cycloSal-pronucleotides of 2',3'-dideoxyadenosine and 2',3'-dideoxy- 2',3'-didehydroadenosine: Synthesis and antiviral evaluation of a highly efficient nucleotide delivery system

Article abstract of DOI:10.1021/jm981096z

The synthesis, hydrolysis, and antiviral evaluation of novel, lipophilic cycloSal-ddAMP (9a-d) and cycloSal-d4AMP (10a-d) derivatives of the antiviral purine dideoxynucleoside analogues 2',3'-dideoxyadenosine (ddA) (2) and 2',3'-dideoxy-2',3'-didehydroadenosine (d4A) (3) are reported. These potential pronucleotides release ddAMP (7) or d4AMP (8) selectively by a controlled, chemically induced tandem reaction. All new compounds 9 and 10a- d were synthesized in good yields using our previously reported phosphorus(III) method starting from substituted salicyl alcohols 14a-h. The phosphotriesters 9 and 10 were obtained with a stereochemical preference of 2:1 with respect to the configuration at the phosphorus center. In an 1- octanol/water mixture phosphotriesters 9 and 10 exhibited 7-43-fold higher lipophilicity than the parent nucleosides ddA (2) and d4A (3) as judged by their log P values. In hydrolysis studies, 9 and 10 decomposed under mild aqueous basic conditions releasing solely ddAMP (7) and d4AMP (8), as well as the diols 14. Further hydrolysis studies under acidic conditions showed a marked increase in stability with respect to the acid-catalyzed cleavage of the glycosyl bond. Phosphotriesters 9 and 10 exhibited antiviral potencies against wild-type HIV-1 and HIV-2 strains in human T-lymphocyte (CEM/O) cells that were, respectively, 100- and 600-fold higher than those of ddA (2) and d4A (3). Furthermore, all triesters 9 and 10 were markedly more active than the corresponding ddI compounds 11 and 12, which supports the concept of the delivery of the adenine nucleotides. Studies with adenosine deaminase (ADA) and adenosine monophosphate deaminase (AMPDA) showed that the triesters were not substrates for enzymatic deamination. The studies reported herein demonstrate conclusively that the cycloSal triesters deliver exclusively the nucleotides ddAMP and d4AMP, not only under chemical-simulated hydrolysis but also under intracellular conditions fulfilling the adenosine deaminase bypass premise.

Full text of DOI:10.1021/jm981096z

1604
J . Med. Chem. 1999, 42, 1604-1614
cycloSa l-P r on u cleotid es of 2′,3′-Did eoxya d en osin e a n d
2′,3′-Did eoxy-2′,3′-d id eh yd r oa d en osin e: Syn th esis a n d An tivir a l Eva lu a tion of a
High ly Efficien t Nu cleotid e Deliver y System
Chris Meier,*,†,‡ Tina Knispel,†,‡ Erik De Clercq,§ and J an Balzarini§
Institute of Organic Chemistry, J ulius-Maximilians-University Wu¨rzburg, Am Hubland, D-97074 Wu¨rzburg, Germany, and
Rega Institute for Medical Research, Katholieke Universiteit Leuven, B-3000 Leuven, Belgium
Received October 13, 1998
The synthesis, hydrolysis, and antiviral evaluation of novel, lipophilic cycloSal-ddAMP (9a -
d ) and cycloSal-d4AMP (10a -d ) derivatives of the antiviral purine dideoxynucleoside analogues
2′,3′-dideoxyadenosine (ddA) (2) and 2′,3′-dideoxy-2′,3′-didehydroadenosine (d4A) (3) are
reported. These potential pronucleotides release ddAMP (7) or d4AMP (8) selectively by a
controlled, chemically induced tandem reaction. All new compounds 9 and 10a -d were
synthesized in good yields using our previously reported phosphorus(III) method starting from
substituted salicyl alcohols 14a -h . The phosphotriesters 9 and 10 were obtained with a
stereochemical preference of 2:1 with respect to the configuration at the phosphorus center. In
an 1-octanol/water mixture phosphotriesters 9 and 10 exhibited 7-43-fold higher lipophilicity
than the parent nucleosides ddA (2) and d4A (3) as judged by their log P values. In hydrolysis
studies, 9 and 10 decomposed under mild aqueous basic conditions releasing solely ddAMP (7)
and d4AMP (8), as well as the diols 14. Further hydrolysis studies under acidic conditions
showed a marked increase in stability with respect to the acid-catalyzed cleavage of the glycosyl
bond. Phosphotriesters 9 and 10 exhibited antiviral potencies against wild-type HIV-1 and
HIV-2 strains in human T-lymphocyte (CEM/O) cells that were, respectively, 100- and 600-
fold higher than those of ddA (2) and d4A (3). Furthermore, all triesters 9 and 10 were markedly
more active than the corresponding ddI compounds 11 and 12, which supports the concept of
the delivery of the adenine nucleotides. Studies with adenosine deaminase (ADA) and adenosine
monophosphate deaminase (AMPDA) showed that the triesters were not substrates for
enzymatic deamination. The studies reported herein demonstrate conclusively that the cycloSal
triesters deliver exclusively the nucleotides ddAMP and d4AMP, not only under chemical-
simulated hydrolysis but also under intracellular conditions fulfilling the adenosine deaminase
bypass premise.
In tr od u ction
RT (Ki ) 0.22 µM) to that of AZTTP (Ki ) 0.1 µM).7
Furthermore, in marked contrast with the pyrimidine
2′,3′-dideoxynucleosides (i.e., d4T, AZT), ddA (2) is
extensively catabolized.3,4,8-10 The main catabolic path-
way is deamination into ddI (1) by the ubiquitous
cellular enzyme adenosine deaminase (ADA).11 The
inosine derivative 1 is either inactivated by cleavage to
hypoxanthine (Hx) by purine nucleoside phosphorylase
(PNP) or further phosphorylated to the monophosphate
ddIMP (5) by 5′-nucleotidase,12 a cellular enzyme whose
activity remains nearly constant during all phases of
the cell cycle.13,14 No further phosphorylation of 5 to the
corresponding triphosphate has been observed. There-
fore, ddIMP (5) enters the adenosine anabolic pathway
through the action of adenylosuccinate synthetase and
adenylosuccinate lyase to give 2′,3′-dideoxyadenosine
monophosphate (ddAMP, 7).11,15 However, the enzy-
matic conversion of 5 into 7 is not very efficient, and
both intervening enzymes become rate-limiting steps.
Further phosphorylation by cellular kinases converts
ddAMP to the triphosphate level (Scheme 1).3,10 It can
be assumed that d4A (3) may follow a similar pathway
via d4I (4) and d4IMP (6) to 2′,3′-dideoxy-2′,3′-didehy-
droadenosine monophosphate (d4AMP, 8). However,
this has never been demonstrated.
Nucleoside analogues, e.g., 3′-azido-2′,3′-dideoxythy-
midine (AZT) or 2′,3′-dideoxy-2′,3′-didehydrothymidine
(d4T), are used as antiviral agents in the treatment for
AIDS.1,2 The only purine nucleoside analogue clinically
approved to date is 2′,3′-dideoxyinosine (ddI, 1; Didano-
sine, Videx), the deamination product of 2′,3′-dideoxy-
adenosine (ddA, 2).3,4 Like all nucleoside analogues that
are active against human immunodeficiency virus (HIV),
ddA (2) as well as the unsaturated 2′,3′-dideoxy-2′,3′-
didehydroadenosine (d4A, 3) need to be converted to
their corresponding 5′-triphosphates intracellularly be-
fore they can express their biological activity as inhibi-
tors of the target enzyme (reverse transcriptase, RT).5
The first phosphorylation step is predominantly cata-
lyzed by nucleoside kinases and/or 5′-nucleotidase, but
ddA (2) and d4A (3) have very poor affinity for their
activating enzymes.5,6 Indeed, in the case of ddA (2),
only very low intracellular levels of ddATP are detected.
This metabolite shows a similar inhibitory effect of HIV-
* To whom correspondence should be addressed.
† J ulius-Maximilians-University of Wu¨rzburg.
§ Katholieke Universiteit Leuven.
‡ New address: Institute of Organic Chemistry, University of
Hamburg, Martin-Luther-King-Platz 6, D-20146 Hamburg, Germany.
10.1021/jm981096z CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/09/1999

cycloSal-NMP of ddA and d4A
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 9 1605
Sch em e 1. Metabolism of the Purine Nucleosides ddA
and d4A
Ch a r t 1. Structures of ddI (1), ddA (2), d4A (3), d4I (4),
Their Corresponding Monophosphates 5-8, and Title
cycloSal Phosphotriesters 9-12
Consequently, the direct administration of ddAMP (7)
and d4AMP (8) could in principle bypass the metabolic
limiting processes of ADA deamination (ADA bypass)
and thus improve the therapeutic potential of these
drugs. However, a well-known problem in antiviral
therapy is the poor bioavailability of highly polar
nucleotides due to low membrane penetration. There-
fore, intracellular delivery of nucleotides, i.e., ddAMP
(7) and d4AMP (8), from neutral, lipophilic pronucleo-
tides16-18 could bypass the rate-limiting ADA f 5′-
nucleotidase route to improve biological activity (pro-
nucleotide approach, Scheme 1). Several types of masked
nucleoside 5′-monophosphate prodrugs containing two
phosphate blocking groups have been synthesized to
achieve intracellular nucleotide delivery.19 The major
differences of these approaches are the nature of the
delivery mechanisms. An additional drawback in the use
of ddA (2), and especially d4A (3), is their high acid
instability which complicates oral administration of the
drugs. In one report, it was shown that the introduction
of an electron-withdrawing phosphate ester group at the
5′-position of ddA had a stabilizing effect on the glycosyl
bond.17
A preliminary account on the ddA derivatives was
published recently.23
As part of our ongoing program to develop efficient
pronucleotide delivery systems, we were interested in
a completely new nucleotide delivery approach based
on pH-driven selective chemical hydrolysis:20 the cy-
closaligenyl-pronucleotides (cycloSal-NMP).21 This con-
cept has been successfully introduced with other anti-
HIV nucleosides such as d4T. Indeed, the cycloSal-
d4TMP triesters proved to be successful thymidine
kinase bypass agents.22
Here we report the synthesis, lipophilic and hydrolytic
properties, and biological activity of cycloSal-2′,3′-
dideoxyadenosine monophosphates (9, cycloSal-ddAMP)
and cycloSal-2′,3′-dideoxy-2′,3′-didehydroadenosine mono-
phosphates (10, cycloSal-d4AMP) which we propose
function as neutral pronucleotides of ddAMP (7) and
d4AMP (8), as well as efficient ADA bypass agents
(Chart 1). Furthermore, the effect of the cycloSal
modification on the acid stability of the pronucleotide
as well as the lack of enzymatic deamination by ad-
enosine deaminase (ADA) and adenosine monophos-
phate deaminase (AMPDA) are reported. The data
obtained for 9 and 10 were compared to those of the
corresponding ddIMP (11c) and d4IMP (12c) triesters.
Resu lts a n d Discu ssion
Ch em istr y. ddA (2) was synthesized from 2′-deoxy-
adenosine by (1) regioselective 5′-O-benzoylation with
benzoic acid catalyzed by N,N′-bis(2-oxooxazolidin-3-yl)-
phosphorodiamidic chloride (BOPDC) in pyridine24 (65%
yield), (2) regioselective 3′-O-esterification with thio-
carbonyl diimidazole and subsequent reaction with
methanol (77% yield; note: esterification using phenoxy-
thiocarbonyl chloride was not sufficiently regioselective),
and (3) Barton deoxygenation to 5′-benzoyl-2′,3′-dideoxy-
adenosine (88% yield), followed by a 12-h treatment with
ammonia in methanol (30% overall yield).23 d4A (3) was
prepared from adenosine via 3′-bromo-3′-deoxy-2′,5′-di-
O-acetyladenosine by a reductive elimination with
freshly prepared Zn/Cu couple using a combination of
procedures published by Walker et al.25 Deprotection
of the 5′-position with cold saturated methanolic am-
monia yielded d4A (3) in 35% overall yield.
For comparative studies, vide infra, we also prepared
in nearly quantitative yields ddI (1) and d4I (4) by
enzymatic deamination of ddA (2) and d4A (3) using

1606 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 9
Meier et al.
Sch em e 2. Synthetic Pathways to the Title cycloSal Triesters 9-12a
a
Reaction conditions: (a) PCl3, pyridine, Et2O, -10 °C, 2 h; (b) ddA (2), DIPEA, DMF/THF (2:1), -40 °C, 30 min; (c) TBHP, DMF/THF
(2:1), -40 °C-rt, 1 h; (d) d4A (3), DIPEA, DMF/THF (2:1), -40 °C, 30 min; (e) from 13c: ddI (1), DIPEA, DMF/THF (2:1), -40 °C, 30 min;
(f) from 13c: d4I (4), DIPEA, DMF/THF (2:1), -40 °C, 30 min.
adenosine deaminase (ADA).26 All nucleoside derivatives
Sch em e 3. Chemical Hydrolysis Pathways of cycloSal
1-4 synthesized here showed spectroscopic data identi-
Phosphotriesters Shown for the cycloSal-ddAMP 9
cal to those reported before.27,28
The synthesis of the cycloSal triesters 9 and 10 was
carried out without protection of the exocyclic amino
group of the adenine nucleosides 2 and 3. The commonly
used base-labile protecting groups had to be avoided due
to the potential instability of the target triester deriva-
tives 9 and 10, and acid-labile protecting groups could
have resulted in the acid-catalyzed cleavage of the
glycosyl bond of ddA and d4A. The cycloSal residue was
introduced using our previously reported reactive chlo-
rophosphanes 13, which were prepared from the ap-
propriate salicyl alcohols 14.29 The only differences from
our previously reported reaction sequence were the
solvent used and the reaction temperature. Due to low
solubility of 2 or 3 in acetonitrile, a solvent mixture of
dimethylformamide and tetrahydrofuran (DMF/THF,
2:1) was used. The regioselective 5′-O-phosphorylation
reaction of 2 and 3 was achieved at -40 °C to enhance
the O- vs N-selectivity in the presence of the unprotected
amino group. Following this procedure, the cycloSal-
ddAMP (9a -d ) and cycloSal-d4AMP (10a -d ) triesters
were obtained in good yields and with a regioselectivity
of 8:1 in favor of the 5′-O-modification. The reaction
sequence is outlined in Scheme 2.
could be separated by analytical RP silica gel columns
(C18), but in a few cases the two peaks overlapped
partially or completely. For the biological evaluation,
small amounts of the cycloSal phosphotriesters 9a -d
and 10a -d were additionally purified by semiprepara-
tive HPLC using acetonitrile/water eluents in order to
eliminate any traces of nucleoside impurities. Only in
the case of the 5-methoxy-substituted cycloSal-d4AMP
10b was the separation of the diastereomers possible
by means of semipreparative HPLC. After lyophiliza-
tion, all phosphotriesters 9 and 10 were obtained as
white fluffy solids. To compare the properties and
antiviral potencies of the cycloSal-ddAMP and cycloSal-
d4AMP triesters with the corresponding inosine deriva-
tives, the 3-methyl-cycloSal-nucleotides of ddI and d4I
(11c and 12c) were prepared by similar methods.
Deter m in a tion of th e P a r tition Coefficien ts (log
P Va lu es). The partition coefficients (log P values) of
the cycloSal-NMPs 9-12, as well as those of the parent
nucleoside analogues 1-4, were determined in 1-oc-
tanol/water by our previously reported HPLC
method.22a,29 The P values of all the phosphotriester
derivatives were >1, and all log P values are positive.
cycloSal-ddAMP phosphotriesters 9 exhibited P values
that were 7.5-43-fold higher than those of the parent
nucleoside ddA (2) (log P ) -0.28), and cycloSal-
In contrast to our previous results with cycloSal-
d4TMP, where 1:1 diastereomeric mixtures were ob-
tained, the title compounds 9 and 10 were isolated as
1:1.6 diastereomeric mixtures with respect to the con-
figuration at the phosphorus center after purification
on a chromatotron. Most probably, this uneven ratio is
due to kinetic control during the low-temperature reac-
tions. The structure of the diastereomers was character-
1
ized by extensive H, 13C, and 31P NMR spectroscopy,
as well as by electrospray mass spectrometry (ESI,
positive mode) and UV spectroscopy. As expected, the
phosphotriesters displayed two closely spaced signals
in the 31P NMR spectra. The purity was checked by
analytical reversed-phase high-performance liquid chro-
matography (RP-HPLC) using three distinct eluant
gradients (water/acetonitrile gradients). In almost all
cases, the two expected peaks of the diastereoisomers

cycloSal-NMP of ddA and d4A
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 9 1607
Ta ble 1. Hydrolysis Half-Lives (t1/2) in Different Aqueous
Buffers and log P Values of the Nucleosides 1-4 and the
cycloSal-NMPs 9-12
d4AMPs 10 were 7-28-fold more lipophilic than d4A
(3) (log P -0.49). As expected, 3-methyl-cycloSal-ddAMP
(9c) was 5-fold more lipophilic than the corresponding
inosinate derivative 11c. Compared to AZT (log P of
0.025),29-31 which enters mammalian cells by passive,
nonfacilitated diffusion,32 the new ddA- and d4A-pro-
nucleotides showed 3.6-21-fold (cycloSal-ddAMPs 9)
and 2.1-8.4-fold (cycloSal-d4AMPs 10) increases in
lipophilicity. According to the higher log P values, an
improved cellular uptake by passive diffusion was
anticipated.
hydrolysis (t1/2, h)b
in aqueous buffers at 37 °C
compd subst X pH 7.3a pH 3.0c pH 1.5d
log P
e
9a
9b
9c
9d
2
11c
1
10a
10b
10c
10d
3
H
17.0
23.4
44.3
58.2
12.9
12.2
13.9
10.3
1.28
3.0
0.60
0.80
0.53
0.73
0.03
0.20
0.03
0.59
0.64
0.88
5-OMe
3-Me
3,5-diMe
1.35
f
na
-0.2831
0.14
3-Me
44.5
f
na
0.2
-1.2531
0.36
Kin etic Stu d ies. The cycloSal-pronucleotide concept
has been designed to release selectively the nucleotides
and the masking group by a controlled, chemically
induced tandem reaction. In contrast to other prodrug
concepts based on enzymatic hydrolysis,19 our approach
involves the successive coupled cleavage of the phenyl
and benzyl esters of the cycloSal phosphotriester. The
degradation pathway has been proven by different
methodologies and has been applied to different nucleo-
sides, e.g., d4T22 and ACV.33 In both cases, the intra-
cellular delivery of the corresponding nucleotides has
been clearly demonstrated by the observed biological
activity even in the case of thymidine kinase-deficient
cell lines or virus strains. Surprisingly, the approach
failed to deliver AZTMP in an in vitro test system.29
Here, we applied the concept to adenosine analogues 2
and 3 with the intent to bypass adenosine deamination
by ADA and thereby avoid the rapid conversion of
adenine nucleosides into the corresponding hypoxan-
thine derivatives.
H
10.3
14.5
32.3
44.7
0.14
0.14
0.15
0.14
0.01
0.04
,0.01
0.010
0.011
0.0098
0.010
,0.01
,0.01
,0.01
5-OMe
3-Me
3,5-diMe
0.48(f)/0.41(s)
0.63
0.95
-0.49
0.078
f
na
12c
4
AZT
3-Me
29.7
f
na
na
-0.89
0.02531
f
f
f
na
na
a
b
pH 7.3: 25 mM phosphate buffer (So¨rensen). Half-lives were
determined from the decreasing peak of the starting phosphotri-
ester and are the mean of duplicate experiments; values are given
in hours (h). c pH 3.0: 25 mM glycine/HCl buffer. pH 1.5: 25
d
mM glycine/HCl buffer. e log P: log of the partition coefficient.f na,
not available.
commensurate with the electron-donating ability of the
substituent. However, at this pH value the half-lives
were considerably longer than previously detected for
the pyrimidine nucleoside cycloSal derivatives22a,29 (e.g.,
about 3.8-fold (9) and 2.5-fold (10) longer than those of
the corresponding d4TMP triesters) despite the common
cycloSal motif. We have no explanation for this marked
increase in stability. It was also apparent that the
d4AMP derivatives 10 were less stable than the ddAMP
derivatives 9. However, the nucleotides ddAMP (7) and
d4AMP (8), plus the salicylic alcohol, were the only
products detected at 260 nm. Further hydrolysis studies
of 9 and 10 were performed in RPMI culture medium
with and without 10% heat-inactivated FCS. Under
these conditions, a slight decrease in half-lives was
observed. This is most probably due to the more basic
pH of the medium used (e.g., 11.4 h (9a ), 15.4 h (9b),
and 33.8 h (9c); data not shown). As before, only the
monophosphates were observed as hydrolysis products.
A very important disadvantage of purine 2′,3′-
dideoxynucleosides is their high instability against acid-
catalyzed cleavage of the glycosyl bond.34 To investigate
this issue, the ddAMP (9a -d ) and d4AMP (10a -d )
phosphotriesters, as well as the corresponding ddIMP
(11c) and d4IMP (12c) derivatives, were incubated in
25 mM glycine/HCl buffer (pH 3.0 and 1.5), and their
half-lives were compared to those of the parent nucleo-
side analogues, ddA, d4A, and ddI. Within each series,
9a -d or 10a -d , the t1/2 values remained constant
around 13 and 0.15 h, respectively, at pH 3.0 (Table 1).
This suggests that the proton-induced cleavage of the
glycosyl bond is independent of the substitution pattern
in the cycloSal residue.34 Furthermore, the 86-fold
increase in lability of the d4AMP triesters 10 compared
to triesters 9 reflect the higher stability of the allylic
carbocation of the latter, which is formed after cleavage
of the glycosyl bond. However, as can be concluded from
the half-lives at pH 3.0 and 1.5, the introduction of the
cycloSal moiety at the 5′-position of the dideoxyribose
residue stabilized the glycosyl bond considerably (10-
The rates of hydrolysis of the cycloSal-ddAMP and
-d4AMP derivatives 9 and 10 were measured under a
variety of conditions using different aqueous buffers and
pH values. Hydrolyses rates in 25 mM phosphate buffer,
pH 7.3, allowed us to compare the effect of the purine
base on the cycloSal motif relative to the pyrimidine
nucleotide analogues described before.22a,29 On the other
hand, the effect of 5′-phosphorylation on the stability
of the glycosyl bond and the acid stability of the cycloSal
moiety itself were determined in a glycine/HCl buffer
at pH values of 3.0 and 1.5, respectively. These studies
were conducted with all cycloSal phosphotriesters 9 and
10 and showed a clear correlation between electronic
properties and hydrolytic stability of the triesters. From
our previous work, we selected the 3,5-dimethyl, the
3-methyl, and the 5-methoxy, as well as the unsubsti-
tuted cycloSal phosphotriesters, because these groups
and substitution patterns gave the best biological
results. The overall behavior of the cycloSal phospho-
triester in terms of chemical hydrolysis correlated well
with successful membrane penetration. The 3-methyl-
cycloSal derivatives of ddIMP (11c) and d4IMP (12c)
were also included in the study. The degradation of the
cycloSal-ddAMPs 9 and cycloSal-d4AMPs 10 was fol-
lowed by RP-HPLC analysis. The half-lives were deter-
mined by integration of the decreasing peaks of the
triesters versus time. The half-lives are summarized in
Table 1.
As in the case of the cycloSal-d4TMP compounds, a
clear correlation between hydrolytic stability and the
substituents on the aromatic ring was observed. In
phosphate buffer (pH 7.3), the half-lives increased

1608 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 9
Meier et al.
15-fold) with respect to ddA (2) and d4A (3) (Table 1).
A 10-fold increase in stability has also been observed
for the bisSATE-ddAMP triesters at pH 2.0.17 Table 1
also shows that all 2′,3′-dideoxyinosine derivatives are
considerably less stable when compared to their 2′,3′-
dideoxyadenosine counterparts. The higher lability of
ddI relative to ddA has already been observed before;
however, the cycloSal-dd4IMP triesters are 15-fold more
stable than the parent nucleosides. As expected, a pH
dependence was also observed. At pH 1.5, the rate of
cleavage of the glycosyl bond of the cycloSal phospho-
triesters 9 was ca. 23-fold higher than at pH 3.0. This
was less than the 43-fold increase observed for ddA (2).
The corresponding reaction rates of the d4AMP triesters
10 were too fast to be determined under the experimen-
tal conditions used. The decrease of the triester peak
in the HPLC chromatograms was caused nearly exclu-
sively by glycosyl bond cleavage and not by chemical
hydrolysis of the phosphate esters. This was also veri-
fied for the cycloSal derivative of F-ara-ddA, an acid-
stable analogue of ddA, which under identical reaction
conditions at pH 3 showed a t1/2 of 284 h (compared to
13 h for 9) and at pH 1.5 a t1/2 of 19 h (compared to 0.54
h for 9).
The stability experiments using rabbit muscle AMP-
DA (EC 3.5.4.6) were carried out with 2′-deoxyadenosine
monophosphate (dAMP) and 2′,3′-dideoxyadenosine
monophosphate (ddAMP) as reference compounds in the
presence of 0.5 and 1 U of AMPDA, respectively, in 10
mM phosphate buffer (pH 6.8). The 3-methyl-cycloSal-
ddAMP (9) and -d4AMP (10) were incubated with 2 U
of AMPDA under the same conditions. As before, the
enzymatic reactions were followed by UV spectroscopy
(260 nm) at 37 °C. Under these conditions, dAMP and
ddAMP were completely converted to dIMP and ddIMP
within 60 min. On the other hand, 3-methyl-cycloSal-
ddAMP (9c) and -d4AMP (10c) were completely stable
against AMPDA deamination for 200 min, even at twice
the concentration of enzyme. The presence of the intact
6-amino group in the cycloSal compounds after the
incubation time was demonstrated by the characteristic
UV absorption maximum at 259 nm. These experiments
showed that the cycloSal derivatives described here are
completely stable against both ADA and AMPDA and
should therefore deliver only the adenine nucleotides
ddAMP and d4AMP intracellularly, and not ddIMP or
d4IMP.
An tivir a l Eva lu a tion . The successful thymidine
kinase bypass by cycloSal-d4TMP22 and cycloSal-
ACVMP phosphotriesters33 has been shown before.
These results demonstrated (1) a pronounced structure-
bioactivity correlation with respect to the substituents
on the cycloSal moiety, (2) the ready membrane pen-
etration of the cycloSal-d4TMP triesters, (3) the efficient
intracellular delivery of d4TMP, and (4) the complete
independence from TK.22 In sharp contrast, the corre-
sponding cycloSal-AZTMP derivatives lost nearly all the
antiviral activity observed in wild-type CEM/O cells29
when tested in TK-deficient CEM cells. The parent
nucleosides ddA (2) and d4A (3), as well as the corre-
sponding cycloSal compounds 9 and 10, were evaluated
for their ability to inhibit the replication of HIV-1 and
HIV-2 in rapidly dividing human T-lymphocytic cells
(CEM cells). The compounds were also tested against
Moloney murine sarcoma virus (MSV)-induced trans-
formation of C3H/3T3 embryo murine fibroblasts. Fur-
thermore, to demonstrate the efficiency of the ADA
bypass, 3-methyl-cycloSal-ddIMP (11c) and 3-methyl-
cycloSal-d4IMP (12c) were included in the antiviral
tests since these compounds can only deliver ddIMP and
d4IMP, respectively. To get to bioactive nucleoside
triphosphates from ddIMP and d4IMP, two enzyme-
catalyzed conversions into ddAMP and d4AMP have to
take place prior to further metabolism to the corre-
sponding triphosphate level (Scheme 1). It is generally
accepted that these conversions are not very efficient.15
The test compounds 9a -d and 10a -d were free of the
parent nucleosides 2 and 3, as verified by analytical RP-
HPLC (>99% purity). The antiviral activities of the
clinically used nucleoside analogues, d4T and AZT, are
given for comparison. The results obtained are displayed
in Table 2.
En zym a tic Stu d ies. As already discussed, ddA and
d4A are highly susceptible to deamination by ADA
which produces inactive inosine derivatives. It was
assumed that this enzymatic deamination was respon-
sible for a considerable loss in biological potency because
the required reamination at the ddIMP and d4IMP
levels becomes rate-limiting. Since our goal was to
bypass the deamination by ADA by 5′-modification of
ddA and d4A, while at the same time developing a
nucleotide delivery system, it was necessary to study
the stability of the cycloSal-ddAMPs 9 and cycloSal-
d4AMPs 10 against ADA deamination. In addition, we
studied the enzymatic stability of 9 and 10 against
adenosine monophosphate deaminase (AMPDA) because
this enzyme is normally responsible for the deamination
of AMP.15 The reason for also testing the AMPDA
stability was to exclude the possibility that the cycloSal
triesters are deaminated to give the corresponding
cycloSal-dd(4)IMP counterparts. The concentration of
ddA, d4A, and the respective 3-methyl-substituted cy-
cloSal phosphotriesters was set at 185 µM, and the
deamination reactions were followed by UV spectroscopy
at 260 nm. The nucleosides ddA (2) and d4A (3) were
completely deaminated to the inosine analogues within
2-10 min in the presence of 0.25 U of ADA (EC 3.5.4.4;
from calf intestinal mucosa) in 10 mM phosphate buffer
(pH 6.8) at 37 °C. The close to neutral pH of the
phosphate buffer was chosen to suppress deglycosylation
during the incubation time. As expected,35 all 5′-modi-
fied cycloSal phosphotriesters were completely stable
against ADA deamination after 90 min, even in the
presence of a 2-fold higher concentration of the enzyme
(0.5 U). In one case, we prolonged the incubation time
to 12 h without detecting any deamination reaction.
Both cycloSal-ddAMP and cycloSal-d4AMP triesters
remained unchanged after the incubation time as shown
by their characteristic UV spectra. From this observa-
tion we can clearly exclude the formation of cycloSal-
ddIs and cycloSal-d4Is through deamination of their
adenosine counterparts.
Both parent nucleoside analogues ddA (2) and d4A
(3) proved effective against HIV-1- and HIV-2-induced
cytopathogenicity in CEM cells with a 10-fold difference
in antiviral potency (2, EC50 4.3 and 4.5 µM; 3, 30 and
50 µM). In contrast, the cycloSal-ddAMP triesters 9
exhibited a 90-170-fold increase in antiviral potency

cycloSal-NMP of ddA and d4A
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 9 1609
Ta ble 2. Antiviral Activity and Selectivity Indices of cycloSal Phosphotriesters 9a -d , 10a -d , 11c, and 12c, as Well as the
Nuleosides 1-4, on the Replication of HIV-1 and HIV-2 in Wild-Type CEM/O Cells and against Transformation Induced in
MSV-Infected C3H/3T3 Cells
antiviral activity EC50 (µM)a
cytotoxicity
CEM
HIV-1
CEM
HIV-2
C3H/3T3
MSV
CEM
C3H/3T3
SIe
(CEM)
compd
9a
9b
9c
9d
2
11c
1
10a
10b (mix)
10b (f)
10b (s)
10c
10d
3
12c
4
d4T
AZT
subst X
CC50 (µM)b
MCC (µM)c
H
0.03
0.047
0.047
0.025
4.3
0.13
0.16
0.03
0.025
4.5
2.86
93
87
28
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
ndd
3096
1851
585
383
>57
53
5-OMe
3-Me
3,5-diMe
3.90
1.18
1.14
10
40
>50
>50
0.30
1.21
2.83
12.8
1.24
1.14
>250
3-Me
1.1
3.5
0.7
4.0
58
156
86
77
40
62
49
25
96
94
44
H
0.065
0.08
0.043
0.50
0.065
0.05
30
5.3
4.3
0.18
0.007
0.35
0.065
0.050
0.33
0.19
0.055
50
1329
962
932
125
746
494
3
5-OMe
5-OMe
5-OMe
3-Me
3,5-diMe
35
3-Me
5.0
>50
>50
18
20
311
>14000
>50
87
56
>100
0.26
0.006
15.0
>100
ndd
a
b
d
50% effective concentration. 50% cytotoxic concentration. c Minimal cytotoxic concentration. nd, not determined. e Selectivity index.
f (mix), (f), and (s) refer to diastereomeric mixture and respectively the resolved “fast-eluting” and “slow-eluting” diastereomers.
against HIV-1-induced cytopathogenicity. A similar
increase in potency was also observed against HIV-2-
induced cytopathogenicity. Furthermore, the selectivity
indices (SI) of the cycloSal phosphotriesters 9 increased
by 50-fold. As expected, due to the fast enzymatic
deamination of ddA into ddI, ddI was equally antivirally
potent and showed a similar SI value as ddA (2) in the
same assay. Interestingly, 3-methyl-cycloSal-ddIMP
(11c) was 4 times more potent than ddI. This enhance-
ment in bioactivity could be attributed to an increased
intracellular delivery of ddIMP (5). The significantly
lower potency against HIV-1 of 3-methyl-cycloSal-
ddIMP (11c) compared to 3-methyl-cycloSal-ddAMP (9c)
(EC50: 1.1 and 0.047 µM, respectively) clearly shows
that in the former case the delivery of ddIMP (5) is not
sufficient for optimal antiviral activity, while in the
latter case the delivery of ddAMP (7) is responsible for
the strong increase in antiviral potency. This effect was
also reflected in the SI values which showed that for
the ddIMP compound 11c it was about 50, whereas for
the ddAMP triester 9c it was about 650. This 13-fold
increase in SI is likely the result of a different intra-
cellular metabolic profile for each compound, since the
basic structure and delivery pathways of 9 and 10 are
identical. The lowest cytotoxic concentration observed
for cycloSal-ddAMP 9a was 86.4 µM resulting in the
highest SI of all the cycloSal compounds tested (SI
3096).
Although ddA (2) was 10 times more potent than d4A
(3), the cycloSal-d4AMP triesters 10 exhibited compa-
rable potencies when compared to the cycloSal-ddAMP
derivatives 9. This represents a 460-600-fold increase
in antiviral potency relative to d4A. At the same time,
the SI values increased from 3, for the parent nucleoside
analogue d4A (3), to 500, 750, and 1300 for the cycloSal
derivatives 10a ,c,d , respectively. Previously, we have
observed that biological activity was dependent on the
configuration at the asymmetric phosphorus atom.22a In
this study we were able to separate the 5-methoxy-
cycloSal-d4AMP diastereoisomers 10b. As seen in Table
2, both diastereomers exhibited different antiviral activ-
ity with the “fast”-eluting isomer being 11- and 6-fold
more potent than the “slow”-eluting diastereomer. Es-
sentially the same difference has been observed with
the cycloSal-d4TMP triesters before.22a However, the
chemical hydrolysis of those compounds showed only a
2-fold difference in the half-lives. The antiviral potency
and the SI of the diastereomeric mixture were inter-
mediate between the values of the diastereomerically
pure compounds.
The antiviral activity of 3-methyl-cycloSal-d4IMP
(12c) showed almost exactly the same antiviral potency
as the parent nucleoside d4I (4) and was about 6 times
more potent than d4A (3). Surprisingly, d4I (4) was also
7-fold more potent than d4A (3) (Table 2). Again, the
results show that even the intracellular delivery of
d4IMP (6) does not produce a comparable biological
effect to that obtained with cycloSal-d4AMP 10.
An interesting difference from our previous findings
with the cycloSal-d4TMP triesters is the virtual inde-
pendence of the biological activity of the cycloSal
compounds studied here relative to the substitution
pattern of the aromatic ring of the cycloSal moiety.
Earlier, we noted a clear correlation between antiviral
activity and the cycloSal substituents and, furthermore,
with the hydrolysis rates which were influenced by the
electronic properties of the substituents. Although a
similar correlation between hydrolysis rates and the
substitution pattern of the prodrugs was observed in the
present study using (1) phosphate buffer (pH 7.3), (2)
RPMI culture medium with 10% heat-inactivated FCS,
and (3) glycine/HCl buffer, the antiviral activity was the
same, within the experimental errors of the test system,
for all prodrugs 9 and 10. For the first time, these
findings seem to support a different activation pathway
for the ddA and d4A phosphotriesters when compared
to the corresponding d4T derivatives. Although the
cycloSal concept was based on a chemical hydrolysis
process to give the monophosphates (which has been
proven above), one possible explanation can be a con-
current enzyme-catalyzed initial step which operates
before the spontaneous second step of the tandem

1610 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 9
Meier et al.
reaction to lead to the monophosphates. This could
explain, in part, the above-mentioned independency of
the antiviral activity from the substituents of the
cycloSal residue. However, this assumption has been
proven by separate hydrolysis studies using isolated
enzymes (e.g., lysate, carboxyesterase) or cell extracts.
The SI values decreased relative to the electron-donat-
ing properties of the substituent resulting in higher
cytotoxicity (CC50 value) and lower SI values. The
reason for this increase in cytotoxicity remains unclear.
This trend can be seen in both series 9 and 10. It should
be added that the salicylic alcohols have been tested for
comparison and showed no antiviral activity and no
toxicity at concentrations >500 µM (data not shown).
In addition to their antiviral activity in HIV-infected
CEM cells, the cycloSal prodrugs 9 and 10 were evalu-
ated against MSV-induced transformation of C3H/3T3
embryo murine fibroblasts. As summarized in Table 2,
both series of ddAMP and d4AMP phosphotriesters
improved bioactivity up to 35-fold for the cycloSal-
ddAMPs and 30-fold for the cycloSal-d4AMPs. The
minimal cytotoxic concentration (MCC) in this assay
was >50 µM in all cases. As in the HIV/CEM assay, no
important differences in bioactivity were observed with
the different substituents on the cycloSal moiety. Ad-
ditionally, ddI, d4I, and the cycloSal-dd(4)IMP triesters
were devoid of any biological activity in the MSV assay
(EC50 > 50 µM).
for the observed pronounced improvement in antiviral
potency can be additionally rationalized in terms of
bypassing the need to convert hypoxanthine-bearing
nucleosides and nucleotides into adenine nucleotides
and the circumvention of enzymatic catabolism by PNP.
In conclusion, at least four enzymatic reactions that are
normally necessary to metabolize ddA/d4A into ddAMP/
d4AMP can be circumvented by the use of the described
cycloSal-pronucleotide approach. Noteworthy are the
considerably higher selectivity indices of some the
compounds. The phosphotriesters described here will be
investigated in the future for their hydrolytic behavior
in crude cell extracts. Also, cellular uptake and metabo-
lism of the prodrug will have to be monitored with the
use of radiolabeled derivatives.
Exp er im en ta l Section
All experiments involving water-sensitive compounds were
conducted under scrupulously dry conditions (argon atmo-
sphere). Solvents: Anhydrous dimethylformamide (DMF) and
anhydrous tetrahydrofuran (THF) were obtained in a Sure/
Seal bottle from Fluka and stored over 4 Å molecular sieves;
ethyl acetate, methylene chloride, and methanol employed in
chromatography were distilled before used. Triethylamine was
distilled from CaH2 prior to use. The solvents for the HPLC
were obtained from Merck (acetonitrile, HPLC grade) and
Riedel-de-Hae¨n (water, HPLC grade). Evaporation of solvents
was carried out on a rotary evaporator under reduced pressure
or using a high-vacuum pump. Chromatography: Chroma-
totron (Harrison Research 7924), silica gel 60PF (Merck,
gipshaltig); UV detection at 254 nm. Column chromatogra-
phy: Merck silica gel 60 (40-60 µm). TLC: analytical thin-
layer chromatography was performed on Merck precoated
aluminum plates 60 F254 with a 0.2-mm layer of silica gel
containing a fluorescent indicator; sugar-containing com-
pounds were visualized with the sugar spray reagent (0.5 mL
of 4-methoxybenzaldehyde, 9 mL of ethanol, 0.5 mL of con-
centrated sulfuric acid, and 0.1 mL of glacial acetic acid) by
heating with a fan or a hot plate. HPLC: (Merck-Hitachi)
semipreparative HPLC, LiChroCART 250-10 containing Li-
Chrospher 100 RP-18 (10 µm); analytical HPLC, EcoCART
125-3 with LiChrospher 100 RP-18 endcapped (5 µm), gradient
I 0-100% CH3CN (0-20 min), 100% CH3CN (20-22 min), 0%
CH3CN (22.1-35 min), flow 0.5 mL, UV detection at 260 nm;
gradient II (standard gradient) 11-100% CH3CN (0-20 min),
100% CH3CN (20-22 min), 0% CH3CN (22.1-35 min), flow
0.5 mL, UV detection at 260 nm; gradient III 0-80% CH3CN
(0-20 min), 100% CH3CN (20.1-22 min), 0% CH3CN (22.1-
35 min), flow 0.5 mL, UV detection at 260 nm. NMR spectra
were recorded using (1H NMR) Bruker DMX 600 at 600 MHz,
AMX 400 at 400 MHz, or Bruker WH 270 at 270 MHz (DMSO
as internal standard); (13C NMR) Bruker AMX 400 or DMX
600 at 100.6 or 150.9 MHz, respectively (CDCl3 or DMSO as
internal standard); (31P NMR) Bruker AMX 400 at 162 MHz
Con clu sion
From the antiviral data presented, it can be concluded
that within each series the adenine nucleotides are
being released intracellularly. This is due to the in-
ability of the cycloSal triesters to serve as substrates
for ADA and AMPDA. Additionally, it seems that the
compounds are taken up intracellularly by passive
diffusion, a property that is significantly enhanced
relative to the parent nucleosides (log P < 0) by the
higher log P values (log P > 0). Compared to the
cycloSal-d4TMP phosphotriesters, the compounds pre-
sented here showed a marked improvement in biological
activity, thus fulfilling the premise of the cycloSal
concept. Moreover, the antiviral activity data are com-
parable to those achieved using alternative prodrug
approaches.16-18 As shown by the chemical hydrolysis
studies, the difference between the cycloSal approach
reported here and these other alternatives is the
intrinsic independence of the cycloSal system from
hydrolytic enzymes, as for example with the bisSATE17
approach (applied to ddA) and the phosphoramidates
approach16,18 (applied to ddA and d4A). However, in
contrast to the results of the corresponding cycloSal-
d4TMP derivatives,22a the results reported here may
also point to an enzymatic contribution to the triester
cleavage in addition to or instead of the chemical
hydrolysis.
1
(H3PO4 as external standard). All H and 13C NMR chemical
shifts (δ) are quoted in parts per million (ppm) downfield from
tetramethylsilane, (CD3)(CDH)SO being set at δH 2.49 as a
reference. 31P NMR chemical shifts are quoted in ppm using
H3PO4 as external reference. The spectra were recorded at
room temperature, and all 13C and 31P NMR spectra were
recorded in the proton-decoupled mode. UV spectra were taken
with a Varian Cary 1E UV/VIS spectrophotometer. Infrared
spectra were recorded with a Perkin-Elmer 1600 series FT-
IR spectrometer in KBr pellets. Mass spectra were obtained
with a Fisons electrospray VG platform II spectrometer in
positive mode (ESI+). The test compounds were isolated as
mixtures of diastereomers arising from the mixed stereochem-
istry at the phosphate center. The resulting lyophilized tri-
esters did not give useful microanalytical data, most probably
due to incomplete combustion of the compounds, but were
found to be pure by rigorous HPLC analysis (three different
It can be concluded that the reported cycloSal phos-
photriesters are suitable to selectively deliver ddAMP
and d4AMP. This leads to a marked increase (up to 600-
fold) in bioactivity against HIV-1 and HIV-2 in infected
CEM cell cultures. The mechanism of action of the
cycloSal compounds is based on the successful bypass
of ADA, an enzyme responsible for a fast deamination
of adenine-containing nucleoside analogues. The reason

cycloSal-NMP of ddA and d4A
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 9 1611
gradients), high-field multinuclear NMR spectroscopy, and
electrospray mass spectrometry.
NH2); 7.21-7.16 (m, 2H, H-aryl); 7.07-7.00 (m, 4H, H-aryl);
6.26-6.21 (m, 2H, H1′); 5.43 (dd, 1H, HA-benzyl); 5.36 (dd, 1H,
HA-benzyl); 5.33 (dd, 1H, HB-benzyl); 5.30 (dd, 1H, HB-benzyl);
4.39-4.19 (m, 6H, H4′, H5′); 2.50-2.42 (m, 4H, H2′); 2.18 (s,
3H, CH3); 2.12 (s, 3H, CH3), 2.20-2.06 (m, 4H, H3′); 13C NMR
(63 MHz, DMSO-d6) δ 155.98 (C6); 152.46 (C2); 148.86 (C4);
147.87, 147.77 (d, C2-aryl); 138.91, 138.82 (C8); 130.75 (C4-
aryl); 126.84, 126.72 (C3-aryl); 123.81 (C5-aryl); 123.40 (C6-
aryl); 120.90, 120.84 (d, C1-aryl); 119.07 (C5); 84.31, 84.21
(C1′); 78.60, 78.49 (C4′); 69.23, 69.14 (C5′); 68.38, 68.27 (d, CH2-
benzyl); 30.71, 30.63 (C3′); 25.81, 25.73 (C2′); 14.81, 14.72
(CH3-C3-aryl); 31P NMR (162 MHz, DMSO-d6) δ -9.11; -9.21;
(ESI+) m/z 418.1; UV (CH3CN) λmax 259.87 nm, 195.98 nm; λmin
230.58 nm; IR (KBr) ν 3332.2, 3179.0, 2955.2, 1647.9, 1598.4,
1575.7, 1473.8, 1415.1, 1367.9, 1296.7, 1190.2, 1090.6, 1017.3,
939.8, 883.3, 823.0, 774.5, 723.9, 696.8, 651.6, 532.7; Rf value
0.38 (CH2Cl2/MeOH, 9:1); analytical HPLC tR 16.15 min, 16.29
min (99.4%, gradient I); 12.09 min, 12.48 min (99.4%, gradient
II); 18.59 min, 18.77 min (99.4%, gradient III).
Gen er a l P r oced u r e for th e P r ep a r a tion of th e cy-
cloSal-ddAMP s 9, cycloSal-d4AMP s 10, 3-Meth yl-cycloSal-
d d IMP (11c), a n d 3-Meth yl-cycloSa l-d 4IMP (12c). The
reactions were performed in an argon atmosphere under
anhydrous conditions. To a solution of the nucleoside 1, 2, 3,
or 4 (0.20 mmol) in 2 mL of DMF and 1 mL of THF, cooled to
-40 °C, was added DIPEA (2.0 equiv, 0.40 mmol). Then the
chlorophosphanes 13a -d 28 (2.0 equiv, 0.4 mmol) were added
within 0.5 h, and the solutions were stirred for 20 min to
complete the reaction (TLC analysis). For the oxidation of the
intermediate cyclic phosphites, tert-butyl hydroperoxide (2.0
equiv, 0.4 mmol) was added to the reaction mixture at -40
°C. After stirring for 0.5 h, the reaction mixture was warmed
to room temperature, and the solvent was removed under
reduced pressure. The residues were purified twice by chro-
matography on silica gel plates on a chromatotron, first using
a gradient of CH3OH in ethyl acetate (0-30% methanol)
followed by a gradient of CH3OH in CH2Cl2 (0-20%), to yield
the title compounds 9-12.
cycloSa lige n yl-5′-O-(2′,3′-d id e oxya d e n osin yl)p h os-
p h a te (cycloSa l-d d AMP , 9a ): yield 47%; 1H NMR (400 MHz,
DMSO-d6) δ 8.31 (s, 1H, H8); 8.27 (s, 1H, H8); 8.19 (s, 1H,
H2); 8.18 (s, 1H, H2); 7.41 (t, 1H, H4-aryl); 7.37 (t, 1H, H4-
aryl); 7.31 (s, 2H, NH2); 7.30 (s, 2H, NH2); 7.27 (d, 1H, H5-
aryl); 7.26 (d, 1H, H5-aryl); 7.23 (d, 2H, H6-aryl); 7.16 (d, 1H,
H6-aryl); 7.08 (d, 1H, H6-aryl); 6.29 (dd, 1H, H1′); 6.28 (dd,
1H, H1′); 5.51 (dd, 1H, HA-benzyl); 5.45 (dd, 1H, HA-benzyl);
5.43 (dd, 1H, HB-benzyl); 5.39 (dd, 1H, HB-benzyl); 4.44-4.27
(m, 6H, H4′, H5′); 2.61-2.45 (m, 2H, H2′); 2.22-2.10 (m, 2H,
H3′); 13C NMR (63 MHz, DMSO-d6) δ 156.03 (C6); 152.53 (C2);
149.34, 149.29 (d, C2-aryl); 148.93 (C4); 138.98, 138.92 (C8);
129.68 (C6-aryl); 125.97 (C4-aryl); 124.34 (C5-aryl); 121.02,
120.86 (d, C1-aryl); 119.08 (C5); 118.18, 118.04 (C3-aryl); 84.35,
84.28 (C1′); 78.50, 78.46 (C4′); 69.15, 69.05 (C5′); 68.40, 68.30
(t, CH2-benzyl); 30.65 (C3′); 25.80 (C2′); 31P NMR (162 MHz,
DMSO-d6) δ -9.77; -9.83; (ESI+) m/z 404.1; UV (CH3CN) λmax
259.62.5 nm, 210.07 nm; λmin 230.24 nm, 203.10 nm; IR (KBr)
ν 3325.9, 3175.7, 2954.4, 1651.3, 1598.5, 1574.1, 1488.3,
1459.5, 1415.3, 1370.0, 1296.8, 1245.4, 1222.8, 1192.0, 1093.5,
1043.8, 1019.5, 991.1, 938.2, 881.4, 842.5, 799.5, 760.0, 724.2,
696.8, 652.8, 579.4, 533.5; Rf value 0.35 (CH2Cl2/MeOH, 9:1);
analytical HPLC tR 17.11 min (99.6%, gradient I); 13.51 min
(99.6%, gradient II); 19.49 min (99.5%, gradient III).
cyclo(3,5-Dim eth ylsa ligen yl)-5′-O-(2′,3′-d id eoxya d en os-
in yl)p h osp h a te (3,5-d iMe-cycloSa l-d d AMP , 9d ): yield 50%;
1H NMR (600 MHz, DMSO-d6) δ 8.22 (s, 1H, H2); 8.19 (s, 1H,
H2); 8.11 (s, 1H, H8); 8.10 (s, 1H, H8); 7.24 (s, 2H, NH2); 7.22
(s, 1H, NH2); 7.00 (s, 1H, H4-aryl); 6.96 (s, 1H, H4-aryl); 6.81
(s, 1H, H6-aryl); 6.78 (s, 1H, H6-aryl); 6.22 (dd, 1H, H1′); 6.20
(dd, 1H, H1′); 5.36 (dd, 1H, HA-benzyl); 5.29 (dd, 1H, HA-
benzyl); 5.28 (dd, 1H, HB-benzyl); 5.23 (dd, 1H, HB-benzyl);
4.34-4.15 (m, 6H, H4′, H5′); 2.48-2.43 (m, 2H, H2′); 2.20 (s,
6H, CH3-C5-aryl); 2.12 (s, 6H, CH3-C3-aryl); 2.13-2.08 (m,
4H, H3′); 13C NMR (150.9 MHz, DMSO-d6) δ 155.92 (C6);
152.41, 152.39 (C2); 148.82, 148.78 (C4); 145.67, 145.63 (d, C2-
aryl); 138.86, 138.78 (C8); 132.89, 132.85 (C5-aryl); 131.17,
131.15 (C6-aryl); 126.39, 126.34 (C4-aryl); 123.41, 123.39 (C3-
aryl); 120.37, 120.32 (d, C1-aryl); 119.00, 118.96 (C5); 84.22,
84.13 (C1′); 78.51, 78.46 (d, C4′); 69.04, 69.01 (C5′); 68.32, 68.20
(d, CH2-benzyl); 31P NMR (162 MHz, DMSO-d6) δ -9.076;
-9.18; (ESI+) m/z 432.4, 453.9 (M + Na+); UV (CH3CN) λmax
259.16 nm, 197.96 nm; λmin 231.02 nm; IR (KBr) ν 3329.3,
3185.7, 2927.7, 1647.0, 1600.3, 1573.1, 1483.4, 1415.4, 1363.7,
1331.2, 1299.7, 1245.1, 1202.1, 1150.0, 1096.8, 1045.8, 993.4,
931.9, 858.9, 799.1, 698.2, 665.4, 582.7, 487.9; Rf value 0.39
(CH2Cl2/MeOH, 9:1); analytical HPLC tR 117.95 min, 18.09 min
(98.8%, gradient I); 15.16 min, 15.40 min (98.8%, gradient II);
20.53 min, 20.69 min (98.9%, gradient III).
cycloSa ligen yl-5′-O-(2′,3′-d id eoxy-2′,3′-d id eh yd r oa d e-
1
n osin yl)p h osp h a te (cycloSa l-d 4AMP , 10a ): yield 54%; H
cyclo(5-Meth oxysa ligen yl)-5′-O-(2′,3′-d id eoxya d en osi-
n yl)p h osp h a te (5-OMe-cycloSa l-d d AMP , 9b): yield 67%; 1H
NMR (400 MHz, DMSO-d6) δ 8.33 (s, 1H, H8); 8.30 (s, 1H,
H8); 8.23 (s, 1H, H2); 8.21 (s, 1H, H2); 7.34 (s, 2H, NH2); 7.33
(s, 2H, NH2); 7.12 (d, 1H, H3-aryl); 7.04 (d, 1H, H3-aryl); 6.98
(dd, 1H, H4-aryl); 6.93 (dd, 1H, H4-aryl); 6.89 (d, 1H, H6-aryl);
6.87 (d, 1H, H6-aryl); 6.32 (m, 2H, H1′), 5.50 (dd, HA-benzyl);
5.43 (dd, HA-benzyl); 5.41 (dd, HB-benzyl); 5.37 (dd, HB-benzyl);
4.44-4.27 (m, 6H, H4′, H5′); 3.81 (s, 3H, OCH3); 3.41 (s, 3H,
OCH3); 2.61-2.49 (m, 4H, H2′); 2.27-2.11 (m, 4H, H2′); 13C
NMR (63 MHz, DMSO-d6) δ 156.02 (C6); 155.51 (C5-aryl);
152.55 (C2); 148.93 (C4); 142.92, 142.85 (d, C2-aryl); 138.94
(C8); 121.75, 121.71 (d, C1-aryl); 119.07 (C3-aryl); 118.98 (C5);
115.09, 115.00 (C6-aryl); 110.54, 110.46 (C4-aryl); 84.28 (C1′);
78.62, 78.52 (C4′); 69.13 (C5′), 68.37, 68.25 (d, CH2-benzyl);
55.55 (OCH3), 30.73 (C3′); 25.80 (C2′); 31P NMR (162 MHz,
DMSO-d6) δ -9.57; -9.63; (ESI+) m/z 434.1; UV (CH3CN) λmax
259.39 nm, 195.02 nm; λmin 236.18 nm; IR (KBr) ν 3331.2,
3177.6, 2954.0, 2837.9, 1647.7, 1598.7, 1575.3, 1497.9, 1472.9,
1431.2, 1364.5, 1297.8, 1244.1, 1198.4, 1159.9, 1094.0, 1024.7,
948.9, 914.9, 866.8, 798.9, 759.6, 726.8, 695.9, 650.0, 568.6,
524.3, 480.6; Rf value 0.36 (CH2Cl2/MeOH, 9:1); analytical
HPLC tR 16.53 min, 16.75 min (98.6%, gradient I); 12.33 min,
12.79 min (98.6%, gradient II); 19.01 min, 19.21 min (98.6%,
gradient III).
NMR (600 MHz, DMSO-d6) δ 8.141 (s, 1H, H2); 8.143 (s, 1H,
H2); 7.92 (s, 1H, H8); 7.93 (s, 1H, H8); 7.29 (s, 2H, NH2); 7.28
(s, 2H, NH2); 7.35 (t, 1H, H4-aryl); 7.27 (t, 1H, H4-aryl); 7.23
(dd, 1H, H6-aryl); 7.19 (dd, 1H, H6-aryl); 7.16 (dt, 1H, H5-
aryl); 7.15 (dt, 1H, H5-aryl); 7.11 (dd, 1H, H3-aryl); 6.97 (dd,
1H, H3-aryl); 6.94 (ddd, 1H, H1′); 6.92 (ddd, 1H, H1′); 6.48
(ddd, 1H, H3′) 6.43 (ddd, 1H, H3′); 6.25 (ddd, 1H, H2′); 6.22
(ddd, 1H, H2′); 5.46 (dd, 1H, HA-benzyl); 5.38 (dd, 1H, HA-
benzyl); 5.35 (dd, 1H, HB-benzyl); 5.32 (dd, 1H, HB-benzyl);
5.09-5.06 (m, 2H, H4′); 4.36-4.26 (m, 4H, H5′); 13C NMR
(150.9 MHz, DMSO-d6) δ 155.96 (C6); 152.64, 152.61 (C2);
149.28, 149.23 (d, C2-aryl); 149.13, 149.12 (C4); 138.51, 138.44
(C8); 132.16, 132.09 (C3′); 129.67, 129.58 (C6-aryl), 126.64
(C2′); 125.93 (C4-aryl); 124.34, 124.28 (C5-aryl); 120.99, 120.81
(d, C1-aryl); 118.76 (C5); 118.08, 118.03 (d, C3-aryl); 87.79,
87.77 (C1′); 84.85, 84.83 (d, C4′); 68.39, 68.22 (d, C5′); 68.29,
68.27 (d, CH2-benzyl); 31P NMR (162 MHz, DMSO-d6) δ -9.65;
-9.84; MS (ESI+) m/z 401.9, 423.9 (M + Na+); UV (CH3CN)
λmax 259.06 nm, 194.00 nm; λmin 228.10 nm; IR (KBr) ν 3333.0,
3182.9, 1650.0, 1598.9, 1488.5, 1418.2, 1371.0, 1293.1, 1242.7,
1193.3, 1093.7, 1021.0, 949.7, 846.2, 828.5, 761.3, 720.8, 650.3,
578.2; Rf value 0.35 (CH2Cl2/MeOH, 9:1); analytical HPLC tR
15.72 min (98.6%, gradient I); 9.43 min, 9.87 min (98.5%,
gradient II); 18.03 min (98.6%, gradient III).
cyclo(3-Meth ylsaligen yl)-5′-O-(2′,3′-dideoxyaden osin yl)-
p h osp h a te (3-Me-cycloSa l-d d AMP , 9c): yield 58%; 1H NMR
(400 MHz, DMSO-d6) δ 8.24 (s, 1H, H8); 8.21 (s, 1H, H8); 8.14
(s, 1H, H2); 8.13 (s, 1H, H2); 7.27 (s, 2H, NH2); 7.26 (s, 2H,
cyclo(5-Meth oxysa ligen yl)-5′-O-(2′,3′-d id eoxy-2′,3′-d id e-
h yd r oa d en osin yl)p h osp h a t e (5-OMe-cycloSa l-d 4AMP ,
10b): yield 50%; 1H NMR fast diastereoisomer (600 MHz,
DMSO-d6) δ 8.16 (s, 1H, H2); 7.97 (s, 1H, H8); 7.30 (s, 2H,

1612 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 9
Meier et al.
NH2); 6.94 (ddd, 1H, H1′); 6.91(dd, 1H, H4-aryl); 6.82-6.81
(m, 2H, H3-aryl., H6-aryl); 6.45 (dt, 1H, H3′); 6.24 (ddd, 1H,
H2′); 5.43 (dd, 1H, HA-benzyl); 5.31 (dd, 1H, HB-benzyl); 5.08
(m, 1H, H4′); 4.33 (ddd, 1H, HB5′); 4.29 (ddd, 1H, HA5′); 3.73
(s, 3H, OCH3); 31P NMR fast diastereoisomer (162 MHz,
DMSO-d6) δ -9.44; 1H NMR slow diastereoisomer (600 MHz,
DMSO-d6) δ 8.16 (s, 1H, H2); 7.99 (s,1H, H8); 7.31 (s, 2H, NH2);
7.07 (d, 1H, H3-aryl); 6.96 (ddd, 1H, H1′); 6.92 (dd, 1H, H4-
aryl); 6.79 (d, 1H, H6-aryl); 6.50 (dt, 1H, H3′); 6.26 (ddd, 1H,
H2′); 5.35 (ddd, 1H, HA-benzyl); 5.28 (ddd, 1H, HB-benzyl); 5.08
(m, 1H, H4′); 4.30 (m, 2H, H5′); 3.74 (s, 3H, OCH3); 31P NMR
slow diastereoisomer (162 MHz, DMSO-d6) δ -9.65; 13C NMR
(150.9 MHz, DMSO-d6) δ 155.97 (C6); 155.50, 155.43 (C5-aryl);
152.69, 152.66 (C2); 149.15, 149.13 (C4); 142.85, 142.83 (d, C2-
aryl); 138.52, 138.48 (C8); 132.17, 132.10 (C3′); 126.64 (C2′);
121.68, 121.56 (d, C1-aryl); 118.95, 118.89 (C3-aryl); 118.74
(C5); 115.08, 114.85 (C6-aryl); 110.57, 110.42 (C4-aryl); 87.77,
87.75 (C1′); 84.88, 84.83 (C4′); 68.36, 68.23 (d, C5′); 68.29, 68.19
(d, CH2-benzyl); 55.50, 55.47 (OCH3); MS (ESI+) m/z 431.9,
453.9 (M + Na+); UV (CH3CN) λmax 258.99 nm, 195.88 nm;
λmin 236.23 nm; IR (KBr) ν 3336.2, 3185.9, 2946.8, 1647.5,
1598.8, 1576.8, 1497.9, 1473.3, 1420.2, 1293.2, 1241.5, 1200.5,
1125.9, 1094.2, 1026.1, 994.3, 951.5, 911.1, 827.9, 799.3, 760.4,
721.1, 649.4; Rf value 0.35 (CH2Cl2/MeOH, 9:1); analytical
HPLC tR 16.25 min, 16.39 min (98.7%, gradient I); 11.88 min,
12.12 min (98.7%, gradient II); 18.44 min (98.7%, gradient III).
1328.9, 1287.0, 1238.8, 1202.8, 1150.0, 1125.9, 1094.9, 1028.0,
993.4, 970.4, 951.6, 932.9, 850.2, 828.7, 798.9, 721.3, 665.4,
583.2, 495.4; Rf value 0.35 (CH2Cl2/MeOH, 9:1); analytical
HPLC tR 17.48 min (98.0%, gradient I); 14.51 min (98.0%,
gradient II); 20.03 min (98.1%, gradient III).
cyclo(3-Met h ylsa ligen yl)-5′-O-(2′,3′-d id eoxyin osin yl)-
p h osp h a te (3-Me-cycloSa l-d d IMP , 11c): yield 35%; 1H NMR
(600 MHz, DMSO-d6) δ 12.33 (s, 2H, NH); 8.17 (s, 1H, H2);
8.14 (s, 1H, H2); 8.01 (s, 1H, H8); 8.00 (s, 1H, H8); 7.20 (t, 1H,
H4-aryl); 7.18 (m, 1H, H4-aryl); 7.06-7.02 (m, 4H, H5-aryl,
H6-aryl); 6.20 (dd, 1H, H1′); 6.19 (dd, 1H, H1′); 5.43 (dd, 1H,
HA-benzyl); 5.38 (dd, 1H); 5.33 (dd, 1H, HB-benzyl); 5.31 (dd,
1H, HB-benzyl); 4.35-4.17 (m, 6H, H5′, H4′); 2.48-2.37 (m 4H,
H2′); 2.18 (s, 3H, CH3); 2.16 (s, 3H, CH3); 2.12-2.01 (m, 4H,
H3′); 13C NMR (100.6 MHz, DMSO-d6) δ 156.48 (C6); 147.82,
147.78 (C2-aryl); 147.62 (C4); 145.55 (C2); 138.01, 137.96 (C8);
130.74 (C4-aryl); 126.81, 126.73 (C3-aryl); 12441, 124.37 (C5),
123.79 (C5-aryl); 123.40 (C6-aryl); 120.81, 120.74 (d, C1-aryl);
84.47, 84.38 (C1′); 78.87, 78.79 (C4′); 69.06, 69.00 (C5′); 68.31,
68.22 (d, CH2-benzyl); 31.10, 31.07 (C3′); 25.59, 25.56 (C2′);
14.78, 14.70 (CH3-C3-aryl); 31P NMR (162 MHz, DMSO-d6) δ
-9.14; -9.20; (ESI+) m/z 419.3; 441.1 (M + Na+); UV (CH3-
CN) λmax 250.53 nm, 196.02 nm; λmin 225.97 nm; IR (KBr) ν
3437.5, 3055.8, 2955.5, 1697.4, 1622.1, 1586.5, 1545.7, 1511.4,
1471.9, 1412.9, 1370.0, 1346.4, 1295.7, 1213.0, 1190.2, 1125.3,
1090.3, 1025.0, 1028.1, 990.9, 939.0, 881.6, 818.1, 789.6, 719.8,
688.0, 651.1, 606.6, 528.0; Rf value 0.29 (CH2Cl2/MeOH, 9:1);
analytical HPLC tR 16.35 min (96.7%, gradient I); 12.21 min
(96.7%, gradient II); 18.92 min (96.7%, gradient III).
cyclo(3-Meth ylsa ligen yl)-5′-O-(2′,3′-d id eoxy-2′,3′-d id e-
h yd r oin osin yl)p h osp h a te (3-Me-cycloSa l-d 4IMP , 12c):
yield 35%; 1H NMR (600 MHz, DMSO-d6) δ 12.36 (s, 2H, NH);
8.05 (s, 1H, H2); 8.04 (s, 1H, H2); 7.87 (s, 1H, H8); 7.86 (s,
1H, H8); 7.23-7.12 (m, 2H, H-aryl); 7.07-7.02 (m, 4H, H-aryl);
6.91-6.88 (m, 2H, H1′); 6.50 (ddd, 1H, H3′); 6.45 (ddd, 1H,
H3′); 6.24-6.21 (m, 2H, H2′); 5.45 (dd, 1H, HA-benzyl); 5.41
(dd, 1H, HA-benzyl); 5.29 (dd, 1H, HB-benzyl); 5.28 (dd, 1H,
HB-benzyl); 4.33-4.22 (m, 4H, H5′); 13C NMR (150.1 MHz,
DMSO-d6) δ 156.46, 156.44 (C6); 147.88 (C4); 147.71, 147.68
(d, C2-aryl); 145.86 (C2); 137.80, 137.74 (C8); 132.41, 132.36
(C3′): 130.76, 130.75 (C4-aryl); 126.79, 126.72 (d, C3-aryl);
126.37, 126.35 (C2′); 124.13, 124.11 (C5); 123.80 (C6-aryl);
123.41, 123.36 (C5-aryl); 120.78, 120.58 (d, C1-aryl); 88.07,
88.06 (C1′); 85.14, 85.06 (d, C4′); 68.29. 68.22 (C5′); 68.18, 68.03
(d, CH2-benzyl,); 14.79, 14.64 (CH3-C3-aryl); 31P NMR (162
MHz, DMSO-d6) δ -9.01; -9.23; (ESI+) m/z 417.1; UV (CH3-
CN) λmax 250.00 nm, 245.04 nm; 203.04 nm; λmin 248.59 nm;
226.84 nm; IR (KBr) ν 3447.8, 1686.1, 1588.4, 1544.6, 1474.2,
1289.6, 1207.3, 1189.7, 1090.9, 1025.4, 998.3, 954.2, 827.6,
784.4, 649.7; Rf value (CH2Cl2/MeOH, 9:1) 0.33; analytical
HPLC tR 16.20 min, 16.07 min (95.7%, gradient I); 10.80 min,
11.29 min (95.7%, gradient II); 18.47 min, 18.57 min (95.8%,
gradient III).
Det er m in a t ion of t h e P a r t it ion Coefficien t s (log P
Va lu es). log P values were determined as follows: a sample
of the compounds 9-12 was dissolved in 0.3 mL of 1-octanol.
To this solution was added 0.3 mL of water. After the phases
were mixed extensively for 10 min (vortex) and separated by
centrifugation (2 min at 9000 rpm), aliquots of each phase were
analyzed by analytical HPLC (Merck EcoCART column filled
with Lichrospher 100 reversed-phase silica gel RP-18 end-
capped (5 µm), gradient 0-100% CH3CN in water (0-20 min),
100% CH3CN (20-22 min), 0-100% CH3CN in water (22.1-
35 min), flow 0.5 mL, UV detection at 260 nm). The P values
were calculated by integration of the peaks of the aqueous and
organic phases.
Kin etic Da ta . DMSO stock solution of the compound (9 µL,
50 mM) was diluted in 312 µL of water (c ) 1.44 mM); 0.3 mL
of this solution was added to 0.3 mL (37 °C) of aqueous buffer
(50 mM phosphate buffer (So¨rensen mixture), pH 7.3; 50 mM
glycine/HCl, pH 3.0; 50 mM glycine/HCl, pH 1.5) containing 5
µL of an aqueous AZT solution (AZT as internal standard).
The final concentration of the test compounds was 0.71 mM
in 24.8 mM aqueous buffer. For kinetic data, aliquots of 40
cyclo(3-Meth ylsa ligen yl)-5′-O-(2′,3′-d id eoxy-2′,3′-d id e-
h ydr oaden osin yl)ph osph ate (3-Me-cycloSal-d4AMP , 10c):
yield 56%; 1H NMR (600 MHz, DMSO-d6) δ 8.133 (s, 1H, H2);
8.130 (s, 1H, H2); 7.96 (s, 1H, H8); 7.95 (s, 1H, H8); 7.26 (s,
2H, NH2); 7.25 (s, 2H, NH2); 7.22-7.01 (m, 6H, aryl. H); 6.94
(m, 2H, H1′); 6.48 (dt, 1H, H2′); 6.43 (dt, 1H, H2′); 6.24 (ddd,
1H, H3′); 5.42 (dd, 1H, HA-benzyl); 5.34 (dd, 1H, HA-benzyl);
5.33 (dd, 1H, HB-benzyl); 5.28 (dd, 1H, HB-benzyl); 5.08 (m,
2H, H4′); 4.31-4.26 (m, 4H, H5′); 2.19 (s, 3H, CH3); 2.10 (s,
3H, CH3); 13C NMR (150.9 MHz, DMSO-d6) δ 155.96 (C6);
152.65 (C2); 149.13 (C4); 147.83, 147.73 (d, C2-aryl); 138.48,
138.39 (C8); 132.14, 132.07 (C3′); 130.78 (C4-aryl); 126.82 (C3-
aryl); 126.64 (C2′); 123.82 (C5-aryl); 123.43, 123.37 (C6-aryl);
120.92, 120.71 (d, C1-aryl); 118.77, 118.74 (C5); 87.80, 87.77
(C1′); 84.91, 84.84 (d, C4′); 68.44, 68.30 (d, C5′); 68.24, 68.19
(d, CH2-benzyl); 14.83, 14.65 (C3-aryl-CH3); 31P NMR (162
MHz, DMSO-d6) δ -9.44; -9.65; MS (ESI+) m/z 415.9, 437.9
(M + Na+); UV (CH3CN) λmax 258.96 nm, 194.38 nm; λmin
228.12 nm; IR (KBr) ν 3328.4, 3182.1, 2944.0, 2886.7, 1649.9,
1597.9, 1574.8, 1472.9, 1418.7, 1368.0, 1328.9, 1293.7, 1238.8,
1203.8, 1190.4, 1126.2, 1090.7, 1023.7, 993.4, 945.0, 880.8,
828.7, 775.4, 720.8, 648.3, 614.1, 583.8, 505.1; Rf value 0.35
(CH2Cl2/MeOH, 9:1); analytical HPLC tR 16.21 min, 16.32 min
(97.1%, gradient I); 11.99 min, 12.35 min (97.2%, gradient II);
19.03 min (97.0%, gradient III).
cyclo(3,5-Dim eth ylsa ligen yl)-5′-O-(2′,3′-d id eoxy-2′,3′-d i-
deh ydr oaden osin yl)ph osph ate (3,5-diMe-cycloSal-d4AMP ,
10d ): yield 51%; 1H NMR (600 MHz, DMSO-d6) δ 8.13 (s, 2H,
H2); 7.96 (s, 1H, H2); 7.95 (s, 1H, H8); 7.27 (s, 2H, NH2); 7.25
(s, 2H, NH2); 7.00 (s, 1H, H4-aryl); 6.97 (s, 1H, H4-aryl); 6.92
(m, 2H, H1′); 6.82 (s, 1H, H6-aryl); 6.77 (s, 1H, H6-aryl); 6.47
(ddd, 1H, H2′); 6.43 (ddd, 1H, H2′); 6.24 (ddd, 1H, H3′); 6.22
(ddd, 1H, H3′); 5.38 (dd, 1H, HA-benzyl), 5.28 (dd, 1H, HA-
benzyl); 5.27 (dd, 2H, HB-benzyl), 5.22 (dd, 2H, HB-benzyl); 5.06
(m, 2H, H4′); 4.31 (ddd, 1H, H5′); 4.28 (ddd, 1H, H5′); 4.25
(ddd, 1H, H5′′); 4.22 (ddd, 1H, H5′′); 2.20 (s, 3H, CH3-C5-aryl);
2.19 (s, 3H, CH3-C5-aryl); 2.14 (s, 3H, CH3-C3-aryl); 2.05 (s,
3H, CH3-C3-aryl); 13C NMR (150.9 MHz, DMSO-d6) δ 155.95
(C6); 152.63 (C2); 149.11, 149.09 (C4); 145.63, 145.58 (d, C2-
aryl); 138.47, 138.38 (C8); 132.94, 132.90 (C3′); 132.16, 132.05
(C5-aryl); 131.22 (C6-aryl); 126.62, 126.60 (C2′); 126.38, 126.35
(C4-aryl); 123.46, 123.40 (C3-aryl); 120.48, 120.30 (d, C1-aryl);
118.75, 118.73 (C5); 87.76 (C1′); 84.91, 84.82 (d, C4′); 68.30,
68.25 (d, C5′); 68.35, 68.10 (d, CH2-benzyl); 20.05, 20.01 (CH3-
C5-aryl); 14.75, 14.59 (CH3-C3-aryl); 31P NMR (162 MHz,
DMSO-d6) δ -8.96; -9.17; (ESI+) m/z 430.2; UV (CH3CN) λmax
258.98 nm, 197.25 nm; λmin 230.98 nm; IR (KBr) ν 3334.2,
3177.8, 2917.5, 1642.2, 1597.9, 1573.8, 1483.8, 1417.9, 1364.4,

cycloSal-NMP of ddA and d4A
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 9 1613
µL of the hydrolysis mixture were stopped by freezing in liquid
air and subsequently analyzed by analytical HPLC after
thawing (Merck EcoCART column filled with Lichrosphere 100
reverse-phase silica gel RP-18 endcapped (5 µm), gradient 11-
100% CH3CN in water (0-20 min), 100% CH3CN (20-22 min),
11% CH3CN (22.1-35 min), 0.5 mL flow, UV detection at 260
nm). The degradation of 9-12 was followed by integration of
the peak areas in the HPLC chromatograms. The rate con-
stants k were determined from the slope of the logarithmic
Forschungsgemeinschaft (DFG-Grant Me 1161/2-1), the
Fonds der Chemischen Industrie, the Biomedical Re-
search Programme of the European Commission, the
Fonds voor Wetenschappelijk Onderzoek (FWO)-Vlaan-
deren, and the Belgian Geconcerteerde Onderzoecksac-
ties. We thank Mrs. Ann Absillis and Mrs. Lizette van
Berckelaer for excellent technical help. We gratefully
acknowledge Dr. David G. J ohns for critical reading and
helpful suggestions to the manuscript.
degradation curve of the title compounds. The half-lives (t1/2
)
were calculated using the rate constants k. Values are means
of duplicate experiments.
Su p p or tin g In for m a tion Ava ila ble: 1H, 13C, and 31P
NMR spectra and detailed tables of coupling constants for the
title compounds 9a -d , 10a -d , 11c, and 12c. This information
is available free of charge via the Internet at http://pubs.
acs.org.
En zym a tic Stu d ies. The experiments with the parent
nucleosides were carried out using 0.25 U of adenosine
deaminase (ADA; EC 3.5.4.4; from calf intestinal mucosa) and
a 10 mM phosphate buffer, pH 7.3 (So¨rensen) at 37 °C,
whereas for 3-methyl-cycloSal-ddAMP (9c) and 3-methyl-
cycloSal-d4AMP (10c) 0.5 U of enzyme and a 10 mM phosphate
buffer (pH 6.8) were used. The concentration of the compounds
in the above-mentioned buffers was 185 µM. To 1.3 mL of this
preincubated solution (37 °C) was added ADA. The enzymatic
deamination was followed by UV spectroscopy (Varian, Cary
1E) at 260 nm for 90 min. The integrity of 9c and 10c was
demonstrated after comparison of the recorded UV spectra
before (t0) and after 90 min (t90) incubation with ADA for 90
min.
The deamination of the monophosphates, dAMP and ddAMP,
by adenosine monophosphate deaminase (AMPDA; EC 3.5.4.6;
from rabbit muscle) was carried out using 1 and 2 U of
AMPDA, respectively, in a 10 mM phosphate buffer (pH 6.8,
So¨rensen) at 37 °C, whereas for both triesters 3-methyl-
cycloSal-ddAMP (9c) and 3-methyl-cycloSal-d4AMP (10c) 2 U
of enzyme was used. The concentration of the compounds in
the above-mentioned buffers was 185 µM. To 1.3 mL of this
preincubated solution (37 °C) was added AMPDA. The deami-
nation reactions were followed by UV spectroscopy (Varian,
Cary 1E) at 260 nm for 200 min. The integrity of 9c and 10c
was demonstrated after comparison of the UV spectra before
(t0) and after 200 min (t200) incubation with AMPDA for 200
min.
An tir etr ovir a l Eva lu a tion . Human immunodeficiency
virus type 1 [HIV-1 (HTLV-IIIB)] was originally obtained from
a persistently HIV-infected H9 cell line, as described previ-
ously,36 which was kindly provided by Dr. R. C. Gallo (then at
the National Institutes of Health, Bethesda, MD). Virus stocks
were prepared from the supernatants of HIV-1-infected MT-4
cells. HIV-2 (strain ROD) was kindly provided by Dr. L.
Montagnier (Pasteur Institute, Paris, France), and virus stocks
were prepared from the supernatants of HIV-2-infected MT-4
cells. CEM cells were obtained from the American Tissue
Culture Collection (Rockville, MD). CEM cells were infected
with HIV as previously described.37 Briefly, 4 × 105 cells/mL
were infected with HIV-1 or HIV-2 at ∼100 CCID50 (50% cell
culture infective dose) per mL of cell suspension. Then, 100
µL of the infected cell suspension was transferred into 96-well
microtiter plate wells and mixed with 100 µL of the appropri-
ate dilutions of the test compounds. After 4-5 days, giant cell
formation was recorded microscopically in the HIV-infected
cell cultures.
Refer en ces
(1) (a) De Clercq, E. Chemotherapeutic Approaches of the Treatment
of the Acquired Immunodeficiency Syndrome. J . Med. Chem.
1986, 29, 1561-1569. (b) De Clercq, E. Towards improved Anti-
HIV Chemotherapy: Therapeutic Strategies for Intervention
with HIV Infections. J . Med. Chem. 1995, 38, 2491-2517.
(2) (a) Balzarini, J . Metabolism and Mechanism of Antiretroviral
Action of Purine and Pyrimidine Derivatives. Pharm. World Sci.
1994, 16, 113-126. (b) Herdewijn, P.; Balzarini, J .; De Clercq,
E. 2′,3′-Dideoxynucleoside Analogues as Anti-HIV Agents. In
Advances in Antiviral Drug Design; De Clercq, E., Ed.; J AI Press
Inc.: Greenwich, London, 1993; Vol. 1, pp 233-318. (c) Balzarini,
J .; Kang, G.-J .; Dalal, M.; Herdewijn, P.; De Clercq, E.; Broder,
S.; J ohns, D. G. The Anti-HTLV-III (Anti-HIV) and Cytotoxic
Activity of 2′,3′-Didehydro-2′,3′-dideoxyribonucleosides: A Com-
parison with their Parental 2′,3′-Dideoxyribonucleosides. Mol.
Pharmacol. 1987, 32, 162-167.
(3) J ohnson, M. A.; Ahluwalia, G.; Conelly, M. C.; Cooney, D. A.;
Broder, S.; J ohns, D. G.; Fridland, A. Metabolic Pathways for
the Activation of the Antiretroviral Agent 2′,3′-Dideoxyadenosine
in Human Lymphoid Cells. J . Biol. Chem. 1988, 263, 15354-
15357.
(4) Shelton, M. J .; O’Donell, M.; Morse, G. D.; Didanosine. Ann.
Pharmacother. 1992, 26, 660-670.
(5) Cooney, D. A.; Ahluwalia, G.; Mitsuya, J .; Fridland, A.; J ohnson,
M.; Hao, Z.; Dalal, M.; Balzarini, J .; Broder, S.; J ohns, D. G.
Initial Studies of the Cellular Pharmacology of 2′,3′-Dideoxy-
adenosine, an Inhibitor of HTLV-III Infection. Biochem. Phar-
macol. 1987, 36, 1765-1768.
(6) Carson, D. A.; Haertle, T.; Wasson, D. B.; Richman, D. D.
Biochemical Genetic Analysis of 2′,3′-Dideoxyadenosine Metabo-
lism in Human T Lymphocytes. Biochem. Biophys. Res. Com-
mun. 1988, 151, 788-793.
(7) Hao, Z.; Cooney, D. A.; Hartman N. R.; Perno, C. F.; Fridland,
A.; DeVico, A. L.; Sarngadharan, M. G.; Broder, S.; J ohns, D. G.
Factors Determining the Activity of 2′,3′-Dideoxynucleosides in
Suppressing Human Immunodeficiency Virus in vitro. Mol.
Pharmacol. 1988, 34, 431-435.
(8) McGowan, J . J .; Tomaszewski, J .; Cracock, J .; Hoth, D.; Grie-
shaber, C. K.; Broderand, S.; Mitsuya, H. Overview of the
Preclinical Development of an Antiretroviral Drug, 2′,3′-
Dideoxyadenosine. Rev. Infect. Dis. 1990, 12, S513-S521.
(9) McLaren, C.; Datema, R.; Knupp, C. A.; Buroker, R. A.
Didanosine. Antiviral Chem. Chemother. 1991, 2, 321-328.
(10) Fridland, A.; J ohnson, M. A.; Cooney, D. A.; Ahluwalia, G.;
Marquez, V. E.; Driscoll, J . S.; J ohns, D. G. Metabolism in
Human Leukocytes of Anti-HIV Dideoxypurine Nucleosides.
Ann. N.Y. Acad. Sci. 1990, 616, 205-216.
(11) Ahluwalia, G.; Cooney, D. A.; Mitsuya, H.; Fridland, A.; Flora,
K. P.; Hao, Z.; Dalal, M.; Broder, S.; J ohns, D. G. Initial Studies
on the Cellular Pharmacology of 2′,3′-Dideoxyinosine, an Inhibi-
tor of HIV Infectivity. Biochem. Pharmacol. 1987, 36, 3797-
3800.
(12) J ohnson, M. A.; Fridland, A. Phosphorylation of 2′,3′-Dideoxy-
inosine by Cytosolic 5′-Nucleotidase of Human Lymphoid Cells.
Mol. Pharmacol. 1989, 36, 291-295.
(13) Sommadossi, J . P. Nucleoside Analogues: Similarities and
Differences. Clin. Infect. Dis. 1993, 16, S7-S15.
(14) Gao, W. Y.; Shirasaka, T.; J ohns, D. G.; Broder, S.; Mitsuya, H.
Differential Phosphorylation of Azidothymidine, Dideoxycyti-
dine, and Dideoxyinosine in Resting and Activated Peripheral
Blood Mononuclear Cells. J . Clin. Invest. 1993, 91, 2326-2333.
(15) Nair, V.; Sells, T. B. Interpretation of the Roles of Adenylosuc-
cinate Lyase and of AMP Deaminase in the Anti-HIV Activity
of 2′,3′-Dideoxyadenosine and 2′,3′-Dideoxyinosine. Biochim.
Biophys. Acta 1992, 1119, 201-204.
C3H/3T3 cells were seeded at 20 000 cells/mL into wells of
tissue culture cluster plates (48 wells/plate). Following a 24-h
incubation period, cell cultures were infected with 80 focus-
forming units of MSV during 120 min, whereafter the culture
medium was replaced by 1 mL of fresh medium containing
appropriate concentrations of the test compounds. After 6 days,
transformation of the cells was examined microscopically. The
EC50 was defined as the compound concentration required to
inhibit MSV-induced cell transformation by 50%. The MCC
was defined as the minimum cytotoxic concentration required
to cause a microscopically visible morphological alteration of
the C3H/3T3 fibroblast cultures.
Ack n ow led gm en t. We gratefully acknowledge the
support of the Adolf-Messer-Stiftung, the Deutsche

1614 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 9
Meier et al.
(16) McGuigan, C.; Wedgwood, O.; De Clercq, E.; Balzarini, J .
Phosphoramidate Derivatives of 2′,3′-Didehydro-2′,3′-dideoxy-
adenosine (d4A) Have Markedly Improved Potency and Selectiv-
ity. Bioorg. Med. Chem. Lett. 1996, 6, 2359-2362.
(17) Perigaud, C.; Aubertin, A.-M.; Benzaria, S.; Pelicano, H.; Giradet,
J .-L.; Maury, G.; Gosselin, G.; Kirn, A.; Imbach, J .-L. Equal
Inhibition of the Replication of Human Immunodeficiency Virus
in Human T-Cell Culture by ddA Bis(Sate)phosphotriester and
3′-Azido-2′,3′-dideoxythymidine. Biochem. Pharmacol. 1994, 48,
11-14.
(18) Balzarini, J .; Kruining, J .; Wedgwood, O.; Pannecouque, C.;
Aquaro, A.; Perno, C.-F.; Naesens, L.; Witrouw, M.; Heijtink,
R.; De Clercq, E.; McGuigan, C. Conversion of 2′,3′-Dideoxyad-
enosine (ddA) and 2′,3′-Didehydro-2′,3′-dideoxyadenosine (d4A)
to their Activity against Human Immunodeficiency Virus and
Hepatitis B Virus. FEBS Lett. 1997, 410, 324-328.
(19) (a) Meier, C. Pro-Nucleotides - Recent Advances in the Design
of Efficient Tools for the Delivery of Biologically Active Nucleo-
side Monophosphates. Synlett 1998, 233-242. (b) Pe´rigaud, C.;
Giradet, J .-L.; Gosselin, G.; Imbach, J .-L. Comments on Nucle-
otide Delivery Forms. In Advances in Antiviral Drug Design; De
Clercq, E., Ed.; J AI Press Inc.: Greenwich, London, 1996; Vol.
2, pp 147-172.
(27) Beach, J . W.; Kim, H. O.; J eong, L. S.; Nampalli, S.; Islam, Q.;
Ahn, S. K.; Babu, R.; Chu, C. K. A Highly Stereoselective
Synthesis of Anti-HIV 2′,3′-Dideoxy- and 2′,3′-Didehydro-2′,3′-
dideoxynucleosides. J . Org. Chem. 1992, 57, 3887-3894.
(28) Chu, C. K.; Bhadti, V. S.; Doboszewski, B.; Gu, Z. P.; Kosugi,
Y.; Pullaiah, K. C.; Van Roey, P. General Syntheses of 2′,3′-
Dideoxynucleosides and 2′,3′-Didehydro-2′,3′-dideoxynucleosides.
J . Org. Chem. 1989, 54, 2217-2225.
(29) Meier, C.; De Clercq, E.; Balzarini, J . Nucleotide Delivery from
cycloSaligenyl-3′-Azido-3′-Deoxythymidine Monophosphates (cy-
cloSal-AZTMP). Eur. J . Org. Chem. 1998, 837-846.
(30) (a) Using a different methodology, a
P value of 0.96 was
determined for AZT: Balzarini, J .; Cools, J . M.; De Clercq, E.
Estimation of the Lipophilicity of Anti-HIV Nucleoside Ana-
logues by Determination of the Partition Coefficient and Reten-
tion Time on a LiChrosphere 60 RP-8 HPLC Column. Biochem.
Biophys. Res. Commun. 1989, 158, 413-422. (b) Lien, E. J .; Gao,
H.; Prabhakar, H. Physical Factors Contributing to the Partition
Coefficient and Retention Time of 2′,3′-Dideoxynucleoside Ana-
logues. J . Pharm. Sci. 1991, 80, 517-521.
(31) Also using a HPLC methodology, the log P values for AZT (0.05),
ddA (-0.29), and ddI (-1.24) have been determined: Barchi, J .
J .; Marquez, V. E.; Driscoll, J . S.; Ford, H., J r.; Mitsuya, H.;
Shirasaka, T.; Aoki, S.; Kelley, J . A. Potential Anti-AIDS Drugs.
Lipophilic, Adenosine Deaminase-activated Prodrugs. J . Med.
Chem. 1991, 34, 1647-1655.
(32) Zimmermann, T. P.; Mahony, W. B.; Prus, K. L. 3′-Azido-3′-
deoxythymidine, an Unusual Nucleoside Analogue that Perme-
ates the Membrane of Human Erythrocytes and Lymphocytes
by Nonfacilitated Diffusion. J . Biol. Chem. 1987, 262, 5748-
5754.
(20) Meier, C. 4H-1.3.2-Benzodioxaphosphorin-2-nucleosyl-2-oxide -
A New Concept for Lipophilic, Potential Prodrugs of Biologically
Active Nucleoside Monophosphates. Angew. Chem. 1996, 108,
77-79; Angew. Chem., Int. Ed. Engl. 1996, 35, 70-72.
(21) Meier, C.; Knispel, T.; Lorey, M.; Balzarini, J . CycloSal-Pro-
Nucleotides: The Design and Biological Evaluation of a New
Class of Lipophilic Nucleotide Prodrugs. Int. Antiviral News
1997, 5, 183-186.
(22) (a) Meier, C.; Lorey, M.; De Clercq, E.; Balzarini, J . cycloSal-
d4TMP: Synthesis and Antiviral Evaluation of a New d4TMP
Delivery System. J . Med. Chem. 1998, 41, 1417-1427. (b) Meier,
C.; Lorey, M.; De Clercq, E.; Balzarini, J . Cyclic Saligenyl
Phosphotriesters of 2′,3′-Dideoxy-2′,3′-didehydrothymidine (d4T)
- A New Pro-Nucleotide Approach. Bioorg. Med. Chem. Lett.
1997, 7, 99-104.
(23) Meier, C.; Knispel, T.; De Clercq, E.; Balzarini, J . ADA-Bypass
by Lipophilic CycloSal-ddAMP Pro-Nucleotides. A Second Ex-
ample of the Efficiency of the cycloSal-Concept. Bioorg. Med.
Chem. Lett. 1997, 7, 1577-1582.
(24) Liguori, A.; Perri, E.; Sidona, G.; Uccella, N. Regioselective 5′-
O-Benzoylation of Deoxyadenosine with Benzoic Acid Catalyzed
by N′,N′-Bis-(2-oxo-oxazolidin-3-yl)phosphordiamidic chloride
(BOPDC). Tetrahedron 1988, 44, 229-234.
(25) Ratnakar, R.; Talekar, P. L.; Coe, L.; Walker, R. T. An Efficient
and General Synthesis of 5′-Esters of 2′,3′-Didehydro-2′,3′-
Dideoxynucleosides: A Facile Opening of 2′,3′-Orthoacetates of
Ribonucleosides Followed by Reductive Elimination of the Ha-
logeno-acetates. Synthesis 1993, 303-306.
(26) Webb, R. R., II; Wos, J . A.; Martin, J . C.; Brodfuehrer, P. R.
Synthesis of 2′,3′-Dideoxyinosine. Nucleosides Nucleotides 1988,
7, 147-153.
(33) Meier, C.; Habel, L.; Haller-Meier, F.; Lomp, A.; Herderich, M.;
Klo¨cking, R.; Meerbach, A.; Wutzler, P. Chemistry and Anti-
HSV-1 Evaluation of cycloSal-Nucleotides (cycloSal-NMP) of
Acyclic Nucleoside Analogues. Antiviral Chem. Chemother. 1998,
9, in press.
(34) York, J . L. Effect of the Structure of the Glycon on the Acid-
Catalyzed Hydrolysis of Adenine Nucleosides. J . Org. Chem.
1981, 46, 2171-2173.
(35) Bloch, A.; Robins, M. J .; McCarthy, J . R., J r. The Role of the
5′-Hydroxyl Group of Adenosine in Determining Substrate
Specificity for Adenosine Deaminase. J . Med. Chem. 1967, 10,
908-912.
(36) Popovic, M.; Sarngadharan, M. G.; Read, E.; Gallo, R. C.
Detection, Isolation, and Continuous Production of Cytopathic
Retroviruses (HTLV-III) from Patients with AIDS and pre-AIDS.
Science 1984, 224, 497-500.
(37) Balzarini, J .; Naesens, L.; Slachmuylders, J .; Niphuis, H.;
Rosenberg, I.; Holy, A.; Schellekens, H.; De Clercq, E. 9-(2-
Phosphonylmethoxyethyl)adenine (PMEA) Efficiently Inhibits
Retrovirus Replication in vitro and Simian Immunodeficiency
Virus Infection in Rhesus Monkeys. AIDS 1991, 5, 21-28.
J M981096Z