New modified nucleoside 5′-triphosphates: Synthesis, properties towards DNA polymerases, stability in blood serum and antiviral activity

Article abstract of DOI:10.1039/a900336c

A series of new nucleoside 5'-triphosphate mimetics, 2,3,5,6, 8-10, modified at the glycone and all three phosphate residues, have been synthesised and studied. These compounds only bear the enzymatically labile anhydride bond between the α and β phosphor

Full text of DOI:10.1039/a900336c

View Online / Journal Homepage / Table of Contents for this issue
New modified nucleoside 5Ј-triphosphates: synthesis, properties
towards DNA polymerases, stability in blood serum and antiviral
activity
Alexander V. Shipitsin,a Lyubov S. Victorova,b Elena A. Shirokova,a Natalya B. Dyatkina,a
Lyudmila E. Goryunova,c Robert Sh. Beabealashvilli,c Chris J. Hamilton,d
Stanley M. Roberts*e and Alexander Krayevskya
a Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, 32 Vavilov Str.,
Moscow, 117984, Russia. E-mail: aak@imb.imb.ac.ru; Tel: ϩ7 095 135 22 55;
Fax: ϩ7 095 135 14 05
b University of Oslo, Centre for Medical Studies, Moscow, Russia
c Institute of Experimental Cardiology, National Cardiology Research Centre,
15A Cherepkovskaya Str., Moscow, 121522, Russia
d Chemistry Building, University of Exeter, Stocker Rd, Exeter, UK EX4 4QD
e Department of Chemistry, University of Liverpool, PO Box 147, Liverpool, UK L69 7ZD.
E-mail: sj11@Liverpool.ac.uk; Tel: ϩ44 151 794 3501; Fax: ϩ44 151 794 3587
Received (in Glasgow) 16th December 1998, Accepted 14th January 1999
A series of new nucleoside 5Ј-triphosphate mimetics, 2, 3, 5, 6, 8–10, modified at the glycone and all three phosphate
residues, have been synthesised and studied. These compounds only bear the enzymatically labile anhydride bond
between the α and β phosphorus atoms. The preparative chemistry involved the preparation of phosphonic salts
30, 31 and 32 and coupling of these species to the morpholidate 33. The mechanism of formation of some of the
intermediates ‘en route’ to 27 and 28 is discussed. All of the target compounds demonstrated high stability in
human blood serum with half lives towards hydrolysis of up to 4.5 days. Some of these nucleoside triphosphonates
have been shown to be selective inhibitors of DNA synthesis catalysed by retroviral reverse transcriptases and
terminal deoxynucleotidyl transferases. They inhibited replication of the artificial virus containing Moloney murine
leukemia virus reverse transcriptase in infected cell culture, probably due to the inhibition of a reverse transcription
step of a genomic RNA. Compared to the triphosphonates, the corresponding monophosphonates demonstrated
decreased antiviral activity by 1–2 orders of magnitude. This implies that the triphosphonates inhibit virus replication
directly, rather than by a two-step mechanism based on their hydrolysis to the monophosphonates and subsequent
intracellular diphosphorylation. Being totally independent of the enzymatic phosphorylation pathways of the host
cell, the compounds under study may also be able to inhibit retrovirus reproduction both in kinase deficient cell lines
and in the intercellular blood media.
modified and unmodified dNTP analogues enhances their
selectivity whilst preserving their substrate properties towards
Introduction
The study of DNA and RNA metabolism often employs
deoxynucleoside 5Ј-triphosphates (dNTP), ribonucleoside 5Ј-
triphosphates (rNTP) and their analogues as substrates or
inhibitors. However, the usage of these compounds in cell
biology is limited due to their rapid dephosphorylation in
intercellular media and during their diffusion into cells. For
the same reason, modified dNTP and rNTP cannot be used
as antiviral agents, hence the need for more stable derivatives.
The design of dNTP analogues with increased stability
in vivo could also produce a new group of highly effective
inhibitors of HIV reproduction. The advantages of such
compounds would be as follows: (i) Independence of the
nucleoside phosphorylation pathways, catalysed by intra-
cellular kinases, and direct inhibition of proviral DNA syn-
thesis. (ii) The lack of the cell cycle effect on their activity. (iii)
The ability to inhibit reverse transcription in blood plasma or
seminal fluids [in the case of human immunodeficiency virus
(HIV)].1 It is also desirable that such dNTP/rNTP analogues
act as substrates only towards viral enzymes and be hydro-
phobic enough to penetrate into the cell or be able to bind to
membrane proteins which could facilitate their transportation
into cells.
the reverse transcriptases of HIV (HIV-rt) and avian
myeloblastosis virus (AMV-rt).2 Similar results were obtained
when the α-phosphate residue in dNTPs were substituted by
the methylphosphonate, phenylphosphonate,3,4 or α-methyl-
phosphonate units,5,6 the β,γ-pyrophosphate residue by di-
phosphonate units7–11 and all the three phosphate groups
by phosphonates.12,13 Some of these modified dNTP anal-
ogues displayed enhanced stability towards dephosphorylating
enzymes in human blood serum as well as towards alkaline
phosphatase.2,12,13 They also showed a significant degree of
hydrophobicity.2 These properties suggest that such com-
pounds have promising potential as substrates or inhibitors of
polymerase-dependent DNA synthesis in cell cultures. We
present herein the preparation of the carbocyclic -dideoxy-
NTP analogues (ddNTP) 1–10 together with the -stereoisomer
of compound 4, i.e. (ent)-4, and the biological evaluation
of these compounds in cell free systems with HIV-rt,
AMV-rt, terminal deoxynucleotidyl transferase (TDT) and
DNA polymerases α and β. The efficacy of these compounds as
inhibitors of viral replication in Rat1 cell culture infected by
artificial retrovirus containing Moloney murine leukemia virus
(Mu-MLV) and their stability in human blood serum is also
reported. The monophosphonates 11 and 12 were used as
reference compounds in the experiments with cell cultures.
We recently reported that the replacement of the γ-phosphate
moiety by a methyl- or phenylphosphonate unit in both ring
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050
1039

View Online
Table 1 Study of the effects of phenyl and methyl functionalised
phosphonochloridates on product distribution when reacting with 21
O
P
O
P
O
P
B
R
O
O
X
O
P
O
P
O
P
O
P
HO
OH OH
F
F
THF
– 78 oC
R
EtO
EtO
OR'
Cl
EtO
EtO
+
1
2
3
4
5
6
7
8
9
B=Adenine, X=O, R=OH
B=Adenine, X=CBr2, R=OH
B=Adenine, X=CF2, R=OH
B=Guanine, X=O, R=OH
B=Guanine, X=CBr2, R=OH
B=Guanine, X=CF2, R=OH
B=Adenine, X=O, R=Ph
CF2Li
R
OR'
30 min
R=Ph, R'=iPr 13
R=Me, R'=iPr 14
21
B=Adenine, X=CH2, R=Me
B=Guanine, X=CF2, R=Me
Electrophile
10 B=Guanine, X=CF2, R=Ph
Reaction products
13a
14a
O
P
O
N
O
HO
HO
B
O
17 (8%)
17 (45%)
N
N
P
EtO
NH
O
O
P
O
P
CF2H
EtO
P
HO
HO
O
NH2
11 B=Adenine
12 B=Guanine
O
O
OH OH
O
P
O
P
F
F
16 (37%)
15 (53%)
(ent)-4
R
EtO
EtO
OR'
Results
O
O
Chemistry: Synthesis and mechanistic considerations
F
F
18 (30%)
19 (12%)
20 (13%)
18 (2%)
P
P
The syntheses of compounds 1,4 2,13 4,5 5,13 6,15 1114 and 125
have been reported elsewhere. The major new elements of
chemistry investigated in connection with this paper involved
the preparation of the requisite phosphorylphosphonate
moieties which were to be appended to the morpholidate
derivative of the known carbocyclic monophosphonate 12.
The phosphorylphosphonate triesters 15 and 16 were key
intermediates and these compounds were prepared from the
chlorophosphates 13 and 14, respectively, as outlined in Scheme
1. Thus treatment of diethyl difluoromethylphosphonate 17
EtO
EtO
OEt
OEt
O
P
O
P
F
F
b
R
R
—
R′O
OR'
O
b
P
R
—
CF2H
R′O
O
P
O
P
a Product ratios as determined by 19F NMR. b Not detectable.
(i)
Cl
Cl
OiPr
Cl
R
R
ture. In addition to the desired phenylphosphoryl phosphonate
16, a number of other by-products were also isolated and
characterised. These were the bisphosphonate 18, the difluoro-
methylenebis[isopropyl(phenyl)phosphonate] 19 and the (di-
fluoromethyl)phenylphosphonate 20; a small amount of the
starting material 17 was also recovered (Scheme 1). A detailed
study of these two reactions was carried out, by 19F NMR
analysis of the crude reaction mixtures, in order to compare
the influence of the methyl/phenyl substituent on the product
ratios. The results are shown in Table 1.
R = Me, Ph
R = Ph
R = Me
13
14
O
P
O
P
O
O
P
O
P
F
F
F
F
(ii)
P
Me
EtO
EtO
EtO
EtO
EtO
EtO
OEt
OEt
+
CF2H
17
OiPr
(iii)
15
18
(ii)
(iv)
The most notable difference between the reactions of 21 with
either methyl or phenyl phosphonochloridates is the extent of
by-product formation. When the phenylphosphonochloridate
13 is used, significant by-product formation is observed but in
the same reaction with the methylphosphonochloridate 14, the
degree of side product formation is much lower. In the former
transformation there are side-reactions taking place which
account for the poorer yields of the desired product 16
(Scheme 2).
O
P
R2
O
O
F
F
R1
P
R4
R3
P
Ph
i
+
CF2H
PrO
20
R1=R2=OEt, R3=Ph, R4=OiPr
R1=R2=R3=R4=OEt
16
18
Thus initially, nucleophilic substitution of 21 onto the
phosphonochloridate 13 gives the desired product 16, step (i).
Following initial product formation, some of the unreacted
anion 21 can further react with 16 at Pα to give the bisphosphon-
ate 18 and the phosphoryl carbanion 22 which is evidently a
good leaving group in the presence of such hard nucleophiles,
step (ii). The phosphoryl carbanion 22 can also react with some
of the unconsumed phosphonochloridate 13 to give 19 as a 1:1
mixture of the racemic and meso products (by 19F NMR), step
(iii). Unreacted carbanions, 21 and 22 may be protonated (to
form 17 and 20) following an aqueous work up. Reactions were
carried out at Ϫ78 ЊC in order to avoid the degradation of 21
via the loss of difluorocarbene.
R1=R3=Ph, R2=R4=OiPr
19
i
Scheme 1 (i) PrOH (1 equiv.), Et3N, PhMe, 0–20 ЊC, 90 min; (ii)
LDA, THF, Ϫ78 ЊC, 30 min; (iii) (14), THF, Ϫ78 ЊC, 90 min; (iv) (13),
THF, Ϫ78 ЊC, 30 min.
with LDA followed by the addition of isopropyl methylphos-
phonochloridate 14 gave the phosphorylphosphonate triester
15 in low yields. A small amount of tetraethyl difluoromethyl-
enediphosphonate 18 was isolated as a by-product.16
Treatment of the phenylchlorophosphonate 13 under the
same reaction conditions gave a more complex reaction mix-
1040
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050

View Online
O
P
O
P
O
P
i PrO
O
P
O
P
O
O
P
F
(i)
Nu–
– HCl
Ph
OiPr
Cl
OEt
OEt
–
–
EtO
EtO
P
Z –
OiPr
24
Nu
+
HZ
CF2
Ph
Z
O
iPr
Nu
Cl
i PrO
F
23
21
13
14
O
P
16
Methylene-
oxophosphorane
–
OEt
OEt
F2C
Z=CH2
O
P
Nu–= 21
+
NuH
21
–
•
(ii)
+
CF2
•
EtO
EtO
Scheme 3
unreacted phosphonochloridate 14 was isolated crude (from the
aqueous layer during work-up) as the deuteriated phosphonic
acid 26 (Scheme 4).
O
P
OiPr
Cl
Ph
13
O
P
OiPr
O
P
O
Thus one possibility, considered initially, was that the form-
ation of 23 is energetically disfavoured and simple deproton-
ation of the methyl phosphonochloridate 14 forms the anion
25. This process would be reversible but if the equilibrium
should favour the conjugate acid 17 only a low concentration of
21 would be maintained in situ (Scheme 4). Hence 17 would be
isolated in its protonated form following a deuterium quench.
Likewise, if 21 can only react, in this manner, through the AE
pathway then the isolated product 15 would also be fully pro-
tonated and the only deuteriated product would be the methyl-
phosphonic acid 26, as is observed.
F
F
Ph
i PrO
P
Ph
OiPr
Ph
–
F2C
(iii)
22
19
+
F
O
P
O
P
EtO
EtO
OEt
OEt
F
18
This mechanistic proposition certainly fits in with the results
of the deuteriation study but there are several reasons why one
should seek an alternative explanation. Firstly, if this was the
predominant mechanism then it would be reasonable to expect
the reaction to be driven further towards completion as none
of 21 is irreversibly consumed unless it reacts as a nucleophile
with 14 to give the substitution product 15. This is not what is
observed as half of 21 is still unconsumed after 30 minutes.
Secondly, it is also unlikely that the acid–base equilibria of 21
with its conjugate acid 17 would strongly favour the formation
of the latter in the presence of 14 as the α-methyl group of 14
will be significantly less acidic than the difluoromethyl group of
17. This means that the equilibrium (Scheme 4) will favour the
presence of 21 and 14 and, on quenching with deuterium oxide,
most of the unconsumed nucleophile 21 would be isolated as
the deuteriated species and the phosphonomethyl group of any
unconsumed electrophile 14 would be undeuteriated, contrary
to our observations. Therefore, notwithstanding the lack of
literature precedent, we were forced to the conclusion that 23
was indeed a viable intermediate. Hence we believe it is feasible
that formation of 23 from 14 and reaction with 21 furnishes 24
which deprotonates unreacted 14 to generate the fully proton-
ated product and additional 23; hydrolysis of residual 23 with
deuterium oxide would yield the deuteriated phosphonic acid
26 as observed.
Although it is difficult to deduce the exact mechanistic details
of this reaction, from the information available, it is possible
that the elimination of HCl from the phosphonochloridate 14
leads to a reduction in the proportion of reactive anion 21
which is present in the reaction mixture at any one time.
Hence only small traces of by-products are evident from the
reaction with the methylphosphonochloridate 14. In contrast,
the phenylphosphonochloridate 13 has no acidic protons and
hence, as stated earlier, the reaction can only proceed by the
simple AE mechanism (Scheme 2). When the first traces of
the initial condensation product 16 are formed, there will be a
large excess of unreacted nucleophile available to initiate the
side product formation that is observed.
Scheme 2
As by-product formation is initiated by further reaction of
the phosphorylphosphonate product with unconsumed nucleo-
phile 21 at Pα, the increased susceptibility of the phenylphos-
phorylphosphonate 16 to nucleophilic attack, by 21, could
account for the greater proportion of by-products arising from
the reaction of 21 with 13.17 Alternatively, the differences in
the reaction profiles for the methyl- and phenyl-phosphono-
chloridates may be due to significant differences in the mechan-
isms of their reactions with nucleophiles such as 21. Thus the
reaction of 21 with 13 probably proceeds via an associative
addition–elimination (AE) mechanism whereby nucleophilic
displacement proceeds through a trigonal bipyramidal (TBP)
intermediate.18 The same mechanism is also possible for the
reactions involving the methylphosphonochloridate 14 but the
presence of the relatively acidic α-methyl protons allows the
possibility of an alternative elimination–addition (EA) reaction
pathway. This unimolecular process proceeds through a high
energy three-coordinate P() intermediate; the formation of
which is strongly dependent on the acidity of ligand HZ—
relative to the basicity versus nucleophilicity of the attacking
nucleophile. The generation of such phosphorane type inter-
mediates in nucleophilic reactions at phosphoryl centres is well
documented for substrates with strongly acidic HZ-ligands
where Z = oxygen,19 sulfur,20 or nitrogen21 but in nucleophilic
substitution reactions there is very little evidence for the form-
ation of methylene(oxo)phosphorane intermediates such as 23
where Z is an unsaturated carbon atom.22 If the nucleophile 21
is basic enough to deprotonate 14, eliminating HCl, to give the
methylene(oxo)phosphorane 23 then any unreacted carbanion
21 could then attack this electrophilic intermediate to give the
product anion 24 (Scheme 3).
To differentiate between the two possible mechanisms, the
reaction of 21 with 14 was repeated and quenched, after 30
minutes, with deuterium oxide. If the reaction proceeds through
the AE mechanism then any unreacted carbanion 21 would be
deuteriated upon quenching. Conversely, if the EA mechanism
predominates then the product would be mono-deuteriated and
the recovered starting material 17 protonated (the in situ proton
source for 21 being the deprotonation of 14). Partial deuteri-
ation of both 17 and 15 would indicate that both reaction
pathways were operating. The results of this experiment sur-
prisingly showed both 17 and 15 to be fully protonated and the
Tetraethyl difluoromethylenediphosphonate 18 was prepared
from 17 as shown in Scheme 5.23 Deprotection of the phos-
phoric/phosphonic esters 15, 16 and 18 was accomplished with
bromotrimethylsilane to give the free acids 27, 28 and 29
(Scheme 5) which were subsequently converted to their respect-
ive tributylammonium salts 30, 31 and 32.
The final coupling steps for the preparation of the new NTP
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050
1041

View Online
O
P
O
P
O
P
O –
P
H3C OiPr
O
P
O
P
F
CF2
Cl
EtO
EtO
H3C
–
EtO
EtO
EtO
EtO
CH3
OiPr
CF2
Cl
i PrO
+
F
21
14
15
25
14
21
O
P
O
P
O
P
O
P
O
– Cl –
F
– H2C
–
P
EtO
EtO
CH2
OiPr
EtO
EtO
OiPr
Cl
CF2H
+
i PrO
F
23
25
17
24
D2O
D2O
O
D
P
OD
i PrO
26
Scheme 4
Table 2 Semiquantitative data on the terminating activity of the newly
synthesised compounds towards different DNA polymerases
O
O
P
O
P
(i)
F
EtO
EtO
P
OEt
OEt
EtO
EtO
CF2H
Enzymesa
F
17
18 (34%)
Compound
HIV-rt
<0.002
AMV-rt
<0.02
<2.0
<0.2
<0.02
<0.002
<2.0
<0.2
<0.02
<2.0
α
β
TDT
<0.2
<0.02
<0.2
<0.02
<0.02
<0.2
O
P
O
P
O
P
O
O
P
O
1
>600
>600
>600
>200
>600
>600
>600
>200
>200
>200
<200
>600
>600
<20
<200
>600
>600
<2
(ii)
(iv)
F
F
F
F
F
F
EtO
EtO
R
HO
HO
P
R
HO
HO
P
R
2
>600
(iii)
3
<0.02
<0.02
<0.002
>2.0
<0.02
<0.02
OR'
OH
OH
•NBu3
ddATP
4
R=Me, R'=iPr 15
R=Me
R=Ph
R=OH
30
31
32
R=Me
R=Ph
R=OH
27 (99%)
28 (91%)
29 (83%)
5
R=Ph, R'=iPr
16
6
<0.2
<0.02
>2.0
18
R=OEt, R'=Et
ddGTP
7
>200
>200
Scheme 5 (i) LDA, diethyl chlorophosphate, THF, Ϫ78 ЊC, 1 h; (ii)
TMSBr, 3 d; (iii) H2O; (iv) NBu3, EtOH(aq)
8
>2.0
<2.0
<2.0
<0.02
>2.0
.
9
<0.2
<0.02
<0.02
10
ent-4
>2.0
>2.0
analogues are summarised in Scheme 6. The morpholidate 33
was prepared from compound 12 in standard fashion5 and
coupled to 30 to give the NTP analogue 9 which was isolated as
its ammonium salt. Following a recent report of improved reac-
tion yields being achieved when using 1H-tetrazole as a catalyst
in pyrophosphate coupling reactions between 5Ј-monophos-
phomorpholidates and sugar-1-phosphates,24 this procedure
was employed in the synthesis of 10. Disappointingly the yield
of 10 was still modest. Compounds 3 and 7 were prepared,
using the carbonyl diimidazole coupling reagent, in similar
yields. The Pγ-substituted pseudo-triphosphate 8 was prepared
from 11, via the phosphorylphosphonate 34, using a 1,2,4-
triazole mediated coupling procedure.
a Concentration required (µM) for 50% incorporation of the terminat-
ing substrate in the template–primer complex.
DNA replication and repairing, respectively. TDT is also of
interest as a unique example of a template independent DNA
polymerase. Table 2, and the complementary data in Figs. 1–3,
summarise the terminating activity of the compounds under
study, based on the analysis of electrophoretic data. It can be
seen that most compounds did not terminate DNA synthesis
catalysed by DNA polymerases α and β but were active in the
case of TDT, though this was one order of magnitude weaker
than their terminating activity with HIV-rt.
The termination pattern of substrates 1–3 in the HIV-rt-
catalysed elongation of the tetradecadeoxynucleotide primer–
RNA template complex (complex 4) was assessed. As can be
seen, the effectivity of the pyrophosphoryl phosphonate 1 was
comparable with that of dATP and could be observed at
concentrations as low as 0.002 µM. The introduction of a
Pβ,Pγ-difluoromethylenediphosphonate linkage, in compound
Biological evaluations
All the compounds were evaluated as terminating substrates for
HIV-rt, AMV-rt, human placenta DNA polymerases α and β
and calf thymus TDT. HIV-rt was chosen because of its con-
tinued importance, AMV-rt was under investigation as a second
example of this group of enzymes and human DNA poly-
merases α and β are known as the most significant enzymes of
1042
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050

View Online
O
P
O
P
O
P
O
P
HO
F
F
O
(i)
Gua
R
O
Gua
O
N
HO
OH OH
O
(ii)
R=OH
R=Me
R=Ph
6 (24%)
9 (46%)
10 (36%)
33
O
O
O
O
HO
HO
Ade
P
O
(iii)
P
R
P
P
Ade
O
O
X
HO
OH OH
11
R=OH, X=CF2, 3 (52%)
R=Me, X=CH2 8 (35%)
(iv)
O
P
O
P
O
P
O
P
O
P
Ade
Ph
HO
O
Ade
(v)
O
O
O
O
HO
OH OH
OH OH
34 (16%)
7 (56%)
Scheme 6 (i) 30 or 32, DMSO 5–13 h room temp.; (ii) 31, DMSO, 1H-tetrazole (2.0 equiv.), 14 h at room temp.; (iii) N,NЈ-carbonyldiimidazole,
DMF, 12 h, then add 32 or (methylphosphinyl)methylene phosphonate, 12 h at 37 ЊC; (iv) (Me)2NCH(OMe)2, DMF, 20 h, then N,NЈ-
carbonyldiimidazole, DMF, 12 h, then H3PO4ؒNBu3, 3 h at room temp.; (v) 1,2,4-triazole, phenylphosphonic dichloride, DMF, 5 min, 3 ЊC.
Fig. 2 Primer extension catalysed by AMV-rt in the presence of 6:
Lane 1, template–primer (complex 1) ϩ enzyme; lane 2, template-
primer (complex 1) ϩ enzyme ϩ dTTP (2 µM); lanes 3–5, as lane
2 ϩ ddGTP [0.02 µM, (3)], [0.2 µM, (4)] and [2 µM, (5)]; lanes 6–8,
as lane 2 ϩ 6 [0.02 µM, (6)], [0.2 µM, (7)] and [2 µM, (8)].
Fig. 1 Primer extension catalysed by AMV-rt in the presence of 4, 5
and 2: Lane 1, template–primer (complex 1) ϩ enzyme; lane 2,
template–primer (complex 1) ϩ enzyme ϩ dTTP (2 µM); lane 3, as lane
2 ϩ dGTP (2 µM); lanes 4–6, as lane 2 ϩ 4 [0.01 µM, (4)], [0.1 µM,
(5)] and [1 µM (6)]; lanes 7–9, as lane 2 ϩ 5 [0.2 µM, (7)], [2 µM, (8)]
and [20 µM, (9)]; lane 10, as lane 2 ϩ dGTP and dATP (2 µM each);
lanes 11–13, as lane 2 ϩ dGTP (2 µM) ϩ 2 [0.2 µM, (11)], [2 µM, (12)]
and [20 µM, (13)].
3, reduced the activity with HIV-rt by at least an order of
magnitude when compared with the parent pyrophosphoryl
phosphonate 1. The dibromomethylene analogue 2 was not
utilised when either the RNA template (complex 4) or the DNA
template (complex 2, data not shown) were used, even at con-
centrations of 600 µM. Similar results were obtained for com-
pounds 4–6 in chain elongation experiments (with complex 1)
catalysed by HIV-rt. The activity of the γ-methyl analogue 9
(with HIV-rt and complex 1) was approximately one tenth that
of 10 and triphosphonates 6 and 10 were 10–50 times less active
than their parent pyrophosphoryl phosphonate 4. The relative
efficacies of compounds 4, 6 and 10, determined from these
experiments, are in close agreement with their IC50 values
(relative to AZTTP) which have been determined in an
independent colourimetric assay (Table 3).25
Figs. 1 and 2 demonstrate the terminating substrate prop-
erties of compounds 2 and 4–6 in the process of AMV-rt
catalysed DNA synthesis. Incorporation of triphosphonates 2
and 5 at the 3Ј-position of the primer was 2000-fold weaker
than that of 4 (compare lanes 13 and 9 with lane 4, Fig. 1). The
incorporation of ATP into the 3Ј-terminus of the growing
primer chain (complex 1) in the presence of 2 was inhibited by
Fig. 3 The effect of 2 and 3 on the dephosphorylation rate of 7 in
human blood serum. Curves: (1) 7 (0.5 mM); (2) 7 (2.0 mM); (3) 7
(0.33 mM) ϩ 3 (1.65 mM); (4) 7 (0.33 mM) ϩ 2 (1.65 mM).
not more than 10% when the molar ratio of 2 to ATP was 100:1
(data not shown). It is evident from Fig. 2 that the utilisation
of 6 (lanes 7, 8) is more than 10 times less effective than that of
ddATP (lanes 3, 4). The difluoromethylenediphosphonate
derivative 6 proved to be at least 200 times weaker as a terminat-
ing substrate than its pyrophosphate counterpart 4 (compare
lanes 6–8, Fig. 2, and lanes 4–6, Fig. 1). Triphosphonates 7, 9
and 10 showed weak terminating substrate properties with
AMV-rt and compound 8 was even less effectively utilised by
this enzyme (Table 2).
In the experiments with DNA polymerase β, the ddGTP
analogue 4 was a rather poor substrate for this enzyme and the
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050
1043

View Online
dihalomethylenediphosphonate derivatives 5 and 6 showed no
substrate properties. Similar results were obtained for the aden-
ine derivatives 1–3. None of the analogues 1–6 were utilised by
DNA polymerase α up to concentrations of 600 µM (Table 2).
Since DNA polymerases α and β did not recognise com-
pounds 6–8, we did not evaluate the α,β,γ-phosphate modified
analogues 9 and 10 with these enzymes.
phonates 9 and 10, bearing an additional modification at the
γ-phosphate residue, were only 20 times greater than that of
dGTP. As illustrated by the reversed-phase-HPLC retention
times (Table 5), all of the test compounds showed a marked
increase in hydrophobicity when compared to the ‘natural’
dNTP. This is particularly significant for the Pγ-phenyl substi-
tuted derivatives 7 and 10. The HPLC retention time of 10
lies between those of the monophosphonate and the pseudo-
triphosphate (ent)-4.
Fig. 3 illustrates the enzymatic stability of diphosphonate
phosphate 7 in human blood serum at two concentrations and
in the presence of triphosphonates 2 and 3. The dephosphoryl-
ation rate of 7 decreases by seven-fold in the presence of 2 and
four-fold in the presence of 3 (compare curve 4 with curves 1
and 2, respectively); in these experiments the molar ratio of
either 2 or 3 to 7 is 5:1. Increasing the concentration of 7 to 2
mM (which corresponds to the combined concentrations of 7
and 2/3 in the experiments in the presence of triphosphonates)
decreases the hydrolysis rate of 7 only by two-fold (curve 3).
When the concentration of 2 was about 20 times higher than
that of 7 the extent of hydrolysis of the latter was not more than
5% after 3.5 h (data not shown).
The data on the inhibition of pSG1 virus replication in Rat 1
fibroblasts are presented in Table 6. Monophosphonates 11 and
12 were used as controls and AZT, FLT and d4T were used as
reference standards. Both the triphosphonates 2 and 6 inhibited
virus reproduction with IC50 values of 0.12 µM and 12.67 µM
and the values for compounds 11 and 12 were 7.46 µM and
>100 µM, respectively. Triphosphonates 9 and 10 were inactive
up to the maximum concentrations tested (100 µM).
To evaluate the effects of the compounds on transcription of
the lacZ gene compounds 2, 6, 11 and 12, at concentrations
of 100 and 200 µM, were added to Rat1 cells, infected by pSG1
virus 24 h earlier. By this time reverse transcription was com-
The results of the experiments with TDT demonstrate
that the replacement of the β,γ-phosphate by a dibromo- or
difluoromethylenediphosphonate unit decreases the terminat-
ing substrate properties by about 10-fold as compared with
ddATP and ddGTP (Table 2). Modified triphosphonates 7, 8
and 10 are more than 100 times less effective as terminating
substrates for TDT than the triphosphonates 2, 3, 5 and 6. The
terminating substrate properties of the pyrophosphoryl phos-
phonate 1 and its -stereoisomer under catalysis by reverse
transcriptases and TDT have been demonstrated previously.26
The kinetic parameters for compounds 1–6 in the reactions
catalysed by HIV-rt and AMV-rt are given in Table 4. The Km
values for 2 and 5 in the reaction catalysed by HIV-rt are not
presented since their substrate activity was not evident, even at
concentrations of 600 µM (Table 2). A comparison of the Km
values for 3 and 6 relative to 1 and 4, respectively, demonstrates
the 8–10-fold decrease in the affinity of the former compounds
for HIV-rt. It should be noted that the Km values for com-
pounds 1 and 4 were 2–3 times higher than those of ddATP and
ddGTP, respectively, whereas their reaction rates were close to
Vmax of their respective triphosphonate derivatives, 2, 3 and
5, 6. In the case of AMV-rt the Km and Vmax values of the
compounds under investigation were close to those obtained for
the reactions catalysed by HIV-rt and compounds 2 and 5 also
revealed terminating substrate properties towards AMV-rt.
Table 5 contains data on the enzymatic stability of the test
compounds in human blood serum. Blood serum is an
appropriate medium in which to perform this assay as it con-
tains numerous dephosphorylating enzymes and so provides a
good model system of the extracellular environment in vivo.
The half-lives of 3, 5 and 6 were shown to be 70–100 times
higher than those of the corresponding dNTP; the value for
2 was 200-fold greater. Surprisingly, the half-lives of triphos-
Table 5 Half-life of test compounds in blood serum and retention
time on reversed-phase HPLC
Compound
Half-life time
Retention time/min
2
110 h
30 h
31 h
45 h
50 min
10 h
17.7
17.5
16.4
16.3
22.3
15.8
19.5
16.5
12.6
11.6
3
Table 3 The relative efficacy of compounds 4, 6 and 10 as inhibitors
5
of HIV-rta
6
7
Compound
IC50/µM
9
10
12 h
4
1.3
16.5
49.4
1.65
(ent)-4
dATP
dGTP
dTTP
11
65 min
<30 min
<30 min
5 mina2
6
10
AZTTP
23.4
21.9
a Testing was carried out using the DuPont RT-DetectTM Reverse Tran-
scriptase Assay (DuPont, NEK-070A) and recombinant HIV-1-rt
(Dupont, NEI-490).
12
a Conditions were identical to those used above.
Table 4 Kinetic constants for 1–6 in the one-step primer extension reaction catalysed by HIV-rt and AMV-rt
HIV-rt
AMV-rt
Compound
Km/µM
Vmax/Vmax(ddNTP)
Km/µM
Vmax/Vmax(ddNTP)
ddATPa
0.105 0.001
0.008 0.0008
1
1.15
—
1.05
1
1.10
—
0.095 0.023
0.026 0.004
0.537 0.075
0.091 0.014
0.0139 0.0016
0.0077 0.0001
0.1831 0.059
0.0249 0.0026
0.0377 0.0002
1
1a
1.52
0.41
1.01
1
1.50
0.27
1.01
1.62
2a
—
3a
0.016 0.001
ddGTPb
4b
0.0517 0.0048
0.00107 0.0002
—
0.0078 0.0001
0.0147 0.0001
5b
6b
1.02
1.20
ent-4b
a With the template–primer complex 2. b With the template–primer complex 3.
1044
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050

View Online
Table 6 Inhibition of Mu-MLV by modified nucleoside triphosphate
hydrophobicity of the triphosphonate 10 (by HPLC, Table 5)
analogues
makes it a promising candidate for further development as an
antiviral agent. Although compound 10 proved to be a good
inhibitor of retroviral DNA polymerases in cell free assays, its
activity as an inhibitor of pSG1 virus replication in cell culture
was less significant. Thus further modification of this com-
pound is warranted in order to increase its lipophilicity, to
further facilitate cell penetration, without compromising its
potency as a terminating substrate.
The stability studies of the triphosphonates 2, 3, 5 and 6 in
human blood serum testifies that, in these compounds, the
phosphate bond between Pα and Pβ is essentially more stable
towards enzymatic hydrolysis than that of the natural tri-
phosphates. This could be attributed to these analogues being
poorer substrates for those dephosphorylating enzymes which
function to hydrolyse NTPs at this position. The markedly
increased stability of 2 is noteworthy. The half-life of the
diphosphonate phosphate 7, bearing the β-phosphate group, is
close to that of dATP. A decrease in the rate of hydrolysis of 7
in the presence of either 2 or 3 indicates that 2 and 3 may be
bound by these enzymes as a result of a catalytic process or by
specific adsorption.
Inhibition of pSG1 virus replication in Rat1 cells by the
compounds under study is clearly evident. Triphosphonate 2 is
about 30 times more potent than the corresponding mono-
phosphonate 11. Triphosphonate 6 (with the guanine base)
exhibits an effect similar to monophosphonate 11 (with the
adenine base) whilst monophosphonate 12 (with the guanine
base) shows no activity up to concentrations of 100 µM. These
data support the hypothesis of the direct action of these tri-
phosphonates as inhibitors of viral replication in whole cell
systems and indicate that a two-step molecular mechanism of
inhibition, involving hydrolysis of 2 and 6 to the corresponding
monophosphonates 11 and 12 followed by intracellular diphos-
phorylation, is unlikely. The results obtained are promising
with respect to the development of a new type of active inhibi-
tor of virus reproduction which does not require intracellular
phosphorylation. The 50% inhibition of virus reproduction by
2 in this system is 40 times lower (and 6 4000 times lower) than
that of AZT but 7 times higher than that of d4T. This decrease
in activity could be rationalised by one or more of the following
reasons: (i) Cell membrane permeability is insufficient for the
penetration of 2 and 6 into the cell. (ii) Partial dephosphoryl-
ation of the triphosphonates during the process of diffusion
across the cell membrane may take place. (iii) The triphos-
phonates may show decreased affinity towards retroviral reverse
transcriptases compared to AZTTP.15 We do not know at pres-
ent whether modified triphosphonates inhibit reverse transcrip-
tion of the virus genome in the intercellular media, prior to the
virus penetrating the cell (as has been shown for AZTTP),1 or
whether they do so in the infected cells during the process of
ordinary reverse transcription. To answer this question will
require another model system since the pSG1 virus loses the
infectivity after 20 h of incubation with dNTP at 37 ЊC.30
Compound
IC50/µM
5
0.12 0.05
12.67 1.86
>10
>100
>100
7.46 1.16
>100
0.003 0.0003
0.011 0.001
0.845 0.101
6
8
9
10
11
12
AZTa
FLTb
d4Tc
a AZT = 3Ј-azido-3Ј-deoxythymidine. b FLT = 3Ј-fluoro-3Ј-deoxythym-
idine. c d4T = 2Ј,3Ј-dideoxy-2Ј,3Ј-didehydrothymidine.
pleted (half-life of reverse transcription for pSG1 was deter-
mined as 6 h when 0.1 µM AZT was used to inhibit this
process).27 After an additional incubation for 24 h the cells
were stained with Xgal. The number of blue colonies did not
differ from the control at test compound concentrations of
up to 200 µM indicating that transcription of lacZ gene was
not altered by these compounds. Cytotoxic effects were not
observed with any of the above compounds at concentrations
up to 1 mM.
Discussion
To date, there have been very few reports on the properties of
natural and modified nucleoside triphosphonates. In 1995
AZT-α,β,γ-(bismethylene)triphosphonate was synthesised and
shown to inhibit HIV-rt catalysed oligodeoxythymidine syn-
thesis 1000 times less efficiently than AZTTP.28 This can be
ascribed to the fact that the compounds bearing the hydrol-
ytically stable α,β-methylenediphosphonate linkage are unable
to be incorporated at the 3Ј-terminus of a polydeoxythymidine
chain and thereby terminate further chain elongation.
The experiments reported herein demonstrate the terminat-
ing substrate properties of compounds 1–6 for a range of DNA
polymerases. The introduction of the β,γ-dibromomethylene-
diphoshonate moiety, in compounds 2 and 5, results in a sharp
decrease in affinity to HIV-rt on both DNA and RNA tem-
plates. The difluoromethylene triphosphonate 3 is a more
powerful terminating substrate for AMV-rt than its dibromo-
methylene analogue 2. The affinity of compounds 3 and 6
towards DNA-synthesising complexes is, as a rule, one order of
magnitude lower than those of the natural dNTP (dATP and
dGTP) and the parent pyrophosphoryl phosphonates 1 and 4,
respectively. DNA polymerases α and β do not recognise the
adenine and guanine derivatives 2, 3, 5, 6 and 8 all of which
bear a diphosphonate residue instead of the terminal pyro-
phosphate moiety. In the case of TDT the replacement of the
pyrophosphate unit by either a dibromo- or difluoromethylene-
diphosphonate unit decreases the terminating substrate proper-
ties of the compounds by ten-fold. Compounds 7, 8 and 10 are
more than 100 times weaker than the triphosphonates 2, 3, 5
and 6 which do not possess an alkyl/aryl substituent at Pγ. Thus
triphosphonates 2, 3, 5 and 6 selectively inhibit DNA synthesis
catalysed by AMV-rt and template independent TDT but do
not significantly affect the process catalysed by DNA poly-
merases α and β. The activity of the dibromo-derivatives 2 and
5 towards HIV-rt is 1000-fold lower than that of the parent
pyrophosphoryl phosphonates 1 and 4 whereas the difluoro-
derivatives 3 and 6 only show a ten-fold reduction in activity.
This significant loss of activity, with compounds 2 and 5,
may be a result of greater steric and electronic dissimilarities
between the dibromomethylenediphosphonate group and the
pyrophosphate unit which it replaces; the difluoromethylene
group is evidently a superior bioisostere.29 The enhanced
Conclusion
Taking an overview of the data obtained, it is important to
emphasise the complementary combination of useful properties
of the new compounds. These compounds are examples of a
new generation of dNTP analogues, modified at both the tri-
phosphate side-chain and the glycone, that are able to inhibit
the replication of retroviruses in cell cultures and show high
stability in human blood serum. Not only do these compounds
point the way to potential antiviral chemotherapeutics but also
the use of such compounds in cell biology studies is worthy of
consideration. For these purposes, slowly dephosphorylating
substrates (or terminating substrates) must bear a ligand or
a reporter group in the nucleic base or another part of the
molecule. The synthesis and biological evaluation (in cell free
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050
1045

View Online
systems) of the first compounds in this series is already in
νmax(KBr)/cmϪ1 1269 (P᎐O); λmax(H2O)/nm 263 (14800); δH(400
MHz, D2O) 1.96 (1H, m, 5Ј-βH), 2.92 (1H, m, 5Ј-αH), 3.61
᎐
progress.31
(2H, d, J 9.0, CH2P), 4.66 (1H, s, 4Ј-H), 5.39 (1H, m, 1Ј-H),
5.92 (1H, m, 2Ј-H), 6.13 (1H, m, 3Ј-H), 8.16 (1H, s, 8-H),
8.18 (1H, s, 2-H); δP(162 MHz, D2O) Ϫ2.08 (m, Pβ), 6.37 (m,
J 65.3, Pγ), 11.34 (d, J 32.4, Pα); δF(188 MHz, D2O) Ϫ44.1 (dd,
J 95.1, 81.0); m/z (FAB) 523 (MHϩ ϩ NH3), 540 (MHϩ ϩ
2NH3).
Experimental
General
Petroleum ether had bp 40–60 ЊC (referred to as petrol). TLC
was performed on pre-coated silica gel glass plates (Merck silica
gel 60F 254). The plates were visualised using UV light (254
nm), alkaline potassium permanganate, or ethanolic bromo-
cresol green. The TLC assay of reactions involving the highly
polar phosphate and triphosphate analogues were carried
out as above eluting with either 1 M NH4HCO3(aq)–EtOH–
NH4OH(dilute) (20:50:1) or dioxane–ammonium–water
(6:4:5). Flash chromatography was performed over silica
(Merck silica gel 60 40–63 µm). Anion exchange chrom-
atography was performed on a column of Fractogel EMD
DEAE-650 (M) (Merck, Art. 16886) (40 × 100 mm) eluting
with a linear gradient of aqueous ammonium hydrogen carb-
onate at a flow rate of ca. 200 ml hϪ1. The column was con-
nected to a UV detector (254 nm, LKB 2238, Uvicord SII) and
a fraction collector. NMR spectra were recorded on a Bruker
AM300 and a Bruker Avance DRX 400 spectrometer at the
following frequencies: 300 or 400 MHz for 1H NMR, 75 or 100
MHz for 13C NMR, 162 MHz for 31P NMR (H3PO4 external
reference) and for 19F NMR either 376 MHz (C6F6 external
reference) or 188 MHz (CF3CO2H external reference); J values
are given in Hz. Infrared spectra were recorded on a Perkin-
Elmer 881 Fourier Transform Spectrophotometer. UV spectra
were recorded on a UV4-200 Unicam UV–VIS spectrometer.
Mass spectra were obtained on either a Kratos Profile
HV3, Kratos MS 50TC or a VG7070E/DEC VAX 4000.60
spectrometer.
(1ЈS,4ЈR)-9-[4Ј-(Hydroxy{[(difluoro)(phosphono)methyl]-
hydroxyphosphoryloxy}phosphorylmethoxy)cyclopent-2Ј-enyl]-
guanine (tris-ammonium salt) (6). A solution of the morphol-
idate 33 (0.081 mmol) in DMSO (1 cm3) was added dropwise to
a solution of 32 (0.4 mmol) in DMSO (2 cm3) and left to stir at
room temperature for 13 h. The reaction mixture was diluted
with water (200 cm3) before purification by anion exchange
chromatography eluting with a linear gradient of NH4HCO3(aq)
(0–0.4 M, 800 cm3). The product eluted at 0.22 M. Fractions
were combined and coevaporated several times with water
before further purification by reverse-phase chromatography,
eluting with water. Product fractions were lyophilised to give 6
as a fluffy white lyophilate (11 mg, 24%). νmax(KBr)/cmϪ1 1269
(P᎐O); λmax(H2O)/nm 252.0; δH(400 MHz, D2O) 1.89 (1H, dt,
᎐
J 15.6, 4.4, 5Ј-βH), 2.97 (1H, dt, J 14.0, 7.2, 5Ј-αH), 3.84
(2H, d, J 9.2, PCH2), 4.76–4.81 (1H, br, 4Ј-H), 5.0–5.28 (1H,
br, 1Ј-H), 6.07–6.11 (1H, m, 2Ј-H), 6.33–6.38 (1H, m, 3Ј-H),
7.75–7.87 (1H, br, 8-H); δF(376 MHz, D2O) 42.40–42.90 (br,
CF2); δP(162 MHz, D2O) Ϫ1.69 (br, Pβ), 3.90 (br, Pγ), 11.44 (d,
J 31.0, Pα); m/z (ES) 522 (6%, MHϩ), 328 (15, M Ϫ CH2-
F2O4P2).
(1ЈS,4ЈR)-9-[4Ј-(Hydroxy{[(phenyl)hydroxyphosphoryloxy]-
hydroxyphosphoryloxy}phosphorylmethoxy)cyclopent-2Ј-enyl]-
adenine (tris-ammonium salt) (7). a) To a solution of 11 (120 mg,
0.39 mmol) in DMF (15 cm3) was added dimethylformamide
dimethyl acetal (0.2 cm3) and the reaction was kept for 20 h at
20 ЊC. The solvent was removed in vacuo, the residue diluted
in DMF (10 cm3) and N,NЈ-carbonyldiimidazole (137 mg, 1.15
mmol) was added. After stirring for 12 h at 20 ЊC a solution of
the tributylammonium salt of phosphoric acid (0.5 M in DMF,
0.6 cm3) was added; the suspension was stirred for 3 h. The
reaction mixture was quenched with 15% aqueous ammonium
hydroxide (to pH 10) and the mixture was stirred for 1 h. The
product was isolated and purified as described for 3 to give
34 (52 mg, 16%). δP(162 MHz, D2O) Ϫ8.05 (d, J 25.5), 8.8
(d, J 25.5).
Enzymes and DNA. HIV1-rt was purchased from Worthing-
ton Corp. (USA); DNA polymerases α and β were isolated from
human placenta.32,33 AMV-rt and TDT were from Omutninsk
Chemicals (Russia) and Amersham Corp., respectively.
M13mp10 DNA was isolated from the culture medium of the
recipient E. coli K12XL1 strain.34 Heteropolymeric RNA was
synthesised by run-off transcription of SalGI-digested plasmid
pPV-19 with T7 RNA polymerase.35 The plasmid containing a
fragment of pBR322 DNA between the SalGI and SphI sites
was a gift from Dr S. Kochetkov. The oligonucleotide primers
were labelled at the 5Ј-terminus using [γ-32P]ATP (Radioizotop,
Russia) and T4 polynucleotide kinase.34 The DNA (0.5 mM)
was hybridised with 0.75 mM [5Ј-32P]-labelled primer in the fol-
lowing buffers: 10 mM Tris-HCl (pH 8.2), 5 mM MgCl2, 40
mM KCl and 1 mM dithiothreitol (for reverse transcriptases);
10 mM Tris-HCl (pH 7.4), 6 mM MgCl2 and 0.4 mM dithio-
threitol (for DNA polymerase α); 10 mM Tris-HCl (pH 8.5), 5
mM MgCl2 and 1 mM dithiothreitol (for DNA polymerase β).
b) A solution of 1,2,4-triazole (20.7 mg, 0.3 mmol) and tri-
ethylamine (30.3 mg, 0.3 mmol) in acetonitrile (10 cm3) was
cooled to 3 ЊC and phenylphosphonic dichloride (29.3 mg,
0.15 mmol) was added. The reaction mixture was kept for 30
min at 3 ЊC and then centrifuged. The supernatant was added
to a solution of the tributylammonium salt of 34 (0.1 mmol)
in DMF (2 cm3). The reaction mixture was stirred for 5 min
and then evaporated. The residue was quenched with water (30
cm3) and then purified, as described for 3, to afford 7 (29.7
mg, 56%). δP(162 MHz, D2O) Ϫ19.9 (m, Pβ), 7.1 (d, J 24, Pγ),
Methods
(1ЈS,4ЈR)-9-[4Ј-(Hydroxy){[(difluoro)(phosphono)methyl]-
hydroxyphosphoryloxy}phosphorylmethoxy)cyclopent-2Ј-enyl]-
adenine (tris-ammonium salt) (3). To a solution of 11 (156 mg,
0.5 mmol) in absolute DMF (5 cm3) was added N,NЈ-carbonyl-
diimidazole (162 mg, 1.0 mmol) under an inert atmosphere and
the mixture was stirred for 12 h at 20 ЊC. A solution of 32 (0.5
M in DMF, 2.5 cm3) was then added and the reaction mixture
was stirred at 37 ЊC for 12 h. The reaction mixture was diluted
with water (80 cm3) before purification by anion exchange
chromatography eluting with a linear gradient of NH4HCO3(aq)
(0–0.4 M, 600 cm3). Product fractions were combined and
coevaporated several times with water (3 × 10 cm3) and ethanol
(1 × 5 cm3) before further purification by reversed-phase
chromatography, eluting with water. Product fractions were
lyophilised to give 3 as a white amorphous solid (130 mg, 52%).
8.9 (d,
J
26, Pα); m/z (FAB) 549 (MHϩ ϩ NH3), 566
(MHϩ ϩ 2NH3).
(1ЈS,4ЈR)-9-[4Ј-(Hydroxy{[(methylhydroxyphosphoryl)-
methyl]hydroxyphosphoryloxy}phosphorylmethoxy)cyclopent-
2Ј-enyl]adenine (tris-ammonium salt) (8). This was synthesised,
as described for 3, from 11 and the bis(tributylammonium)
salt of (methylphosphoryl)methylenephosphonate (35% yield).
δH(400 MHz, D2O) 1.32 (3H, d, J 14.5, PCH3), 1.86 (1H, m,
5Ј-βH), 2.16 (2H, dd, J 20, 17, PCH2P), 2.98 (1H, m, 5Ј-αH),
3.78 (2H, d, J 8.5, PCH2O), 4.78 (1H, s, 4Ј-H), 5.45 (1H, m,
1Ј-H), 6.10 (1H, m, 2Ј-H), 6.40 (1H, m, 3Ј-H), 8.10 (1H, s, 8-H),
8.20 (1H, s, 2-H); δP(162 MHz, D2O) 9.3 (Pα ϩ Pβ), 35.9 (Pα);
m/z (FAB) 486 (MHϩ ϩ NH3), 503 (MHϩ ϩ 2NH3).
1046
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050

View Online
(1ЈS,4ЈR)-9-[4Ј-(Hydroxy{[(difluoro)(methylhydroxyphos-
phoryl)methyl]hydroxyphosphoryloxy}phosphorylmethoxy)-
cyclopent-2Ј-enyl]guanine (tris-ammonium salt) (9). A solution
of the morpholidate 33 (0.068 mmol) in anhydrous DMSO (1.5
cm3) was added dropwise to a solution of 30 (0.306 mmol) in
DMSO (1.5 cm3). This was left to stir for 5 h until all of 33 was
consumed (by TLC). The reaction mixture was then diluted in
water (400 cm3) before purification by anion exchange chrom-
atography eluting with a linear gradient of NH4HCO3(aq) (0–0.3
M, 800 cm3). The product eluted at 0.19 M. Product fractions
were coevaporated several times with water before finally being
lyophilised to give 9 as a fluffy white lyophilate (18 mg, 46%).
νmax(KBr)/cmϪ1 3600–2600 (br, NH , OH), 1690, 1624 (C᎐N,
24.02 (d, J 4.3, CH3), 74.11 (d, J 7.7, CH), 128.63 (d, J 16.7,
CH), 130.91 (d, J 11.8, CH), 131.38 (d, J 180.7, C), 133.32
(d, J 3.4, CH); δP(162 MHz, CDCl3) 28.16 [Found: (EI) Mϩ,
218.02642. C9H12Cl35O2P requires 218.02635]; m/z 220 (0.3%,
Mϩ), 218 (1.3, Mϩ), 141 (65.7, M–Ph), 77 (100, Phϩ).
Diethyl difluoro[isopropoxy(phenyl)phosphoryl]methylphos-
phonate (16). To a solution of LDA (6.87 mmol) in THF (8 cm3)
at Ϫ78 ЊC was added a solution of diethyl difluoromethylphos-
phonate (1.09 g, 5.79 mmol) in THF (2 cm3) dropwise over 10
min. This gave a yellow solution which was stirred at Ϫ78 ЊC for
30 min. A solution of 13 (1.45 g, 6.63 mmol) in THF (2 cm3),
pre-cooled to Ϫ78 ЊC, was then added via a cannula and was
left to stir for another 30 min at Ϫ78 ЊC. The reaction was
warmed to 0 ЊC before quenching with a saturated solution of
aqueous KH2PO4 (20 cm3). The aqueous layer was then
extracted with ethyl acetate (3 × 25 cm3). The combined
organics were dried (MgSO4) and solvents were then removed
to give a crude oil which was purified by chromatography elut-
ing with EtOAc–petrol (1:1) to give 16 as a colourless oil (364
᎐
2
C᎐O, C᎐C), 1247, 1195 (P᎐O); λmax(H2O)/nm 252.4; δH(400
᎐
᎐
᎐
MHz, D2O) 1.40 (3H, d, J 14.8, CH3P), 1.86 (1H, dt, J 14.4, 4.4,
5Ј-βH), 2.97 (1H, dt, J 14.5, 7.1, 5Ј-αH), 3.83 (2H, d, J 9.1,
PCH2), 4.76–4.80 (1H, m, 4Ј-H), 5.25–5.31 (1H, m, 1Ј-H), 6.07–
6.11 (1H, m, 2Ј-H), 6.36 (1H, dt, J 5.6, 1.9, 3Ј-H), 7.70–8.95
(1H, br, 8-H); δC(100 MHz, D2O) 13.31 (d, J 98.9, PCH3), 37.65
(C-5Ј), 57.38 (C-1Ј), 65.16 (d, J 162.9, PCH2), 84.61 (d, J 11.6,
C-4Ј), 132.81 (C-2Ј), 136.07 (C-3Ј), 153.66, 158.95 (C); δF(376
MHz, D2O) 39.56 (dd, J 85.1, 75.4); δP(162 MHz, D2O) Ϫ4.44
(br, Pβ), 10.57 (d, J 32.1, Pα), 30.96 (br, Pγ); m/z (ES) 520 (100%,
MHϩ), 328 (88, M Ϫ C2H4F2O4P2).
mg, 17%). νmax(neat)/cmϪ1 1435 (Ph–P), 1275 (α-P᎐O), 1256
᎐
(β-P᎐O); δ (300 MHz, CDCl ) 1.28 (3H, dt, J 7.1, 0.5, CH ),
᎐
H
3
3
1.36 (3H, dt, J 7.1, 0.5, CH3), 1.37 (3H, d, J 6.0, CH3), 1.48
(3H, d, J 6.0, CH3), 5.01–5.12 (1H, m, CH), 7.47–7.55 (2H,
m, 2 × ArH), 7.59–7.67 (1H, m, ArH), 7.86–7.97 (2H, m,
2 × ArH); δC(100 MHz, CDCl3) 16.23 (d, J 5.6, CH3), 16.32 (d,
J 5.7, CH3), 23.84, (d, J 5.1, CH3), 24.58 (d, J 3.0, CH3), 64.96
(d, J 6.5, CH2), 65.12 (d, J 6.2, CH2), 73.65 (d, J 6.9, CH),
126.42 (d, J 141, C), 128.40 (d, J 13.6, CH), 133.25 (d, J 9.9,
CH), 133.62 (d, J 2.9, CH); δP(162 MHz, CDCl3) MX of an
ABMX system, 4.80 (ddd, J 88.4, 86.7, 50.3, Pα), 24.10 (ddd,
J 84.6, 74.3, 56.2, Pβ); δF(376 MHz, CDCl3) AB of an ABMX
system, 42.00 (ddd, J 370.3, 88.5, 74.4), 43.86 (dt, J 370.1, 85.8)
[Found: (EI) Mϩ, 370.09240. C14H22F2O5P2 requires 370.09106];
m/z 370 (2.7%, Mϩ), 141 (96.3, PhP(O)OHϩ), 77 (100, C6H5ϩ),
51 (34.8, CF2Hϩ).
(1ЈS,4ЈR)-9-[4Ј-(Hydroxy{[(difluoro)(phenylhydroxyphos-
phoryl)methyl]hydroxyphosphoryloxy}phosphorylmethoxy)-
cyclopent-2Ј-enyl]guanine (tris-ammonium salt) (10). A solution
of the morpholidate 33 (0.044 mmol) in anhydrous DMSO (1.5
cm3) was added dropwise to a solution of 31 (0.250 mmol)
in DMSO (1.5 cm3) followed by 1H-tetrazole (6.2 mg, 0.088
mmol) in DMSO (0.5 cm3). This was left to stir overnight
by which time all of 33 had been consumed (by TLC). The
reaction mixture was then dissolved in water (400 cm3) before
purification by anion exchange chromatography eluting with
a linear gradient of NH4HCO3(aq) (0–0.3 M, 800 cm3). The
product was eluted at 0.21 M. The product fractions were
coevaporated several times with water before finally being
lyophilised to give 10 as a fluffy white lyophilate (10 mg, 36%).
The following side products were also characterised following
isolation from the appropriate fractions of the chromato-
graphic separation described above. Any additional purification
procedures are as detailed below:
νmax(KBr)/cmϪ1 3600–2700 (br, NH , NH ), 1685, 1632 (C᎐N,
᎐
2
3
C᎐O, C᎐C), 1245, 1208 (P᎐O); λmax(H2O)/nm 252.0; δH(400
᎐
᎐
᎐
Isopropyl methylphosphonochloridate (14). This was prepared
from methylphosphonic dichloride (4.87 g, 0.037 mmol) as
described for 13 and was isolated as a colourless oil (2.58 g,
45%). Boiling point 68–74 ЊC (17 mmHg); νmax(neat)/cmϪ1 1315
MHz, D2O) 1.72 (1H, dt, J 14.5, 4.0, 5Ј-βH), 2.85 (1H, dt,
J 14.5, 7.2, 5Ј-αH), 3.65 (2H, d, J 9.0, PCH2), 5.20–5.26 (1H, m,
1Ј-H), 6.01–6.06 (1H, m, 2Ј-H), 6.27 (1H, m, 3Ј-H), 7.36–7.50
(3H, m, 3 × Ar-H), 7.70–7.85 (3H, m, 2 × Ar-H ϩ 8-H); δC(100
MHz, D2O) 37.50 (C-5Ј), 57.28 (C-1Ј), 64.80 (d, J 164.4, PCH2),
84.52 (d, J 12.5, C-4Ј), 128.11 (d, J 12.5, Ar-CH), 132.13 (C-2Ј),
132.85 (d, J 9.6, Ar-CH), 132.93 (Ar-CH), 135.93 (C-3Ј),
154.10, 159.48 (C); δF(376 MHz, D2O) 43.33 (t, J 80.8); δP(162
MHz, D2O) Ϫ4.82 (br, Pβ), 10.86 (d, J 32.9, Pα), 19.59 (m, JP P
(CH –P), 1269 (P᎐O), 1012 (P–OiPr); δ (300 MHz, CDCl )
᎐
3
H
3
1.39 (3H, d, J 6.1, CH3), 1.40 (3H, d, J 6.1, CH3), 1.94 (3H, d,
J 17.4, CH3), 4.87–5.01 (1H, m, CH); δC(100 MHz, CDCl3)
20.69 (d, J 129.9, PCH3), 23.35 (d, J 5.0, CH3), 23.97 (d, J 4.7,
CH3), 73.32 (d, J 8.0, CH); δP(162 MHz, CDCl3) 39.24 [Found:
(EI) Mϩ, 156.01107. C4H10Cl35O2P requires 156.010696]; m/z
159 (27.9%, MHϩ), 157 (89.7, MHϩ).
β
γ
49.1, Pγ); m/z (ES) 582 (100%, MHϩ), 28 (47, M Ϫ C7H6-
F2O4P2), 260 (65).
Diethyl difluoro[isopropoxy(methyl)phosphoryl]methylphos-
phonate (15). To a solution of LDA (3.20 mmol) in THF (4 cm3)
at Ϫ78 ЊC was added a solution of diethyl difluoromethylphos-
phonate 17 (506 mg, 2.69 mmol) in THF (2 cm3) dropwise over
10 min. This was stirred at Ϫ78 ЊC for 30 min. A solution of 14
(485 mg, 3.09 mmol) in THF (2 cm3) was then pre-cooled to
Ϫ78 ЊC before addition via a cannula; the resultant solution was
left to stir for 30 min at Ϫ78 ЊC. The reaction was warmed to
0 ЊC before quenching with a saturated solution of aqueous
KH2PO4 (10 cm3). This mixture was extracted with ethyl acetate
(3 × 15 cm3), the combined organics were dried (MgSO4) and
solvents were then removed to give a crude oil which was puri-
fied by chromatography, eluting with EtOAc–petrol (3:2), to
give 15 as a colourless oil (209 mg, 25%). νmax(neat)/cmϪ1 1274
(α-P᎐O), 1251 (β-P᎐O); δ (400 MHz, CDCl ) 1.34 (3H, d,
Isopropyl phenylphosphonochloridate (13). To a three necked
round bottomed flask fitted with an overhead mechanical stir-
rer was added a solution of phenylphosphonic dichloride (25.64
g, 0.131 mol) in anhydrous toluene (75 cm3). This was cooled to
0 ЊC and a mixture of triethylamine (18.4 cm3, 0.131 mol) and
anhydrous propan-2-ol (10.1 cm3, 0.131 mol) was slowly added
dropwise whilst stirring vigorously. This reagent addition gave
an exothermic reaction with the formation of a white precipi-
tate. The thick slurry was allowed to warm to room temperature
and left to stir for 90 min. The solids were filtered and washed
with toluene. Solvents were then removed to give a crude yellow
oil which was purified by vacuum distillation to give 13 as a
colourless oil (12.17 g, 42%). Boiling point 102–108 ЊC (1.5
mmHg); νmax(neat)/cmϪ1 1453 (Ph–P), 1275 (P᎐O); δ (300
᎐
᎐
᎐
H
H
3
MHz, CDCl3) 1.44 (3H, d, J 6.2, CH3), 1.48 (3H, d, J 6.3, CH3),
5.02–5.15 (1H, m, CH), 7.46–7.54 (3H, m, Ar-H), 7.83–7.93
(2H, m, Ar-H); δC(100 MHz, CDCl3) 23.53 (d, J 5.2, CH3),
J 7.0, CH3), 1.37 (3H, d, J 7.0, CH3), 1.38 (6H, dt, J 7.5, 0.6,
2 × CH3), 1.68 (3H, ddd, J 15.6, 2.0, 1.1, PCH3), 4.28–4.39 (4H,
m, 2 × CH2), 4.96 (1H, dseptet, J 7.1, 5.5, CH); δC(100 MHz,
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050
1047

View Online
CDCl3) 12.36 (d, J 100.4, PCH3), 16.33 (d, J 5.5, CH3), 23.42 (d,
J 5.5, CH3), 24.81 (d, J 2.2, CH3), 65.20 (d, J 6.7, CH2), 65.42
(d, J 6.5, CH2), 72.98 (d, J 7.0, CH); δP(162 MHz, CDCl3) MR
of an ABMR system, 4.64 (dt, J 85.5, 60.3, Pα), 38.12 (ddd,
J 80.0, 71.5, 60.2, Pβ); δF(376 MHz, CDCl3) AB of an ABMRX3
system, 37.99 (ddd, J 369.7, 84.5, 71.6), 41.72 (dddq, J 369.7,
85.8, 79.9, 1.9) [Found: (EI) Mϩ, 308.07557. C9H20F2O5P2
requires 308.07539]; m/z 308 (0.7%, Mϩ), 210 (100), 171 (7.4,
M Ϫ (EtO)2P(O)), 121 (13.0, M Ϫ (EtO)2P(O)CF2), 51 (1.9,
CF2Hϩ).
(77 mg, 0.27 mmol) under an inert atmosphere and was left to
stir at room temperature for three days. The excess bromosilane
was removed in vacuo before the addition of water (2 cm3)
which was then left to stir for 30 min after which the cloudy
emulsion had cleared. The reaction mixture was washed with
ether (4 × 1 cm3). Aqueous ammonium bicarbonate solution
(1 M, 2 cm3) was added to the aqueous layer and excess volatile
salts were removed by repeated coevaporation with water before
the sample was finally lyophilised to give the product ammo-
nium salt 27 as a glassy white solid in quantitative yield (73 mg).
νmax(KBr)/cmϪ1 3600–2600 (br, NH3), 2300–2100, 2050–1800
Difluoromethylenebis[isopropyl(phenyl)phosphonate]
(19).
(br, POH), 1183 (P᎐O), 1091 and 1124 (d, P᎐O); δ (300 MHz,
᎐
᎐
H
The crude fractions of the isomeric mixture of 19 were further
purified by chromatography using EtOAc–petrol (2:5) as eluent
to give a racemic mixture of 19 (43 mg). The mesomeric isomers
were not isolated although they were evident in the crude mix-
ture. δH(300 MHz, CDCl3) 1.33 (6H, d, J 6.6, 2 × CH3), 1.44
(6H, d, J 6.6, 2 × CH3), 5.05–5.19 (2H, m, 2 × CH), 7.40–7.49
(4H, m, 4 × ArH), 7.53–7.60 (2H, m, 2 × ArH), 7.80–7.90 (4H,
m, 4 × ArH); δC(100 MHz, CDCl3) 23.77 (CH3), 24.57 (CH3),
73.57 (d, J 6.6, CH), 126.80 (d, J 150.6, C), 128.33 (d, J 13.6,
CH), 133.07 (d, J 10.1, CH), 133.41 (CH); δP(162 MHz, CDCl3)
24.74 (t, J 82.0); δF(376 MHz, CDCl3) 44.82 (t, J 82.0).
The 19F NMR data for the meso isomer of 19 is as follows:
δF(376 MHz, CDCl3) AB of an ABX spectra, 41.70 (dt, J 367.0,
75.2), 45.28 (dt, J 367.1, 85.3).
D2O) 1.51 (3H, d, J 15.0, PCH3); δC(100 MHz, D2O) 13.58 (d,
J 96.0, PCH3); δF(376 MHz, D2O) 41.36 (t, J 76.6); δP(162 MHz,
D2O) 32.28 (dt, J 78.4, 48.6, Pβ), 5.00 (dt, J 73.6, 47.3, Pα)
[Found: (FABϪ) M Ϫ HϪ, 208.95768. C2H5F2O5P2 requires
208.95803]; m/z 209 (100%, M Ϫ HϪ).
Difluoro[hydroxy(phenyl)phosphoryl]methylphosphonic acid
(28). Bromotrimethylsilane (2.4 cm3, 18.18 mmol) was added
dropwise to a neat sample of 16 (210 mg, 0.57 mmol) under
nitrogen and left to stir at room temperature for 3 days. The
mixture was cooled to 0 ЊC and quenched with water (3 cm3).
This solution was left to stir for 2 h before extracting with ethyl
acetate (4 × 4 cm3). Decolourising charcoal (100 mg) was then
added to the aqueous layer which was agitated at 50 ЊC for
2 h. The charcoal was filtered and solvents removed to give a
colourless oil. Residual impurities and hydrobromic acid were
removed by reverse-phase chromatography eluting with H2O.
Product fractions were evaporated to give 28 as a white solid
(137 mg, 91%). This was characterised as the free acid, but
resolution of the infra-red spectra was markedly improved
by first converting a sample into the tris-ammonium salt.
νmax(neat)/cmϪ1 2500–3500 (NH3), 2100–2300, 1800–2050 (br,
Isopropoxydifluoromethylphenylphosphine oxide (20). This
was obtained as a crude mixture with 17 (130 mg, purity 85%
w/w). Attempts to remove the contaminant 17 by Kugelrohr
distillation were not successful. νmax(neat)/cmϪ1 1440 (Ph–P),
1249 (P᎐O); δ (300 MHz, CDCl ) 1.37 (3H, d, J 6.1, CH ), 1.46
᎐
H
3
3
(3H, d, J 6.1, CH3), 4.84–4.96 (1H, m, CH), 5.97 (1H, dt, J 50.7,
24.3, PCF2H), 7.50–7.58 (2H, m, 2 × ArH), 7.60–7.68 (1H, m,
ArH), 7.82–7.95 (2H, m, 2 × ArH); δC(100 MHz, CDCl3) 24.13
(d, J 4.0, CH3), 24.32 (d, J 4.1, CH3), 72.71 (d, J 6.6, CH),
113.27 (dt, J 259.9, 147.4, PCF2H), 125.68 (d, J 135.5, C),
128.76 (d, J 13.3, CH), 132.80 (d, J 10.0, CH), 133.78 (d, J 2.9,
CH); δP(162 MHz, CDCl3) X of an ABX system, 23.43 (dd,
J 83.8, 76.9); δF(376 MHz, CDCl3) AB of an ABX system,
28.13 (ddd, J 346.2, 83.8, 49.0), 30.18 (ddd, J 346.2, 76.9, 49.3)
[Found: (EI) Mϩ, 234.06228. C10H13F2O2P requires 234.06213];
m/z 234 (0.1%, Mϩ), 183 (37.3, M Ϫ CF2H), 141 (100, PhP-
(O)OHϩ), 77 (86, C6H5ϩ), 51 (39.3, CF2Hϩ).
P–OH), 1703 (P–OH), 1485, 1457 (PhP), 1186 (α-P᎐O), 1124
᎐
and 1085 (β-P᎐O); δ (300 MHz, D O) 7.60–7.68 (2H, m, 2 ×
᎐
H
2
ArH), 7.70–7.77 (1H, m, ArH), 7.88–7.98 (2H, m, 2 × ArH);
δC(100 MHz, D2O) signals too broad to resolve; δP(162 MHz,
D2O) 4.46 (br, Pα), 21.57 (dt, J 76.6, 50.0, Pβ); δF(376 MHz,
D2O) 41.34 (dd, J 85.2, 78.6) [Found: (FAB) Mϩ, 272.98911.
C7H8F2O5P2 requires 272.98933]; m/z 273 (46%, Mϩ), 77 (46,
C6H5ϩ), 51 (24, CF2Hϩ).
General procedure for the preparation of the tributylammonium
salts
Tetraethyl difluoromethylenediphosphonate (18).23 To a solu-
tion of LDA (1.27 mmol) in THF (1.5 cm3) at Ϫ78 ЊC was
added a solution of diethyl difluoromethylphosphonate 17 (218
mg, 1.16 mmol) in THF (1.5 cm3). This was left to stir at
Ϫ78 ЊC for 30 min before the addition of diethyl chlorophos-
phate (0.17 cm3, 1.16 mmol) in THF (1.5 cm3). The reaction was
then stirred at Ϫ78 ЊC for 60 min before warming to 0 ЊC and
quenching with a saturated solution of aqueous potassium
dihydrogen phosphate. This was extracted with EtOAc (3 × 8
cm3), dried (MgSO4) and the solvents were then removed to give
a crude yellow oil. Purification by chromatography, using
EtOAc–petrol (3:2) as eluent gave a colourless oil which was
further purified by Kugelrohr distillation (impurities and some
of the product distilled over at 130 ЊC, 2 mmHg) to leave the
clean product residue 18 as a colourless oil (128 mg, 34%).
νmax(neat)/cmϪ1 1282 (P᎐O), 1038 (P–OEt); δ (300 MHz,
All the tributylammonium salts of the pyrophosphate anal-
ogues were prepared as follows: To a solution of 2936 (1.32
mmol) in 50% aqueous ethanol (2 cm3) was added tributyl-
amine (1.98 mmol). This was stirred at room temperature for 30
min before removing the solvents to give the tributylammonium
salt 32 as a viscous brown oil. This was stored as a solution
in DMF (0.6 M). Before being used in the subsequent phos-
phorylation reaction, solvents were removed from an appropri-
ately sized aliquot and residual water was removed in vacuo as
an azeotropic mixture (3 × 1 cm3 of pyridine followed by 1 cm3
of benzene).
Recombinant retroviral vector and virus production
The recombinant retrovirus encoding β-galactosidase (pSG1)
used for measuring antiviral activity has been described previ-
ously.27 The pSG1 vector was constructed by introducing a full
size 3.4 kilobase fragment containing the E. coli lacZ gene
from pWB310 plasmid37 to the cloning site of pPS-3-neo
eukaryotic vector derived from the Mu-MLV.38 The structure of
the resulting vector pSG1 is as follows:
᎐
H
CDCl3) 1.39 (12H, t, J 7.1, 4 × CH3), 4.25–4.42 (8H, m, 4 ×
CH2); δC(100 MHz, CDCl3) 16.34 (t, J 2.7, CH3), 65.33 (t, J 3.1,
CH2); δP(162 MHz, CDCl3) 4.29 (t, J 86.3); δF(376 MHz,
CDCl3) 42.40 (t, J 86.3) [Found: (EI) Mϩ, 324.07061. C9H20-
F2O6P2 requires 324.07032]; m/z 324 (2.5%, Mϩ), 212 (100), 188
(36.3, (EtO)2P(O)CF2Hϩ).
Difluoro[hydroxy(methyl)phosphoryl]methylphosphonic acid
(tris-ammonium salt) (27). Bromotrimethylsilane (0.4 cm3, 3.03
mmol) was added dropwise to a neat sample of the triester 15
1048
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050

View Online
The vector containing cis-acting regulatory elements of
Mu-MLV genome [long terminal repeats (LTR) and ψ-region]
which are necessary for viral transcription, packaging and inte-
gration into the host genome. The lacZ gene is transcribed from
5Ј-LTR. The transmitting G418 resistance transposon 5 neo-
mycin phosphotransferase (neo) gene is located downstream
from the lacZ gene and transcribed from the internal simian
virus 40 (SV40) early promoter. The β-lactamase (Ampr) gene
and the origin of replication are derived from the bacterial
pSP64/65 plasmid.38 The vector is replication defective and
allows the analysis of a single cycle of infection, since infectious
progeny is not produced and subsequent cycles of viral repli-
cation do not occur. The pSG1 construct was transferred into
the PA317 amphotropic packaging cell line by the calcium–
phosphate-DNA coprecipitation.39 After 48 h the culture
medium was harvested and used to infect the ψ2 ecotropic
packaging cell line in the presence of polybrene (8 µg mlϪ1).40
Individual ψ2 clones expressing the lacZ gene were isolated
after the selection in the G418 (0.5 mg mlϪ1) medium for 20
days. Titres of lacZ transducing retroviruses were determined
by infecting Rat1 fibroblasts with a series of dilutions of the
culture medium from each clone. After 48 h cells were incu-
bated with 5-bromo-4-chloro-3-indolyl-β--galactopyranoside
(Xgal; Sigma) at 37 ЊC overnight to reveal β-galactosidase activ-
ity41 and blue cell colonies were counted. When assayed on
Rat1 cells, the filtered supernatant from ψ-PA-pSG1 clone con-
tained approximately 104 β-galactosidase focus-forming units
(ffu) per cm3. With murine Balb/3T3 fibroblasts used as targets
for the infection, the apparent titre of pSG1 virus was about
ten-fold higher than that with Rat1 cells. The virus was used
immediately or stored at Ϫ70 ЊC.
mM, final concentration 0.32 mM), 4 µl of 2 or 3 (40 mM, final
concentration 3.2 mM) and 44 µl of human blood serum. Con-
trol assays were performed for 7 at a concentration of 2 mM.
Primer extension assays
For the template-dependent DNA polymerases, the assay mix-
ture (volume 6 µl) contained [5Ј-32P]-labelled-template-primer
(0.01 µM), the compounds under study or dNTPs, the enzyme
(2 activity units of RTs or 1 unit of DNA polymerases α and β)
and the corresponding buffer. For TDT, the assay mixture (vol-
ume 5 µl) contained the primer (0.1 µM), the compound under
study, 2 units of the enzyme, 100 mM sodium cacodylate (pH
7.2), 10 mM MgCl2, 1 mM CoCl2 and 1 mM dithiothreitol. The
reaction was carried out for 30 min at 37 ЊC and terminated by
adding deionised formamide (3 µl containing 0.5 mM EDTA),
2% bromophenol blue and xylene cyanol. The reaction prod-
ucts were separated by electrophoresis in 20% polyacrylamide
gel, and the gels obtained were radioautographed.
Kinetic measurements
These were performed within the linear region of the product
formation, as previously described.26 The reaction time was 2
min.
Cytotoxic assay
To assess cytotoxicity of the compounds under study, a lactate
dehydrogenase (LDH)-release assay (CytoTox 96 Non-
Radioactive Cytotoxicity Assay; Promega, USA) was utilised.
Rat1 cells were incubated for 24 h with the compounds under
study and the colourimetric LDH-release assay was performed
according to the instructions of the manufacturer. Cells incu-
bated in the absence of the NTP analogues were used as a
control. Cytotoxic effects were not observed with any of the test
compounds at concentrations up to 1 mM (data not shown).
Template–primer complexes used in the chain elongation
experiments:
Evaluation of activity
For measuring antiviral activity, Rat1 cells were seeded in 2 cm
wells of 24-well tissue culture plates in 1 cm3 of Dubelco modi-
fied essential medium (ICN Biomedicals. Inc.) with 10% foetal
bovine serum (ICN Biomedicals. Inc.), 2 mM glutamine and
penicillin–streptomycin (Pen/STrep; ICN Biomedicals. Inc.).
After 2 days the medium was replaced with 1 cm3 of 1:10 fresh
viral supernatant containing Polybrene (8 µg mlϪ1), and the
compounds under study were added (as a series of ten-fold
dilutions in the medium) at the time of infection. After incubat-
ing for 24 h (37 ЊC), the old media was exchanged with 1 cm3 of
a fresh one (without test compounds), and after a day of addi-
tional incubation the cells were fixed and stained with Xgal.
The blue cell colonies expressing β-galactosidase were counted.
The results were expressed as the percentage of β-gal-positive
cells in the presence of inhibitor compared to the proportion of
β-gal-positive cells in the control. The results were averaged
from data taken from 3 parallels in 2 independent experiments.
Complex 1:
dGGGTCAGTGCTGCAACATTTTGCTGCCGGTCAC
GGTTC . . . .
[5Ј-32P] dCCCAGTCACGACGT →
Complex 2:
dCTGCAACATTTTGCTGCCGGTCACGGTTCGAAC
CCGA . . . .
[5Ј-32P] dGACGTTGTAAAACG →
Complex 3:
dCATTTTGCTGCCGGTCACGGTTCGAACCCGACG
TCCA . . . .
[5Ј-32P] dGTAAAACGACGGCCAGT →
Complex 4:
The estimation of the dephosphorylation rate
rUACGCUGAGGACGUAAUCCUUCGUCGGGUC
AU . . . .
The assay mixture, containing 2.5 µl of 10 mM solution of the
compound under study and 47.5 µl of 100% foetal blood serum,
was incubated at 37 ЊC for 2.5, 5, 10, 20, 30, 60 min, 2, 3, 4, 5, 8,
12 h, 2, 4 and 7 days. The sampled aliquots (9 µl) were stirred
with an aqueous suspension of phosphocellulose (2 mg per
50 µl) and then kept for 2 h at 20 ЊC. The samples were then
centrifuged for 6 min (12000 rpm) and the supernatants were
analysed by HPLC on a Silasorb C18 column (4 × 250 mm,
10 µ), eluting with a linear gradient of methanol (0–7.5%) in
aqueous potassium phosphate buffer (0.01 M, pH 6.5) over 25
min at a flow rate of 0.7 ml minϪ1. The extent of hydrolysis was
assessed by measuring the accumulation of the phosphonate
11 or 12. The experiments on inhibition of the phosphate
hydrolysis of 7 by triphosphonates 2 and 3 were carried out, in
a similar manner, using either 2.5 µl of 7 (8 mM, final concen-
tration 0.33 mM), 5 µl of 2 or 3 (20 mM, final concentration
1.65 mM) and 52.5 µl of human blood serum or 2.0 µl of 7 (8
[5Ј-32P] dATGCGACTCCTGC →
Acknowledgements
This paper is dedicated to the memory of Professor Ralph
Raphael, a truly inspirational synthetic organic chemist. These
studies were supported by the Russian Foundation for Basic
Research, projects 96-04-48277, 96-04-48278; Russian National
Priorities in Medicine and Public Health Care, AIDS, projects
sp1 and sp2; the National Institute of Health (Career Devel-
opment Award Grant). The Medical Research Council (UK) is
also thanked for its support (studentship to C. J. H.).
References
1 H. Zhang, G. Dornadula and R. Pomerantz, J. Virol., 1996, 70,
2809.
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050
1049

View Online
21 (a) M. Harger, J. Chem. Soc., Perkin Trans. 2, 1991, 1057; (b)
2 A. A. Arzumanov, D. G. Semizarov, L. S. Victorova, N. B. Dyatkina
and A. A. Krayevsky, J. Biol. Chem., 1996, 271, 24389.
3 L. S. Victorova, N. B. Dyatkina, D. J. Mozzherin, A. M. Atrazhev,
A. A. Krayevsky and M. K. Kukhanova, Nucleic Acids Res., 1992,
20, 783.
A. Gerrard and N. Hamer, J. Chem. Soc. (B), 1968, 539; (c) S.
Freeman and M. Harger, J. Chem. Soc., Perkin Trans. 2, 1988, 82;
(d) S. Freeman and M. Harger, J. Chem. Soc., Perkin Trans. 1, 1988,
2737.
4 N. B. Dyatkina, A. A. Arzumanov, L. S. Victorova, M. K.
Kukhanova and A. A. Krayevsky, Nucleosides Nucleotides, 1995, 14,
91.
5 V. Merlo, S. M. Roberts, R. Storer and R. C. Bethell, J. Chem. Soc.,
Perkin Trans. 1, 1994, 1477.
6 D. Coe, S. M. Roberts and R. Storer, J. Chem Soc., Perkin Trans. 1,
1992, 2695.
7 J. Balzarini, P. Herdewijn, R. Pauwels, S. Broder and E. De Clercq,
Biochem. Pharmacol., 1988, 37, 2395.
22 M. Harger and B. Hurman, J. Chem. Soc., Chem. Commun., 1995,
1701.
23 The synthesis of 18 using this procedure has previously been noted
in the following paper: M. Obayashi, E. Ito, K. Matsui and
K. Kondo, Tetrahedron Lett., 1982, 23, 2323. Full experimental
details are reported herein as the experimental procedure was not
described in the original paper and to our knowledge has not been
published since. Characterisation data are available for 18 from
reports of alternative preparations.16,35,41
8 T. J. Kilesso, N. B. Tarussova, E. D. Atrazheva, M. K. Kukhanova,
S. V. Shulenin, A. F. Bobkov, M. M. Garayev, G. A. Galegov and
A. A. Krayevsky, Bioorg. Khim., 1990, 16, 531.
9 T. A. Rozovskaya, A. V. Tischenko, N. B. Tarussova, M. K.
Kukhanova, A. A. Krayevsky and R. Sh. Beabealashvilli, Mol. Biol.
Russian, 1993, 27, 1051.
24 V. Wittmann and C. H. Wong, J. Org. Chem., 1997, 62, 2144.
25 C. J. Hamilton, PhD Thesis, University of Exeter, 1998.
26 D. G. Semizarov, L. S. Victorova, N. B. Dyatkina, M. von Janta
Lipinski and A. Krayevsky, FEBS Lett., 1994, 354, 187.
27 A. Y. Shevelev, O. A. Deyanova, L. E. Goryunova, A. Y. Zhukova
and R. Sh. Beabealashivilli, Mol. Biol. Russian, 1995, 29, 658.
28 P. Labataille, P. Pelicano, J.-P. Imbach and G. Gosselin, Bioorg. Med.
Chem. Lett., 1995, 5, 2315.
10 L. Arabshahi, N. N. Khan, M. Butler, T. Noothy, N. C. Brown and
G. E. Wright, Biochemistry, 1990, 29, 6820.
11 B. I. Martynov, E. A. Shirokova, M. V. Jasko, L. S. Victorova and
A. A. Krayevsky, FEBS Lett., 1997, 410, 423.
12 E. A. Shirokova and N. B. Dyatkina, Collect. Czech. Chem.
Commun., 1996, 61, S158.
13 N. Dyatkina, E. Shirokova, F. Theil, S. M. Roberts and A.
Krayevsky, Bioorg. Med. Chem. Lett., 1996, 6, 2639.
14 N. Dyatkina, F. Theil and M. von Janta Lipinski, Tetrahedron, 1995,
51, 761.
29 G. R. J. Thatcher and A. S. Campbell, J. Org. Chem., 1993, 58, 2272.
30 A. A. Krayevsky, unpublished work.
31 L. A. Alexandrova, A. J. Scoblov, M. V. Jasko, L. S. Victorova and
A. A. Krayevsky, Nucleic Acids Res., 1998, 26, 778.
32 D. Y. Mozzherin, A. M. Atrazhev and M. K. Kukhanova, Mol. Biol.
Russian, 1992, 26, 999.
33 T. A. Kolocheva and G. A. Nevinsky, Mol. Biol. Russian, 1993, 27,
1368.
15 C. J. Hamilton, S. M. Roberts and A. Shipitsin, Chem. Commun.,
1998, 1087.
34 A. V. Krayev, Mol. Biol. Russian, 1988, 22, 1164.
35 J. R. Sampson and O. Uhlenbeck, Proc. Natl. Acad. Sci. USA, 1988,
85, 1033.
36 C. McKenna and P. Shen, J. Org. Chem., 1981, 46, 4573.
37 N. Lehming, J. Sartorius, M. Niemoller, G. Genenger, B. von
Wilken-Bergmann and B. Muller Hill, EMBO J., 1987, 6, 3145.
38 V. S. Prasolov and P. M. Chumakov, Mol. Biol. Russian, 1988, 22,
1371.
16 D. Burton and R. Flynn, J. Fluorine Chem., 1980, 15, 263.
17 The high electrophilic reactivity of the acylation products of 21,
leading to the rapid formation of secondary products, has
previously been reported; G. M. Blackburn, D. Brown, S. Martin
and M. Parrat, J. Chem. Soc., Perkin Trans. 1, 1987, 181.
18 F. Westheimer, Acc. Chem. Res. 1968, 1, 70.
19 (a) E. Lightcap and P. Frey, J. Am. Chem. Soc., 1992, 114, 9750; (b)
J. Friedman, S. Freeman and J. Knowles, J. Am. Chem. Soc., 1988,
110, 1268.
20 (a) P. Cullis, R. Misra and D. Wilkins, J. Chem. Soc., Chem.
Commun., 1987, 1594; (b) P. Cullis and R. Misra, J. Am. Chem. Soc.,
1991, 113, 9679; (c) S. Hamett and G. Lowe, J. Chem. Soc., Chem.
Commun., 1987, 1416.
39 A. D. Miller and C. Buttimore, Mol. Cell Biol., 1986, 6, 2895.
40 R. Mann, R. C. Milligan and D. Baltimore, Cell, 1983, 33, 153.
41 J. R. Sanes, J. L. R. Rubenstein and J. F. Nicolas, EMBO J., 1986, 5,
3133.
Paper 9/00336C
1050
J. Chem. Soc., Perkin Trans. 1, 1999, 1039–1050