F NMR (D2O) δ -119.44 (ddt, 2JF-P ) 81.3 Hz, 2JF-P
2
) 81.3 Hz, JF-P 69.5 Hz, CF2); MS m/z 538 (M - H)-; HPLC
analysis, 98.4% purity.
3′-Azido-3′-deoxythymidine 5′-r-P-Borano-â,γ-methyl-
enetriphosphate (6b). 2-Chloro-4H-1,3,2-benzodioxaphos-
phorin-4-one (101 mg, 0.495 mmol) in anhydrous DMF (0.5
mL) was added via syringe to a stirred solution of 3′-azido-3′-
deoxythymidine (134 mg, 0.5 mmol) in 1 mL of anhydrous
DMF and 0.25 mL of anhydrous pyridine under argon. After
stirring at room temperature for 1 h, tributylamine (0.3 mL)
was added, followed by a mixture of methylenediphosphonic
acid (88 mg, 0.495 mmol) and tributylamine (0.35 mL) in 0.5
mL of anhydrous DMF. The reaction mixture was stirred for
1 h, and 2 mL of borane-diisopropylethylamine complex was
added. After stirring at room temperature for 6 h, the reaction
mixture was cooled, quenched with water (5 mL), and stirred
at room temperature for 3 h. Purification by HPLC yielded
91.45 µmol (18.3%) of 6b-I/II (two diastereomers): 1H NMR
(D2O) δ 1.76, 1.80 (2s, 2 isomers, 5-CH3, 3H), 2.10-2.23 (m,
H-2′, 2H), 2.30-2.34 (m, 2H, â,γ-CH2, 2 isomers) 4.01-4.10
(m, H-4′, H-5′, H-5′′, 3H), 4.39-4.46 (m, H-3′, 1H), 6.12
(apparent q, J ) 7.1 Hz, H-1′, 1H), 7.59-7.60 (2s, 2 isomers,
To a stirred solution of 3′-azido-3′-deoxythymidine (89 mg.
0.33 mmol) in anhydrous DMF (1 mL) and pyridine (0.2 mL)
at 0 °C under argon was added a solution of 2-chloro-4H-1,3,2-
benzodioxaphosphorin-4-one (83 mg, 0.41 mmol). The reaction
mixture was stirred at room temperature for 1 h and cooled
with ice. Tributylamine (0.2 mL) was added, followed by
addition of (dichloromethylene)diphosphonic acid bis(tributy-
lammonium) salt (430 mg, 0.43 mmol) in DMF (1 mL). The
reaction mixture was stirred at room temperature for 1 h and
cooled with ice. Borane-diisopropylethylamine complex (1.32
mL) was added and the resulting mixture was stirred at room
temperature for 6 h, cooled with ice, and quenched by slow
addition of water (3 mL). The mixture was stirred at room
temperature overnight, diluted with water (10 mL), and
extracted with chloroform three times. HPLC purification gave
14.51 µmol (4.4%) of 6e-I and 19.76 µmol (6.0%) of 6e-II. 6e-I:
1H NMR (D2O) δ -0.1 to 0.7 (br, BH3, 3H), 1.80 (s, 5-CH3, 3H),
2.25-2.43 (m, H-2′, 2H), 4.01-4.26 (m, H-4′, H-5′, 3H), 4.49-
4.58 (m, H-3′, 1H), 6.11 (t, J ) 7.0 Hz, H-1′, 1H), 7.61 (s, H-6,
H-6, 1H); 31P NMR (D2O) δ 9.10-9.40 (dd, 2JP-P ) 90 Hz, 2JP-H
,
2
15 Hz, Pâ), 15.93 (dt, JP-P ) 59.3 Hz, Pγ), 84.50 (m, PR); MS
m/z 502 (M - H)-; HPLC analysis, 99.9% purity.
2
1H); 31P NMR (D2O) δ -2.11 (apparent q, JP-P ) 19.2 Hz,
2
3′-Azido-3′-deoxythymidine 5′-(r-P-Borano-â,γ-fluo-
romethylene)triphosphate (6c). Trimethylbromosilane (0.70
mL, 5.28 mmol) was added dropwise to a stirred solution of
tetraisopropyl (fluoromethylene)diphosphonate (320 mg, 0.88
mmol) in 1,2-dichloroethane (5 mL). The resulting solution was
stirred at 40-42 °C for 24 h, and 3 mL of anhydrous toluene
was added. The solution was concentrated to dryness and
coevaporated with toluene once. The residue was redissolved
in DMF (3 mL)/water (2 mL) and concentrated. The residue
was dissolved in a mixture of DMF (2 mL) and tributylamine
(0.42 mL, 1.76 mmol) and then concentrated to dryness. The
residue was coevaporated with anhydrous DMF twice. The
resulting residue was dried in a vacuum oven at 30 °C
overnight to give (fluoromethylene)diphosphonic acid bis-
(tributylammonium) salts as a slightly amber residue (460 mg).
Pâ), 9.33 (d, JP-P ) 18.4 Hz, Pγ), 83.68 (m, PR); MS m/z 570
(M - H)-; HPLC analysis, 96.8% purity. 6e-II: MS m/z 571
(M - H)-; HPLC analysis, 89.0% purity.
3′-Azido-3′-deoxythymidine 5′-r-P-Borano-â,γ-imido-
triphosphate (6f). Commercially available imidodiphosphate
sodium salt (1.0 g) in 10 mL of water was loaded on a column
which was filled with 20 g of DOWEX 50W ×8-100 acidic
resin that was thoroughly washed with water/pyridine and
then water. The column was eluted with water. Fractions
containing imidodiphosphate were collected and made alkaline
with tributylamine (3.6 mL). The solution was concentrated
at room temperature under reduced pressure, coevaporated
with anhydrous DMF three times, and dried under vacuum
overnight to give 2.80 g of imidodiphosphate tetrakis(tributy-
lammonium) salt (3f) as a semisolid.

6910 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Wang et al.
2-Chloro-4H-1,3,2-benzodioxaphosphorin-4-one (93 mg, 0.46
mmol) in anhydrous DMF (1 mL) was added to a stirred
solution of 3′-azido-3′-deoxythymidine (93 mg, 0.35 mmol) in
anhydrous DMF (1 mL) and anhydrous pyridine (0.2 mL) at 0
°C under argon. The reaction mixture was stirred at room
temperature for 1 h and then cooled with ice. Tributylamine
(0.2 mL) was added, followed by addition of imidodiphosphonic
acid tetrakis(tributylammonium) salt (440 mg, 0.48 mmol) in
DMF (1 mL). The reaction mixture was stirred at room
temperature for 1 h and cooled with ice. Borane-diisopropy-
lethylamine complex (1.4 mL) was added, and the resulting
mixture was stirred at room temperature for 6 h, cooled with
ice, and quenched by slow addition of water (3 mL). The
mixture was stirred at room temperature overnight, concen-
trated at room temperature under reduced pressure to a small
volume, diluted with water, and extracted with ethyl acetate
three times. HPLC purification as described in the general
section yielded 5.75 µmol (1.6%) of 6f-I/II (two diastereo-
mers): 1H NMR (D2O, TEA+ form) δ 1.80 (s, 5-CH3, 3H), 2.29-
2.38 (m, H-2′, 2H), 3.99-4.13 (m, H-4′, H-5′, 3H), 4.36-4.46
(m, H-3′, 1H), 6.13 (t, J ) 6.9 Hz, H-1′, 1H), 7.59 (s, H-6, 1H);
31P NMR (D2O) δ -9.73 (m, Pâ), -0.55, -0.42 (2d, 2 isomers,
2JP-P ) 9.8 Hz, Pγ), 83.80 (m, PR); MS m/z 503 (M - H)-; HPLC
analysis, 90.1% purity.
A General Procedure for Preparation of AZT 5′-γ-
Substituted-â,γ-(difluoromethylene)triphosphates 7a-
7i. A freshly prepared solution of 2-chloro-4H-1,3,2-benzodiox-
aphosphorin-4-one (0.165-0.5 mmol, 1.0-1.2 equiv) in anhy-
drous DMF (0.5-1.0 mL) was added to a stirred solution of
3′-azido-3′-deoxythymidine (40-134 mg, 0.15-0.5 mmol) in
anhydrous DMF (0.5-1.0 mL) and anhydrous pyridine (0.1-
0.25 mL) at 0 °C under argon. After stirring at room temper-
ature for 1 h, tributylamine (0.21 to 0.3 mL) was added,
followed by addition of difluoromethylenediphosphonic acid
bis(tributylammonium) salt (0.18-0.65 mmol, 1.2-1.3 equiv)
in anhydrous DMF (0.5-1.0 mL). The resulting mixture was
stirred for 1 h and cooled with ice, and iodine (2 equiv, 0.3-
1.0 mmol) was added. After stirring at room temperature for
1 h, a nucleophilic reagent (neat or in 0.5-1.0 mL of DMF)
was added. The reaction mixture was stirred at room temper-
ature for 1-2 h, cooled with ice, quenched with water, and
stirred at room temperature for 30 min. HPLC purification
yielded 7a-7i, respectively. The products were purified on
anion-exchange and reverse-phase HPLC as described in the
general section.
argon. After stirring at room temperature for 1 h, tributyl-
amine (0.42 mL) was added, followed by a solution of (difluo-
romethylene)diphosphonic acid bis(tributylammonium) salt
(209.6 mg, 0.36 mmol) in anhydrous DMF (1.0 mL). The
mixture was stirred at room temperature for 1 h and cooled
to 0 °C, and iodine (152.2 mg, 0.60 mmol) was added. After
stirring at room temperature for 1 h, ammonia (5 mL, freshly
condensed) was added, and stirring was continued at room
temperature for 1 h. Excess ammonia was evaporated and the
solution was then diluted with water. HPLC purification gave
79.07 µmol (26.4%) of 7b: 1H NMR (D2O) δ 1.78 (s, 5-CH3, 3H),
2.32-2.36 (m, H-2′, 2H), 4.05-4.08 (m, H-4′, H-5′, 3H), 4.41-
4.45 (m, H-3′, 1H), 6.13 (t, J ) 7.0 Hz, H-1′, 1H), 7.61 (s, H-6,
2
1H); 31P NMR (D2O) δ -10.40 (d, JP-P ) 30.0 Hz, PR), -4.04
2
2
(m, Pâ), 9.50 (dt, JP-P ) 59.2 Hz, JP-F ) 79.3 Hz, Pγ); 19F
2
NMR (D2O) δ -121.2 (apparent q, JF-P ) 82.3 Hz, CF2); MS
m/z 539 (M - H)-; HPLC analysis, 100% purity.
3′-Azido-3′-deoxythymidine 5′-â,γ-Difluoromethylene-
γ-P-methylaminotriphosphate (7c). Reaction using 133.6
mg (0.5 mmol) of 3′-azido-3′-deoxythymidine and 2.0 M me-
thylamine in THF (5 mL) as nucleophilic reagent yielded, after
HPLC purification, 0.126 mmol (25.3%) of 7c: 1H NMR (D2O)
δ 1.97 (s, 5-CH3, 3H), 2.32-2.42 (m, H-2′, 2H), 2.74-2.85 (m,
H-5′, 1H), 3.22-3.25 (m, H-5′′, 1H), 3.55 (s, CH3, 3H), 4.04-
4.06 (m, H-4′, 1H), 4.40-4.44 (m, H-3′, 1H), 6.09-6.14 (dd, J
) 6.9, 13.8 Hz, H-1′, 1H), 7.56 (s, H-6, 1H); 31P NMR (D2O) δ
2
-10.45 (d, JP-P ) 78 Hz, PR), -3.89 (m, Pâ), 8.88 (dt, JP-F
)
189 Hz, 2JP-P ) 78 Hz, Pγ); 19F NMR (D2O) -119.22 (ddd, 2JF-P
) 82 Hz, 2JF-P ) 90 Hz, 3JF-F ) 8 Hz, CF2); MS m/z 553 (M -
H)-; HPLC analysis, 92.6% purity.
3′-Azido-3′-deoxythymidine 5′-â,γ-Difluoromethylene-
γ-P-ethylaminotriphosphate (7d). Reaction using 133.6 mg
(0.5 mmol) of 3′-azido-3′-deoxythymidine and 2.0 M ethylamine
in THF (2 mL) as nucleophilic reagent yielded, after HPLC
1
purification, 95.42 µmol (19.1%) of 7d: H NMR (D2O) δ 1.05
(t, CH3, 3H), 1.76, (s, 5-CH3, 3H), 2.19-2.35 (m, H-2′, 2H),
2.77-2.88 (m, CH2), 4.04-4.06 (m, H-4′, H-5′, H-5′′, 3H), 4.40-
4.44 (m, H-3′, 1H), 6.09-6.14 (dd, J ) 6, 15 Hz, H-1′, 1H),
2
7.58 (s, H-6, 1H); 31P NMR (D2O) δ -10.45 (d, JP-P ) 78 Hz,
PR), -3.89 (m, Pâ), 8.88 (dt, JP-F ) 189 Hz, 2JP-P ) 78 Hz, Pγ);
2
2
19F NMR (D2O) -119.39 (ddd, JF-P ) 81 Hz, JF-P ) 81 Hz,
3JF-F ) 12 Hz, CF2); MS m/z 567 (M - H)-; HPLC analysis,
96.4% purity.
3′-Azido-3′-deoxythymidine â,γ-Difluoromethylene-5′-
γ-P-phenylaminotriphosphate (7e). Reaction using 103.8
mg (0.388 mmol) of 3′-azido-3′-deoxythymidine and aniline
(0.55 mL, 3.88 mmol) as nucleophilic reagent yielded, after
HPLC purification, 26.30 µmol (6.8%) of 7e: 1H NMR (D2O) δ
1.72 (s, 5-CH3, 3H), 2.18-2.22 (m, H-2′, 2H), 3.82-4.01 (m,
H-4′, H-5′, 3H), 4.22-4.32 (m, H-3′, 1H), 6.03 (t, J ) 6.3 Hz,
H-1′, 1H), 6.7-6.78 (m, Ph, 1H), 6.97-7.03 (m, Ph, 4H), 7.45
(s, H-6, 1H); 31P NMR (D2O) δ -10.10 (d, 2JP-P ) 30.0 Hz, PR),
3′-Azido-3′-deoxythymidine 5-(â,γ-difluoromethylene)-
triphosphate (7a). A freshly prepared solution of 2-chloro-
4H-1,3,2-benzodioxaphosphorin-4-one (101.27 mg, 0.495 mmol)
in anhydrous DMF (0.5 mL) was added via syringe to a stirred
solution of 3′-azido-3′-deoxythymidine (133.63 mg, 0.495 mmol)
and anhydrous pyridine (0.25 mL) in 1 mL of anhydrous DMF
under argon. After stirring at room temperature for 1 h,
tributylamine (0.3 mL) was added, followed by a solution of
bis(tributylammonium) difluoromethylenediphosphate (co-
evaporated with anhydrous pyridine two times) in 0.5 mL of
anhydrous DMF. The mixture was stirred for 1 h and 253.8
mg (1 mmol) of iodine was added. After stirring at room
temperature for 3.5 h, the reaction was quenched with 5 mL
of water and stirred at room temperature for 4 h. HPLC
purification afforded 22.55 µmol (4.6%) of 7a: 1H NMR (D2O)
δ 1.34 (t, CH3, 3H), 1.98 (s, 5-CH3, 3H), 2.32-2.37 (m, H-2′,
1H), 2.80-2.82 (m, H-2′, 1H), 3.28-3.30 (m, 5′CH2, 1H), 4.06-
4.08 (m, H-4′, H-5′′ 2H), 4.40-4.43 (m, H-3′, 1H), 6.10-6.15
(dd, J ) 6.9, 13.8 Hz, H-1′, 1H), 7.59 (s, H-6, 1H); 31P NMR
2
2
-4.14 (m, Pâ), 2.56 (dt, JP-P ) 58.0 Hz, JP-F ) 80.0 Hz, Pγ);
19F NMR (D2O) δ -119.29 (t, JF-P ) 83.0 Hz, CF2); MS m/z
2
615 (M - H)-; HPLC analysis, 92.5% purity.
3′-Azido-3′-deoxythymidine 5′-γ-P-Azido-â,γ-(difluo-
romethylene)triphosphate (7f). Reaction using 40.09 mg
(0.15 mmol) of 3′-azido-3′-deoxythymidine and sodium azide
(97.5 mg, 1.5 mmol) as nucleophilic reagent yielded, after
HPLC purification, 29.50 µmol (19.7%) of 7f: 1H NMR (D2O) δ
1.74 (s, 5-CH3, 3H), 2.31-2.35 (m, H-2′, 2H), 4.01-4.10 (m,
H-4′, H-5′, 3H), 4.41-4.46 (m, H-3′, 1H), 6.11 (t, J ) 6.8 Hz,
H-1′, 1H), 7.59 (s, H-6, 1H); 31P NMR (D2O) δ -10.40 (d, 2JP-P
) 30.1 Hz, PR), -6.14 (m, Pâ), 4.08 (dt, 2JP-P ) 63.3 Hz, 2JP-F
2
2
) 86.1 Hz, Pγ); 19F NMR (D2O) δ -120.2 (t, JF-P ) 84.2 Hz,
(D2O) δ -10.77 (d, JP-P ) 75.3 Hz, PR), -3.95 (m, Pâ), 4.16
(dt, JP-F ) 144 Hz, 2JP-P ) 75 Hz, Pγ); 19F NMR (D2O) -119.92
CF2); MS m/z 565 (M - H)-; HPLC analysis, 89.3% purity.
2
2
3
(ddd, JF-P ) 84 Hz, JF-P ) 92 Hz, JF-F ) 8 Hz, CF2); MS
3′-Azido-3′-deoxythymidine 5′-â,γ-Difluoromethylene-
γ-P-fluorotriphosphate (7g). Reaction using 80.2 mg (0.30
mmol) of 3′-azido-3′-deoxythymidine and sodium azide (174.2
mg, 3.0 mmol) as nucleophilic reagent yielded, after HPLC
m/z 540 (M - H)-; HPLC analysis, 98.3% purity.
3′-Azido-3′-deoxythymidine 5′-γ-P-Amino-â, γ-(difluo-
romethylene)triphosphate (7b). 2-Chloro-4H-1,3,2-benzo-
dioxaphosphorin-4-one (66.8 mg, 0.33 mmol) in anhydrous
DMF (1.0 mL) was added via syringe to a solution of 3′-azido-
3′-deoxythymidine (80.2 mg, 0.30 mmol) in 1.0 mL of anhy-
drous DMF and 0.10 mL of anhydrous pyridine at 0 °C under
1
purification, 73.35 µmol (24.4%) of 7g: H NMR (D2O) δ 1.78
(s, 5-CH3, 3H), 2.30-2.36 (m, H-2′, 2H), 4.03-4.10 (m, H-4′,
H-5′, 3H), 4.42-4.46 (m, H-3′, 1H), 6.13 (t, J ) 7.0 Hz, H-1′,
2
1H), 7.60 (s, H-6, 1H); 31P NMR (D2O) δ -10.50 (d, JP-P
)

Synthesis of AZT 5′-Triphosphate Mimics
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6911
2
2
29.8 Hz, PR), -6.43 (ddt, JP-F ) 82.6 Hz, JP-P ) 62.7 Hz,
diphosphonic acid bis(tributylammonium) salt (294 mg, 0.504
mmol) in anhydrous DMF (1 mL). The mixture was stirred at
room temperature for 1 h and cooled with ice, and sulfur (24.8
mg, 0.776 mmol) was added. After stirring at room tempera-
ture for 2 h, the reaction mixture was quenched with water
(10 mL) and stirred at room temperature for 30 min. Purifica-
tion by HPLC yielded 0.108 mmol (27.7%) of 13: 1H NMR (D2O)
δ 1.84 (s, 5-CH3, 3H), 2.29-2.35 (m, H-2′, 2H), 4.02-4.12 (m,
H-4′, H-5′, 3H), 4.39-4.43 (m, H-3′, 1H), 6.09 (t, J ) 6.9 Hz,
H-1′, 1H), 7.60 (s, H-6, 1H); 31P NMR (D2O) δ -1.83 (m, Pâ),
6.87 (dt, 2JP-P ) 60.0 Hz, 2JP-F ) 80.4 Hz, Pγ), 100.61 (apparent
2JP-P ) 29.3 Hz, Pâ), 5.23 (ddt, JP-F ) 1024.9 Hz, JP-F
)
1
2
91.3 Hz, 2JP-P ) 63.2 Hz, Pγ); 19F NMR (D2O) δ -121.40 (ddd,
2JF-P ) 91.1 Hz, 2JF-P ) 83.2 Hz, 3JF-F ) 4.0 Hz, CF2), -68.75
(dt, 1JF-P ) 1021.0 Hz, 3JF-F ) 4.0 Hz, F-Pγ); MS m/z 543 (M
- H)-; HPLC analysis, 99.8% purity.
3′-Azido-3′-deoxythymidine 5′-â,γ-Difluoromethylene-
γ-O-methyltriphosphate (7h). Reaction using 40.09 mg (0.15
mmol) of 3′-azido-3′-deoxythymidine and sodium methoxide
(1.5 mmol in methanol) as nucleophilic reagent yielded, after
1
HPLC purification, 18.20 µmol (12.1%) of 7h: H NMR (D2O)
2
2
q, JP-P ) 28.2 Hz, PR); 19F NMR (D2O) δ -121.1 (t, JF-P
)
δ 1.90 (s, 5-CH3, 3H), 2.30-2.38 (m, H-2′, 2H), 4.02-4.07 (m,
H-4′, H-5′, 3H), 4.41-4.46 (m, H-3′, 1H), 6.11 (t, J ) 6.8 Hz,
H-1′, 1H), 7.58 (s, H-6, 1H); 31P NMR (D2O) δ -10.32 (d, 2JP-P
) 30.6 Hz, PR), -5.05 (m, Pâ), 6.15 (dt, 2JP-P ) 59.6 Hz, 2JP-F
81.3 Hz, CF2); MS m/z 556 (M - H)-; HPLC analysis, 100%
purity.
3′-Deoxythymidine 5′-â,γ-Difluoromethylene-r-P-dithio-
triphosphate (14). 2-Chloro-4H-1,3,2-benzodioxaphosphorin-
4-one (101.2 mg, 0.495 mmol) in anhydrous DMF (1 mL) was
added via syringe to a stirred solution of 3′-azido-3′-deoxythy-
midine (103.8 mg, 0.388 mmol) in 1 mL of anhydrous DMF
and 0.20 mL of anhydrous pyridine at 0 °C under argon. After
stirring at room temperature for 1 h, tributylamine (0.3 mL)
was added, followed by (difluoromethylene)diphosphonic acid
bis(tributylammonium) salt (294 mg, 0.504 mmol) in anhy-
drous DMF (1 mL). The mixture was stirred at room temper-
ature for 1 h and cooled with ice, and sulfur (24.8 mg, 0.776
mmol) was added. After stirring at room temperature for 2 h,
lithium sulfide (89.1 mg, 1.94 mmol) was added and stirring
was continued for 1 h. Water (10 mL) was added, and the
mixture was stirred at room temperature for 30 min. Purifica-
tion on HPLC yielded 12.34 µmol (3.2%) of 14: 1H NMR (D2O)
δ 1.83 (s, 5-CH3, 3H), 2.31-2.40 (m, H-2′, 2H), 4.12-4.22 (m,
H-4′, H-5′, 3H), 4.50-4.71 (m, H-3′, 1H), 6.14 (t, J ) 6.8 Hz,
H-1′, 1H), 7.70 (s, H-6, 1H); 31P NMR (D2O) δ -4.75 (m, Pâ),
4.25 (dt, 2JP-P ) 61.2 Hz, 2JP-F ) 81.6 Hz, Pγ), 100.61 (d, 2JP-P
) 81.6 Hz, Pγ); 19F NMR (D2O) δ -119.4 (t, JF-P ) 83.2 Hz,
2
CF2); MS m/z 554 (M - H)-; HPLC analysis, >95% purity.
3′-Azido-3′-deoxythymidine 5′-â,γ-Difluoromethylene-
γ-O-phenyltriphosphate (7i). Reaction using 59.0 mg (0.22
mmol) of 3′-azido-3′-deoxythymidine and sodium phenoxide
(2.2 mmol in DMF) as nucleophilic reagent yielded, after HPLC
purification, 3.65 µmol (1.7%) of 7I: MS m/z 616 (M - H)-;
HPLC analysis, 95.4% purity.
3′-Azido-3′-deoxythymidine 5′-r-P-Borano-â,γ-difluo-
romethylene-γ-O-methyltriphosphate (11). A solution of
the bis(tetrabutylammonium) salts of 6d-I/II (0.12 mmol),
tributylamine (85 µL, 0.36 mmol), and methyl trifluoromethane-
sulfonate (54 µL, 0.48 mmol) in anhydrous acetonitrile (2 mL)
stood at room temperature overnight. The reaction was
quenched with water (2 mL) at 0 °C and then most of the
acetonitrile was evaporated. The aqueous solution was subject
to reverse-phase HPLC purification to give 2.74 µmol (2.3%)
of 11: MS m/z 551 (M - H)-; HPLC analysis, 94.0% purity.
3′-Azido-3′-deoxythymidine 5′-r-P-Borano-â,γ-difluo-
romethylene-γ-O-phenyltriphosphate (12). A slurry of
sodium hydride (320 mg, 8.0 mmol) in anhydrous DMF (10
mL) was treated with phenol (0.9 g, 10.0 mmol) dissolved in
anhydrous DMF (2 mL) under an argon atmosphere. After
stirring for 1 h at room temperature, gas evolution had ceased.
15-Crown-5 (1.59 mL, 8.0 mmol) was added and stirring
continued for 1 h. This mixture was used immediately.
) 43.2 Hz, PR); 19F NMR (D2O) δ -120.6 (apparent q, 2JF-P
)
81.3 Hz, CF2); MS m/z 572 (M - H)-; HPLC analysis, 97.3%
purity.
Preparation of 3′-Azido-3′-deoxythymidine 5′-(Difluo-
romethylene)diphosphate. Trimethylbromosilane (4.3 mL,
33.3 mmol) was added to a stirred solution of tetraisopropyl
(difluoromethylene)diphosphonate (2.10 g, 5.51 mmol) in 1,2-
dichloroethane (20 mL) under argon. The reaction mixture was
continued with magnetic stirring at 40 °C for 24 h. Volatiles
were evaporated under reduced pressure, and the residue was
coevaporated with toluene twice. The resulting residue was
dissolved in dichloromethane, cooled with ice, and quenched
with a DMF and water mixture (3:2). After removal of solvent,
the residue was dissolved in water and cooled to 0 °C.
Tetrabutylammonium hydroxide aqueous solution (40 wt %,
10.72 g, 16.52 mmol) was added, and the resulting solution
was concentrated. The residue was coevaporated with DMF
three times and dried at room temperature under vacuum to
give 5.36 g of (difluoromethylene)diphosphonic acid tris-
(tetrabutylammonium) salt as a semisolid, which was coevapo-
rated with pyridine two times and dried under vacuum: 31P
NMR (DMSO) δ 5.59 (t, JP-F ) 70.4 Hz); 19F NMR (DMSO) δ
-118.90 (t, JF-P ) 69.4 Hz).
To a flask containing the (difluoromethylene)diphosphonic
acid tetrabutylammonium salt prepared above (3.86 g, 4.12
mmol) in acetonitrile (4 mL) under argon was added dropwise
a solution of AZT 5′-tosylate29 (358 mg, 0.85 mmol) in aceto-
nitrile (2.5 mL). The resulting solution was stirred for 27 h at
room temperature, quenched by adding 10 mL of water, and
stirred for 5 min. HPLC purification yielded 160 mg (0.177
mmol) of 3′-azido-3′-deoxythymidine 5′-(difluoromethylene)-
diphosphate.
The triethylammonium salt of 6d-I/II (0.21 mmol) was
coevaporated with anhydrous DMF (3 × 5 mL). For the third
coevaporation, approximately half the volume of DMF was
removed and anhydrous methanol (0.3 mL) was added, fol-
lowed by addition of N,N′-dicyclohexylcarbodiimide (DCC, 200
mg, 0.97 mmol). After stirring at room temperature under
argon for 3 h, the mixture was concentrated under reduced
presure. The above procedure was repeated once more using
DMF (3 × 5 mL), methanol (0.3 mL), and DCC. After removal
of solvents, the residue was resuspended in anhydrous DMF
(3 mL) and treated dropwise with the phenoxide solution from
step 1. After stirring at room temperature under argon for 4
h, the reaction mixture was treated with water (20 mL),
adjusted to pH 6 with hydrochloric acid (1.0 M), and extracted
with diethyl ether (3 × 20 mL). The aqueous portion was
purified on reverse-phase HPLC to give 67.15 µmol (32.0%) of
12: 1H NMR (D2O) δ 0.1-0.7 (br, BH3, 3H), 1.75, 1.76 (2s, 2
isomers, 5-CH3, 3H), 2.18-2.29 (m, H-2′, 2H), 3.92-4.11 (m,
H-4′, H-5′, 3H), 4.28-4.38 (m, H-3′, 1H), 6.07 (apparent q, J
) 7.1 Hz, H-1′, 1H), 6.97-7.08 (m, Ph-meta, para, 3H), 7.15-
7.21 (m, Ph-ortho, 2H), 7.52, 7.54 (2s, 2 isomers, H-6, 1H); 31
P
)
NMR (D2O) δ -5.42 (m, Pâ), 2.20 (dt, 2JP-P ) 59.3 Hz, 2JP-F
87.0 Hz, Pγ), 84.70 (m, PR); 19F NMR (D2O) δ -119.45, -119.50
(2t, 2 isomers, JF-P ) 86.0 Hz, CF2); MS m/z 614 (M - H)-;
2
HPLC analysis, 84.6% purity.
3′-Azido-3′-deoxythymidine 5′-â,γ-Difluoromethylene-
r-P-thiotriphosphate (13). 2-Chloro-4H-1,3,2-benzodioxa-
phosphorin-4-one (101.2 mg, 0.495 mmol) in anhydrous DMF
(1 mL) was added via syringe to a stirred solution of 3′-azido-
3′-deoxythymidine (103.8 mg, 0.388 mmol) in 1 mL of anhy-
drous DMF and 0.20 mL of anhydrous pyridine at 0 °C under
argon. After stirring at room temperature for 1 h, tributyl-
amine (0.3 mL) was added, followed by (difluoromethylene)-
3′-Azido-3′-deoxythymidine 5′-â,γ-Difluoromethylene-
γ-P-methyltriphosphate (15). Triethylamine (37 µL, 0.265
mmol) was added to a stirred solution of 1,2,4-1H-triazole (18.3
mg, 0.265 mmol) in anhydrous acetonitrile (0.5 mL). The
solution was cooled to 0 °C, and a solution of methylphosphonic
dichloride (17.6 mg, 0.133 mmol) in acetonitrile (0.5 mL) was
added dropwise. The reaction mixture was stirred at room
temperature for 40 min and then centrifuged. The supernatant

6912 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Wang et al.
was added to a solution of the tributylammonium salt of 3′-
azido-3′-deoxythymidine 5′-(difluoromethylene)diphosphonate
(89.8 mg, 0.088 mmol) in DMF (1 mL). The reaction mixture
was stirred for 1.5 h and then quenched with water (2 mL).
Purification by reverse-phase HPLC gave 24.30 µmol (27.6%)
of 15: 1H NMR (D2O) δ 1.30 (d, Pγ-CH3, 3H, JCH3-P ) 17.6 Hz),
1.76 (s, 5-CH3, 3H), 2.29-2.33 (m, H-2′, 2H), 4.01 (m, H-4′,
1H), 4.10 (m, H-5′, 2H), 4.38-4.42 (m, H-3′, 1H), 6.06-6.11
(t, H-1′, 1H), 7.56 (s, H-6, 1H); 31P NMR (D2O) δ -5.53 (m, Pâ,
JPâ-γ ) 31.7 Hz, JPâ-R ) 61.0 Hz, JPâ-F ) 86.3 Hz), 4.43 (dt,
PR, JP-P ) 59.2 Hz, JP-F ) 81.4 Hz), 19.47 (d, Pγ, JP-P ) 31.3
Hz); 19F NMR (D2O) δ -119.58 (t, JP-F ) 83.2 Hz); MS m/z
538 (M - H)-; HPLC analysis, 100% purity.
3′-Azido-3′-deoxythymidine 5′-r,â-Difluoromethylene-
γ-P-phenyltriphosphate (16). Triethylamine (37.2 µL, 0.265
mmol) was added to a stirred solution of 1,2,4-1H-triazole (18.4
mg, 0.267 mmol) in anhydrous acetonitrile (0.5 mL). The
solution was cooled to 0 °C, and a solution of phenylphosphonic
dichloride (26 mg, 0.134 mmol) in acetonitrile (0.5 mL) was
added dropwise. The reaction mixture was stirred at 4 °C for
40 min and then centrifuged. The supernatant was added to
a solution of the tributylammonium salt of 3′-azido-3′-deoxy-
thymidine 5′-(difluoromethylene)diphosphonate (90.3 mg, 0.089
mmol) in DMF (1 mL). Similar workup and purification as
described for 15 gave 21.29 µmol (23.9%) of 16: 1H NMR (D2O)
δ 1.71 (s, 5-CH3, 3H), 2.23 (m, H-2′, 2H), 3.93 (m, H-4′, 1H),
4.00 (m, H-5′, 2H), 4.27 (m, H-3′, 1H), 6.03-6.08 (t, H-1′, 1H),
7.27-7.33 and 7.61-7.68 (2m, Ph, 5H), 7.48 (s, H-6, 1H); 31P
NMR (D2O) δ -5.24 (m, Pâ, JPâ-γ ) 33.1 Hz, JPâ-R ) 60.3 Hz,
JPâ-F ) 85.3 Hz), 4.64 (dt, PR, JP-P ) 59.9 Hz, JP-F ) 82.8
Hz), 17.61 (d, Pγ, JP-P ) 32.8 Hz); 19F NMR (D2O) δ -119.4 (t,
JP-F ) 83.2 Hz); MS m/z 600 (M - H)-; HPLC analysis, 100%
purity.
HIV Reverse Transcriptase Inhibition Assays. This
assay was used to measure the ability of the AZT P3Ms to
inhibit the enzymatic synthesis of complimentary strand DNA
from a DNA-primed template of homopolymeric RNA. This
assay is a modification of a published procedure.31 Assay buffer
conditions (50 µL total/reaction): 50 mM Tris-HCl, pH 8.1, 6.5
mM MgCl2, 100 mM, NaCl, 10 mM DTT, 5 µM dTTP (thymi-
dine triphosphate), 1 µg/mL primed-poly(A) RNA, 2 nM
purified HIV reverse transcriptase (type B, 66 kDa subunit).
The compounds were tested at various concentrations up to
500 µM final concentration. DNA polymerase activity was
measured in a reaction buffer containing primed-RNA tem-
plate and dTTP diluted to appropriate concentrations in assay
buffer. The AZT P3Ms were diluted in buffer and pipeted into
the wells of a 96-well plate. The reaction was initiated by
addition of enzyme and allowed to proceed at 37 °C for 10 min
and then quenched by addition of 5 µL of 0.2 M EDTA, pH
8.0. Blank reactions were prepared in parallel with the test
reactions in which either enzyme or dTTP was omitted from
the reactions, substituted by an appropriate volume of enzyme
diluent or assay buffer, respectively. Then 200 µL of diluted
PicoGreen dsDNA Quantitation Reagent (Molecular Probes,
Inc, Eugene, OR) was added to each well of a 96-well plate
and incubated at room temperature for 5 min. Plate wells were
read on a microplate fluorometer (Molecular Devices Corp.,
Sunnyvale, CA). The wells were excited at 480 nm and the
fluorescence emission intensity (RFU) was measured at 538
nm. The percentage of inhibition was calculated according to
the equation: % nhibition ) [1 - (RFU in test reaction - RFU
in blank)/(RFU in control reaction - RFU in blank)] × 100.
Inhibition constants (Ki) were determined for compounds that
exhibited g50% inhibition at 10 µM. Each inhibitor was
titrated over an appropriate range of concentrations, and
inhibition constants were determined using a competitive
inhibition (Synergy Software).
pH 7.4, 0.1 mM MgCl2, 500 µM nucleotide mimic, 10% (v/v)
fetal calf serum. The reaction mixtures were incubated at 37
°C. At appropriate times, aliquots of 25 µL were removed and
added to 65 µL of ice-cold methanol. These solutions were
incubated for at least 1 h at -20°C. After incubation, samples
were centrifuged for 20 min at high speed in a microcentrifuge.
The supernatant was transferred to a clean tube and the
extract was dried under vacuum in a LabConco Centrivap
Concentrator. The dried extracts were resuspended in deion-
ized water and filtered to remove particulates and then
analyzed on reverse phase HPLC. The reverse phase HPLC
column used for the analysis was a Phenomenex C18 Aqua
column (3 × 150 mm). The HPLC flow rate was 0.5 mL/min
with the following buffer system: 5 mM tetrabutylammonium
acetate, 50 mM ammonium phosphate, and an acetonitrile
gradient running from 5% up to as high as 60%. The amount
of remaining parent compound at each time point was used
to determine the half-life of the compound. Time points were
only taken through 48 h so that if greater than 50% of a
compound was still intact after 48 h incubation the half-life
was expressed as >48 h. Unmodified nucleoside triphosphates
were used as positive controls.
Stability Assessment Using Cell Extracts. Cell lysis
buffer was added to cell pellets, and the cells were frozen and
thawed three times using dry ice. The lysis buffer (LB) was
composed of the following: 50 mM Tris-HCl, pH 7.4 (100 µL/
mL 10× stock), 20% glycerol (200 µL/mL), and 0.5% Triton
X-100 (5µL/mL). LB (100 µL) was added to each microfuge tube
containing 107 frozen CEM cells. After the cells were lysed,
the extracts were centrifuged at high speed in a microcentri-
fuge for 5 min, and the clarified cell extract was transferred
to a new tube. The cell stability reaction mixtures contained
concentrations of buffer, magnesium, nucleotide, and cell
extract as shown below: 50 mM Tris-HCl, pH 7.4, 0.1 mM,
MgCl2, 500 µM nucleotide mimic (or control nucleotide), 50%
cell extract (v/v). The reaction mixtures were incubated at 37
°C. At each time point, including at time zero, 12.5 µL aliquots
were added to 40 µL of ice-cold methanol. Typically, time points
were taken after 30 and 60 min and 2, 3, 8, 24, and 48 h. The
samples were incubated in methanol for at least 60 min and
typically overnight at -20°C prior to further processing. The
cell extracts were centrifuged at high speed in a refrigerated
microcentrifuge for 20 min, and the supernatant was trans-
ferred to a new tube. The extract was then dried under vacuum
in a LabConco Centrivap Concentrator. The samples were then
resuspended in 40 or 50 µL of deionized water, filtered to
remove any particulate, and analyzed by reverse phase HPLC.
The reverse phase HPLC columns used for the analysis were
either a Phenomenex C18 Aqua column (2 × 100 mm) or the
Phenomenex C18 Aqua column (3 × 150 mm) used with the
appropriate guard column. The HPLC flow rate was either 0.2
mL/min (for the 2 × 100 mm column) or 0.5 mL/min (for the
3 × 150 mm column) with the following buffer system: 5 mM
tetrabutylammonium acetate, 50 mM ammonium phosphate,
and an acetonitrile gradient running from 5% up to as high
as 60%. The amount of remaining parent compound at each
time point was used to determine the half-life of the compound.
Time points were only taken through 48 h so that if greater
than 50% of a compound was still intact after 48 h incubation
the half-life was expressed as >48 h. Unmodified nucleoside
triphosphates were used as positive controls.
Acknowledgment. The authors wish to thank Drs.
Kandasamy Sakthivel and Gregory J. Ewing, for their
helpful suggestions and discussions of chemistry, and
Dr. John Lambert, for his critical reading and sugges-
tions in the preparation of the manuscript.
Serum Stability Assessment Using Serum. The stability
of nucleotide mimics was assessed in fetal calf serum, generally
following a published procedure.20 Fetal calf serum purchased
from HyClone Corporation was mixed 1:1 with each compound
containing Tris-HCl buffer and MgCl2. The final concentrations
of the reaction components were as follows: 50 mM Tris-HCl,
References
(1) Herdewijn, P.; Balzarini, J.; De Clercq, E. 2′,3′-Dideoxynucleoside
Analogues as Antiviral Agents. In Advances in Antiviral Drug
Design; De Clercq, E., Ed.; JAI Press: Greenwich, 1993; Vol. 1.

Synthesis of AZT 5′-Triphosphate Mimics
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6913
(2) St. Claire, 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.
(3) Parker, W. B.; White, E. L.; Shaddix, S. C.; Ross, L. J.; Buckheit,
R. W., Jr.; Germany, J. M.; Secrist, J. A., III; Vince, R.; Shannon,
W. H. Mechanism of Inhibition of Human Immunodefficiency
Virus Type 1 Reverse Transcriptase and Human DNA Poly-
merases R, â, γ by the 5′-Triphosphates of Carbovir, 3′-Azido-
3′-deoxythymidine, 2′,3′-Dideoxyguanosine, and 3′-dexoythymi-
dine. J. Biol. Chem. 1991, 266, 1754-1762.
(4) Jeffrey, J. L.; Feng, J. Y.; Qi, C. C. R.; Anderson, K. S.; Furman,
P. A. Dioxolane Guanosine 5′-Triphosphate, an Alternative
Substrate Inhibitor of Wild-Type and Mutant HIV-1 Reverse
Transcriptase. J. Biol. Chem. 2003, 278, 18971-18979.
(5) Perigaud, C., Gosselin, G., Girardet, J.-L., Korba, B. E., Imbach,
J. L. The S-Acyl-2-thioethyl Pronucleotide Approach Applied to
Acyclovir: Part I. Synthesis and in vitro Anti-Hepatitis B Virus
Activity of Bis(S-acyl-2-thioethyl)phosphotriester Derivatives of
Acyclovir. Antiviral Res. 1999, 40, 167-78.
(6) Wagner, C. R.; Iyer, V. V.; McIntee, E. J. Pronucleotides: Toward
the In Vivo Delivery of Antiviral and Anticancer Nucleotides.
Med. Res. Rev. 2000, 20, 417-451.
(7) Farquhar, D.; Khan, S.; Srivastva, D. N.; Saunders, P. P.
Synthesis and Antitumor Evaluation of Bis[(pivaloyloxy)methyl]
2′-deoxy-5-fluorouridine 5′-Monophosphate (FdUMP): A Strat-
egy to Introduce Nucleotides into Cells. J. Med. Chem. 1994,
37, 3902-3909.
(8) Lefebvre, I.; Perigaud, C.; Pompon, A.; Aubertin, A.-M.; Girardet,
J.-L.; Kirn, A.; Gosselin, G.; Imbach, J.-L. Mononucleoside
Phosphotriester Derivatives with S-Acyl-2-thioethyl Biorevers-
ible Phosphate-Protecting Groups: Intracellular Delivery of 3′-
Azido-2′,3′-dideoxythymidine 5′-Monophosphate. J. Med. Chem.
1995, 38, 3941-3950.
(9) Siddiqui, A. Q.; McGuigan, C.; Ballatore, C.; Zuccotto, F.; Gilbert,
I. H.; De Clercq, E.; Balzarini, J. Design and Synthesis of
Lipophilic Phosphoramidate d4T-MP Prodrugs Expressing High
Potency Against HIV in Cell Culture: Structure Determinants
for in vitro Activity and QSAR. J. Med. Chem. 1999, 42, 4122-
8.
(10) Meier, C.; Knispel, T.; De Clercq, E.; Balzarini, J. CyscloSal-
Pronucleotides of 2′,3′-Dideoxyadenosin and 2′,3′-Dideoxy-2′,3′-
didehydroadenosine: Synthesis and Antiviral Evaluation of a
Highly Efficient Nucleotide Delivery System. J. Med. Chem.
1999, 42, 1604-1614.
(11) De Clercq, E. Antiviral Drugs: Current State of the Art. J. Clin.
Virol. 2001, 22, 73-89.
(12) Bayes, M.; Rabasseda, X.; Prous, J. R. Gateways to Clinical
Trials. Methods Find. Exp. Clin. Pharmacol. 2003, 25, 53-76.
(13) Selmi, B.; Deval, J.; Borreto, J.; Canard, B. Nucleotide Analogue
Binding, Catalysis and Primer Unblocking in the Mechanisms
of HIV Reverse Transcriptase-Mediated Resistance to Nucleoside
Analogues. Antiviral Therapy 2003, 8, 143-154.
(14) Blackburn, G. M.; Kent, D. E.; Kolkman, F. Three New â,γ-
Methylene Analogues of Adenosine Triphosphate. J. Chem. Soc.
Chem. Commun. 1981, 1188-1190.
(15) Khandazhinskaya, A. L.; Shirikova, E. A.; Skoblov, Y. S.;
Viktorova, L. S.; Goryunova, L. Y.; Beabealashvilli, R. S.;
Pronyaeva, T. R.; Fedyuk, N. V.; Zolin, V. V.; Pokrovsky, A. G.;
Kukhanova, M. K. Carbocyclic Dinucleotide Polyphosphonates:
Interaction with HIV Reverse Transcriptase and Antiviral
Activity. J. Med. Chem. 2002, 45, 1284-1291.
(16) Martynov, B. I.; Shirokova, E. A.; Jasko, M. V.; Victorova, L. S.;
Krayevsky, A. A. Effect of Triphosphate Modifications in 2′-
Deoxynucleoside 5′-triphosphates on Their Specificity towards
Various DNA Polymerases. FEBS Lett. 1997, 410, 423-427.
(17) Arabshahi, L.; Khan, N. N.; Butler, M.; Noonan, T.; Brown, N.
C.; Wright, G. E. (Difluoromethylene)phosphate Nucleosides as
Probes of DNA Polymerases and G Proteins. Biochemistry 1990,
29, 6820-6826.
(19) Oyelere, A. K.; Strobel, S. A. Biochemical Detection of Cytidine
Protonation within RNA. J. Am. Chem. Soc. 2000, 122, 10259-
10267.
(20) Arzumanov, A. A.; Semizarov, D. G.; Victorova, L. S.; Dyatkina,
N. B.; Krayevsky, A. A. γ-Phosphate-Substituted 2′-Deoxy-
nucleoside 5′-Triphosphates as Substrates for DNA Polymerases.
J. Biol. Chem. 1996, 271, 24389-24394.
(21) He, K.; Hasan, A.; Krzyzanowska, B.; Shaw, B. R. Synthesis and
Separation of Ribonucleoside 5′-(R-P-Borano)triphosphates. J.
Org. Chem. 1998, 63, 5769-5773.
(22) Lin, J.; Porter, K. W.; Shaw, B. R. Synthesis And Properties of
Novel Triphosphate Analogues: Ribonucleoside and Deoxyribo-
nucleoside (R-P-Borano, R-P-thio)triphosphates. Nucleosides Nu-
cleotides 2001, 20, 1019-1023.
(23) Trowbridge, D. B.; Yamamoto, D. M.; Kenyon, G. L. Ring
Openings of Trimetaphosphoric Acid and Its Bismethylene
Analogue. Synthesis of Adenosine 5′-Bis(dihydroxyphosphinyl-
methyl)phosphonate and 5′-Amino-5′-deoxyadenosine 5′-Tri-
phosphate. J. Am. Chem. Soc. 1972, 94, 3816-3824.
(24) Meyer, P.; Scheider, B.; Sarfati, S.; Deville-Bonne, D.; Guerreiro,
C.; Boretto, J.; Janin, J.; Veron, M.; Canard, B. Structural Basis
for Activation of R-Boranophosphate Nucleotide Analogues
Targeting Drug-Resistant Reverse Transcriptase. EMBO J.
2000, 19, 3520-3529.
(25) Nahum, V.; Zundorf, G.; Levesque, S. A.; Beaudoin, A. R.; Reiser,
G.; Fischer, B. Adenosine 5′-O-(1-Boranotriphosphate) Deriva-
tives as Novel P2Y1 Receptor Agonists. J. Med. Chem. 2002,
45, 5384-5396.
(26) Ma, Q.-F.; Kenyon, G. L.; Markham, G. D. Specificity of S-
Adenosylmethionine Synthetase for ATP Analogues Mono- and
Disubstituted in Bridging Positions of the Polyphosphate Chain.
Biochemistry 1990, 29, 1412-1416.
(27) Eckstein, E. Oligonucleotides and Analogues; IRL Press: Oxford,
1991; Chapter 1.
(28) Ludwig, J.; Eckstein, F. Synthesis of Nucleoside 5′-O-(1,3-
Dithiotriphosphates) and 5′-O-(1,1-Dithiotriphosphates). J. Org.
Chem. 1991, 56, 1777-1783.
(29) Arzumanov, A. A.; Dyatkina, N. B. An Alternative Route for
Preparation of R-Methylphosphonyl-â,γ-diphosphates of Thymi-
dine Derivatives. Nucleosides Nucleotides 1994, 13, 1031-1037.
(30) Lin, T. S.; Shen, Z. Y.; August, E. M.; Brankovan, V.; Yang, H.;
Ghazzouli, I.; Prusoff, W. H. Synthesis and Antiviral Activity of
Several 2,5′-Anhydro Analogues 3′-Azido-3′-deoxythymidine (AZT),
3′-Azido-2′,3′-dideoxyuridine (AZU), 3′-Azido-2′,3′-dideoxy-5-ha-
louridines, and 3′-Deoxythymidine Against Human Hmmuno-
deficiency Virus (HIV-1) and Rauscher-Murine Leukemia Virus
(R-MULV). J. Med. Chem. 1989, 32, 1891-1895.
(31) Seville, M.; West, A. B.; Cull, M. G.; McHenry, C. S. Fluorometric
Assay for DNA Polymerases and Reverse Transcriptase. Bio-
Techniques 1996, 21, 664-672.
(32) Shaw, B. R.; Dobrikov, M.; Wang, X.; Wan, J.; He, K.; Lin, J.-L.
Li, P.; Rait V.; Sergueeva, Z. A.; Sergueev, D. Reading, Writing
and Modulating Genetic Information with Boranophosphate
Mimics of Nucleotides, DNA, RNA. Ann. N. Y. Acad. Sci. 2003,
1002, 12-29.
(33) Selmi, B.; Boretto, J.; Sarfati, S. R.; Guerreio, C.; Canard, B.
Mechanism-based Suppression of Dideoxynucleotide Resistance
by K65R Human Immunodeficiency Virus Reverse Transcriptase
Using an R-Boranophosphate Nucleotide Analogue. J. Biol.
Chem. 2001, 276, 48488-48472.
(34) Deva, J.; Selmi, B.; Boretto, J.; Egloff, M. P.; Guerreio, C.; Sarfati,
S. Canard, B. The Molecular Mechanism of Multidrug Resistance
by the Q151M Human Immunodeficiency Virus Type 1 Reverse
Transcriptase and its Suppression Using R-Boranophosphate
Nucleotide Analogues. J. Biol. Chem. 2002, 277, 42097-42104.
(35) Dyatkina, N.; Arzumanov, A.; Victorova, L.; Kukhanova, M.;
Krayevsky, A. Synthesis and Substrate Properties of Thymidine
5′-Triphosphate Analogues with Large Hydrophobic Substituent
Groups at R-P Atom. Nucleosides Nucelotides 1995, 14, 91-103.
(18) Ma, Q.-F.; Bathurst, I. C.; Barr, P. J.; Kenyon, G. L. New
Thymidine Triphosphate Analogue Inhibitors of Human Immu-
nodefficiency Virus-1 Reverse Transcriptase. J. Med. Chem.
1992, 35, 1938-1941.
JM040116W