Synthesis, Characterization, and Biological Properties of 8-Azido- and 8-Amino-Substituted 2′,5′-Oligoadenylates

Article abstract of DOI:10.1021/jm030035k

A series of 8-azido- and 8-amino-substituted 2′,5′ -oligoadenylatyes was prepared by a uranylion catalyzed polymerization of the corresponding 8-substituted adenosine phosphorimidazolide. Subsequent 5′-dephosphorylation of the resulting 5′-phosphoryl 2′,5′-linked oligomers with alkaline phosphatase gave the corresponding core oligomers. The CD spectra indicated that the 8-aminoadenosine analogue of the 2′,5′-linked trimer has an anti-orientation as in naturally occurring 2′,5′-oligoadenylates, while 8-azido-substituted 2′,5′-oligoadenylates have a synorientation. The 8-substituted oligomers showed enhanced resistance against digestion by snake venom phosphodiesterase. The 2′,5′-linked 8-azidoadenylate trimer and tetramer displayed strong RNase L binding and activating ability, although the corresponding dimer is devoid of such activities. In contrast, very low or no RNase L binding and activating ability were observed in the 8-aminoadenosine analogue of 2′,5′ -oligoadenylates. Results indicate that the bulkiness and ionic character of the 8-substituting group have significant effects on the ability of these analogues to bind and activate RNase L. Furthermore, the orientation of the glycosidic base in the 2-5A analogues may change from syn to anti during binding to RNase L. The 8-azido-adenosine analogues of 2-5A will be useful tools in the photoaffinity labeling of RNase L, due to their strong RNase L binding ability. In addition, these 8-azidoadenosine compounds may be considered as candidates for experimental therapeutic agents because they have enhanced stability to enzyme degradation while retaining the ability to activate RNase L.

Full text of DOI:10.1021/jm030035k

4926
J . Med. Chem. 2003, 46, 4926-4932
Syn th esis, Ch a r a cter iza tion , a n d Biologica l P r op er ties of 8-Azid o- a n d
8-Am in o-Su bstitu ted 2′,5′-Oligoa d en yla tes
Hiroaki Sawai,*,† Atushi Hirano,†,§ Hiroyuki Mori,†, Kazuo Shinozuka,† Beihua Dong,‡ and
Robert H. Silverman‡
Department of Applied Chemistry, Faculty of Engineering, Gunma University, Kiryu, Gunma 376-8515 J apan, and
Department of Cancer Biology, The Lerner Research Institute, The Cleveland Clinic Foundation, Cleveland, Ohio 44195
Received J anuary 22, 2003
A series of 8-azido- and 8-amino-substituted 2′,5′-oligoadenylatyes was prepared by a uranyl-
ion catalyzed polymerization of the corresponding 8-substituted adenosine phosphorimidazolide.
Subsequent 5′-dephosphorylation of the resulting 5′-phosphoryl 2′,5′-linked oligomers with
alkaline phosphatase gave the corresponding core oligomers. The CD spectra indicated that
the 8-aminoadenosine analogue of the 2′,5′-linked trimer has an anti-orientation as in naturally
occurring 2′,5′-oligoadenylates, while 8-azido-substituted 2′,5′-oligoadenylates have a syn-
orientation. The 8-substituted oligomers showed enhanced resistance against digestion by snake
venom phosphodiesterase. The 2′,5′-linked 8-azidoadenylate trimer and tetramer displayed
strong RNase L binding and activating ability, although the corresponding dimer is devoid of
such activities. In contrast, very low or no RNase L binding and activating ability were observed
in the 8-aminoadenosine analogue of 2′,5′-oligoadenylates. Results indicate that the bulkiness
and ionic character of the 8-substituting group have significant effects on the ability of these
analogues to bind and activate RNase L. Furthermore, the orientation of the glycosidic base in
the 2-5A analogues may change from syn to anti during binding to RNase L. The 8-azido-
adenosine analogues of 2-5A will be useful tools in the photoaffinity labeling of RNase L, due
to their strong RNase L binding ability. In addition, these 8-azidoadenosine compounds may
be considered as candidates for experimental therapeutic agents because they have enhanced
stability to enzyme degradation while retaining the ability to activate RNase L.
In tr od u ction
very low or no RNase L binding and activating abilities.
The difference in their biological activities was partly
explained by the difference in the syn-anti conformation
around the base-glycosydic bonds of 2-5A.8-14 However,
the detailed mechanism of 2-5A and its congener to
bind and activate RNAse L is still unknown. Among the
8-substituted adenosine 2-5A analogues, 8-azidoadeno-
sine analogue is very interesting, because it has strong
biological activities despite its bulky azido group at the
8-position, and it can be useful as a photoaffinity
labeling agent for RNase L to study the binding mech-
anism.15 Enzymatically, an 8-azido analogue of 2-5A
timer is produced from 8-azido ATP by 2-5A syn-
thetase, although the yield and the scale of the enzy-
matic synthesis are very low.10 2-5A trimer bearing
8-azidoadenosine at the 2′-terminal was prepared chemi-
cally by a conventional phosphotriester method, in
which tedious blocking, coupling, and deprotection are
required in each reaction step.11 Thus, we undertook the
preparation of a series of 5′-monophosphorylated 2′,5′-
oligo-8-azidoadenylates from the dimer to the tetramer
by uranyl-ion catalyzed polymerization of 8-azidoadeno-
sine 5′-phosphorimidazolide in aqueous solution, as the
procedure is very easy to perform with high 2′,5′ re-
gioselectivity without using any protecting group. We
have also prepared a series of 8-aminoadenosine ana-
logues of 2-5A by the same methodology to examine
the effect of an amino group at the 8-position on the
biological activity. We have evaluated biological activi-
ties of these 2-5A analogues, along with the corre-
sponding core analogues prepared from the correspond-
2′,5′-Oligoadenylates (2-5A) are formed from ATP
when viral double-stranded RNA binds to 2-5A syn-
thetase in interferon-treated cells. The 2-5A binds and
activates 2-5A dependent RNase (RNase L) which
degrades viral mRNA, resulting in an inhibition of the
protein synthesis.1-6 The 2-5A plays an important role
in the mechanism of interferon’s antiviral activity, but
is also involved in other cellular processes. During the
past two decades, various 2-5A analogues have been
synthesized to study the biological function of 2-5A and
its structure-activity relationship and to explore the
potential for deriving an antiviral agent based on 2-5A.
Several 8-substitued analogues of 2-5A were prepared
and their biological activities were evaluated. 2-5A
analogues with 8-bromo,7,8 8-hydroxy,9 8-azido,10,11 and
8-methyl12 groups in the 2′-terminal residue have been
reported to have strong RNase L binding and activating
ability, equal to or stronger than that of naturally
occurring 2-5A, while the corresponding 2-5A ana-
logues bearing fully 8-substituted adenosine showed 10-
to 100-fold reduction of the biological activities. On the
other hand, 2-5A analogues containing an 8-mercapto13
or 8-hydroxypropyl9 group have been reported to possess
* To whom correspondence should be addressed. Tel: 81-277-30-
1220. Fax: 81-277-30-1224. E-mail: sawai@chem.gunma-u.ac.jp.
† Gunma University.
‡ The Cleveland Clinic Foundation.
§ Present address: Banyu Tsukuba Research Institute, Okubo-3,
Tsukuba, Ibaraki, 300-2611 J apan.
Present address: Fine Chemical Division, J . F. E. Chemical Co.,
9-2 Koukan-chou, Kasaoka, Okayama, 714-0063 J apan.
10.1021/jm030035k CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/04/2003

8-Substituted 2′,5′-Oligoadenylates
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23 4927
Ta ble 1. Synthesis of 8-Azido- and 8-Amino-Substituted
Oligoadenylates by Uranyl Ion-Catalyzed Polymerization of
ImpA-Xa
Sch em e 1. Synthetic Scheme of 2′,5′-Linked
Oligoadenylates with 8-Azido and 8-Amino Groups
yield (%)*
[ImpA-X]/
[uranyl ion] pA-X (pA-X)2 (pA-X)3 (pA-X)4 (pA-x)5 (pA-X)6-8
X-N3
5
10
20
5.6
4.6
2.9
0.6
4.7
18.8
18.6
23.6
30.9
25.8
23.5
24.8
25.9
22.0
18.7
4.8
7.6
9.0
3.9
2.4
2.9
2.7
2.7
0.7
0.2
1.5
2.6
2.1
0.3
50
100
X-NH2
5
10
50
50.6
67.7
78.2
12.8
6.9
2.6
6.6
6.0
2.9
2.1
a
Polymerization of ImpA-X (50 mM) was carried out at 20 °C
b
for 24 h in 0. 2 M N-ethylmorpholine buffer (pH 7.0). No
correction for the hypochromicity of each oligomer was done on
the yield.
reaction. 5′-Phosphoryl 2′,5′-oligoadenylates bearing
8-azido group or 8-amino group from the dimer to the
tetramer were prepared in a large scale under the
optimum reaction condition and purified by DEAE-
Sephadex anion exchange column chromatography fol-
lowed by preparative HPLC or paper chromatography.
The corresponding core analogues were obtained by
dephosphorylation of the 5′-phosphoryl 2′,5′-oligoaden-
ylates analogues with alkaline phosphatase. The struc-
tures of these 2-5A analogues were assigned by ESI-
mass and the analyses of enzyme digestion products,
as well as NMR when sufficient quantity is available.
The results are listed in Tables 2 and 3, with hypochro-
micity data of the 5′-phosphoryl oligonucleotides.
Figure 1 shows the CD spectra of the 2′,5′-linked
trimers of 8-azidoadenosine and 8-aminoadenosine at
several temperatures. The CD spectrum of the trimer
of 8-aminoadenosine showed positive ellipticity at longer
wavelength and negative ellipticity at lower wavelength
around 270 nm, indicative of a right-handed helix with
an anti-orientation about the base-glycoside bonds as a
naturally occurring 2-5A.14,19 The 8-amino group of
adenine interacts with 5′-phosphate in the 2-5A ana-
logue forming anti-conformation. On the contrary, the
CD spectrum of the trimers of 8-azidoadenosine showed
negative ellipticity at longer wavelength and positive
ellipticity at lower wavelength around 280 nm, which
indicates that two adjacent 8-azide adenosine moieties
are in a strongly right-handed stacking arrangement
with a syn-orientation about the base-glycoside bonds.19
The presence of a bulky 8-azido group in the 2-5A
analogues could make syn-conformation as the 8-bro-
moadenosine analogue of 2-5A.
The stability of 8-azido- and 8-aminoadenosine ana-
logues of 2-5A against snake venom phosphodiesterase
(SVPD) and bacterial alkaline phosphatase was as-
sessed using the corresponding 2,′5′-linked dimers.
8-Azido and 8-amino substitution enhanced the stability
against SVPD. The half-lives of the intact 2′,5′-linked
dimers of 8-azido-, 8-amino-, and the unsubstituted
adenosine against SVPD were 330, 37, and 7.5 min,
respectively, under the same condition. Thus 8-azido
group in the 2-5A markedly enhance the stability
against SVPD. The enhancement of stability against
phosphodiesterase has been reported in other 8-substi-
tuted 2-5A analogues.8-12 Both 2-5A analogues were
dephosphorylated at the same rate as the corresponding
ing 5′-phosphorylated analogues with alkaline phos-
phatase.
Resu lts a n d Discu ssion
Syn th esis a n d Ch a r a cter iza tion of 2′,5′-Oligoa d -
en yla te An a logu es Bea r in g a n 8-Azid o or 8-Am in o
Gr ou p . 8-Aminoadenosine phosphorimidazolide was
prepared by the published method from 8-aminoadeno-
sine 5′-phosphate and imidazole using triphenyl phos-
phine and dipyridyl disulfide as a condensing agent.16
However, this method could not be applied for the
synthesis of 8-azidoadenosine 5′-phosphorimidazolide,
because an azido group reacts easily with triphenylphos-
phine forming triphenylphosphine-nitride. Thus, we
employed a carbonyldiimidazole method17 for the forma-
tion of 8-azidoadenosine phosphorimidazolide after tran-
sient protection of the 2′,3′-dihydroxy groups of 8-azi-
doadenosine with phenylboronic acid as shown in Scheme
1. The phenyl boronate ester of 2′,3′-diol was deblocked
by treatment with ethylene glycol and 1,3-propandiol
in methanol without degrading the phosphorimidazolide
bond. We conducted polymerization of 8-azidoadenosine
5′-phosphorimidazolide by an uranyl-ion catalyst in
neutral aqueous solution for the synthesis of 2′,5′-
oligomers of 8-azidoadenosine, as the reaction proceeds
with high 2′,5′ regioselectivity without using any pro-
tecting group.18 Table 1 shows the HPLC yield data of
oligo 8-azidoadenylates under several conditions. Pref-
erential formation of 2′,5′-linked oligomers from the
dimer to the octamer was observed under the optimum
reaction conditions. 2′,5′-Linkage in the resulting oli-
gomers was confirmed by the resistance against nu-
clease P1 digestion. Polymerization of 8-aminoadenosine
5′-phosphorimidazolide took place in the same way,
although yields and chain length of the resulting oligo-
mers were very low as shown in Table 1. The amino
group is likely located close to the phosphoryl group and
catalyzes the hydrolysis of the phosphorimidazolide
bond intramolecularly suppressing the polymerization

4928 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23
Sawai et al.
Ta ble 2. ESI-Mass, Characteristic NMR, and Hypochromicity of 8-Azido- and 8-Amino-Substituted 2′,5′-Oligoadenylates
NMR
ESI-mass in negative mode (calcd)
2-5A analogue
[M - H]*, [M - 2H]/2**, [M - 3H]/3***
H-2
H-1′
hypochromicity (%)
ApA-N3
676.9 (677.1)*
NDa
pApA-N3
757.1 (757.1)*, 378.0 (378.1)**
8.05 (s, 1H)
8.18 (s, 1H)
6.09 (d, 0.6 Hz, 1H)
5.62 (d, 6.8 Hz, 1H)
ND
6.00 (d, 2.2 Hz, 1H)
5.74 (d, 3.1 Hz, 1H)
5.59 (d, 6.9 Hz, 1H)
ND
6.01 (d, 2.7 Hz, 1H)
5.77 (d, 2.9 Hz, 1H)
5.74 (d, 2.6 Hz, 1H)
5.62 (d, 6.9 Hz, 1H)
ND
13
17
21
ApApA-N3
pApApA-N3
1047.0 (1047.2)*, 522.9 (523.1)**
1127.1 (1127.2)*, 563.0 (563.1)**
8.03 (s, 2H)
8.16 (s, 1H)
ApApApA-N3
pApApApA-N3
708.2 (708.2)**
748.1 (748.1)**, 498.6 (498.4)***
8.03 (s, 3H)
8.13 (s, 1H)
ApA-NH2
625.3 (625.2)*
pApA-NH2
705.1 (705.1)*, 352.3 (352.1)**
8.05 (s, 1H)
8.16 (s, 1H)
6.20 (d, 6.5 Hz, 1H)
5.85 (d, 6.1 Hz, 1H)
ND
ND
ND
12
ApApA-NH2
969.2 (969.2)*, 484.2 (484.1)**
1049.1 (1049.2)*
1313.5 (1313.3)*, 656.2 (656.2)**
696.2 (696.2)**, 463.8 (463.8)***
pApApA-NH2
ApApApA-NH2
pApApApA-NH2
21
24
ND
a
ND: Not meassured, as the sample quantity is too small.
Ta ble 3. Enzyme Digestion of 8-Azido- and 8-Aminoadenosine Analogues of 2-5A
2-5A analogue
(retention time)
BAP digest product
(retention time)
SVP digest product
(retention time)
NP1 digestion
ApA-N3
pApA-N3 (7.8)
ApApA-N3
pApApA-N3 (10.7)
ApApApA-N3
pApApApA-N3 (13.7)
ApA-NH2
pApA-NH2 (10.8)
ApApA-NH2
pApApA-NH2 (14.8)
ApApApA-NH2
pApApApA-NH2 (20.7)
A-N3 + pA-N3
pA-N3 (2.9)
A-N3 + 2pA-N3
pA-N3 (2.9)
A-N3 + 3pA-N3
pA-N3 (2.9)
A-NH2 + pA-NH2
pA-NH2 (4.4)
A-NH2 + 2 pA-NH2
pA-NH2 (4.4)
A-NH2 + 3pA-NH2
pA-NH2 (4.4)
no digestion
no digestion
no digestion
no digestion
no digestion
no digestion
no digestion
no digestion
no digestion
no digestion
no digestion
no digestion
ApA-N3 (2.8)
ApApA-N3 (10.0)
ApApApA-N3 (15.8)
ApA-NH2 (2.8)
ApApA-NH2 (9.2)
ApApApA-NH2 (16.1)
RNa se L Bin d in g a n d Activa tin g Activity of
8-Azid oa d en osin e a n d 8- Am in oa d en osin e An a -
logu es of 2-5A. Activity of the 8-azido- or 8-amino-
adenosine analogues of 2-5A to bind with RNAse L was
evaluated by the radiobinding assay described previ-
ously using 125I-labeled 2-5A analogue as a probe.20
Mouse liver extract was used as a source of RNase L.
The binding ability is expressed as a IC50, a concen-
tration of the 2-5A analogue to inhibit binding of
the radiolabeled 2-5A probe by 50%. The results
are summarized in Table 4, along with the relative
binding ability to the corresponding unmodified
2-5A, pA2′p5′A2′p5′A. The trimer and tetramer of
5′-phosphoryl 8-azidoadenosine analogues of 2-5A
were bound to the RNAse L slightly more effectively
than the unmodified 2-5A trimer, pA2′p5′A2′p5′A,
and the corresponding core analogues had a slightly
weaker binding ability. The dimer of 2′,5′-linked 8-azi-
doadenosine showed very low or no binding ability. On
the other hand, 2-5A analogues of 8-aminoadenosine
showed very poor binding ability. The tetramers,
pApApApA-NH2 and ApApApA-NH2, showed over 100-
fold decrease in the binding with RNase L. The trimer
and dimer of 2′,5′-linked 8-aminoadenosine have no
binding ability at 10 µM concentration.
F igu r e 1. CD spectra of 2′,5′-linked triadenylates with 8-azido
and 8-amino groups at 4 s, 22 - - -, 52 ‚‚‚, and 82 °C -‚-.
(A) pApApA-N3 and (B) pApApA-NH2.
Assay for activation of RNase L with the 2-5A ana-
logues was carried out by a previously described method
that uses human recombinant RNase L.21,22 32P-Labeled
oligoribonucleotide, C7UUC12[32P]pCp, was used as an
unsubstituted 2-5A by bacterial alkaline phosphatase.
The half-lives of the intact 5′-phosphoryl dimers were
5-6 min when 0.04 µmol of the substrate and 0.01 unit
of the enzyme were used.

8-Substituted 2′,5′-Oligoadenylates
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23 4929
Ta ble 4. Binding Ability of 8-Azido- and 8-Amino-Substituted
Analogues of 2-5A to RNase L in Mouse Liver Extract
tion similarly when they are bound to RNase L. The
conformation of the active form of 2-5A or 2-5A
analogues is not clear, and NMR, CD, or X-ray crystal-
lographic studies of the bound state with RNase L are
needed for its detailed analysis. 8-Aminoadenosine
analogues of 2-5A that occur in the anti-conformation
in the free state could not activate RNase L. In contrast,
the 8-hydroxyadenosine analogue of 2-5A has been
shown to possess strong biological activities.9 The
8-amino group of the 2-5A analogue likely forms
guanidinium type resonance cation in the 8-aminoimi-
dazole ring under the physiological condition and in-
teracts with the negatively charged 5′-phosphoryl group.
The large diversities in the RNase L binding and
activating of the 8-substituted 2-5A analogues found
in this and other previous studies demonstrate that the
bulkiness and ionic character of the 8-substituent
groups is an important factor for their biological activity.
In conclusion, a series of 8-azido- and 8-aminoadeno-
sine analogues of 2-5A were prepared easily and at
reasonable scales by polymerization of the phosphorimi-
dazolide of 8-azidoadenosine in aqueous solution with-
out employing protecting groups. The 8-azidoadenosine
analogue of 2-5A will be useful as a photoaffinity
labeling agent for the study of 2-5A-RNase L binding
and activation after labeling with 32P at the 5′-phos-
phate with kinase. The 5′-phopshoryl trimer and tet-
ramer analogues of 8-azidoadenosine have significant
RNase L binding and activating ability, in contrast to
the corresponding 8-aminoadenosine analogues which
are devoid of biological activity. Because the 8-azido-
adenosine analogues of 2-5A are relatively stable while
retaining the ability to activate RNase L, they may also
be considered candidates for experimental therapeutic
agents.
oligomer
IC50 (M)*
relative activity to pApApA
ApA-N3
ApApA-N3
>10-5
<0.0004
0.6
0.8
0.0008
2.0
2.0
<0.0004
<0.0004
0.004
<0.0004
<0.0004
0.008
1.0
7 × 10-8
5 × 10-8
5 × 10-6
2 × 10-9
2 × 10-9
>10-5
ApApApA-N3
pApA-N3
pApApA-N3
pApApApA-N3
ApA-NH2
ApApA-NH2
ApApApA-NH2
pApA-NH2
pApApA-NH2
pApApApA-NH2
pApApA
>10-5
1 × 10-6
>10-5
>10-5
5 × 10-7
4 × 10-9
3 × 10-9
pppApApA
1.3
a
Binding ability was determined by radiobinding assay. *IC50
) concentration required to inhibit binding of the 125I-labeled 2-5A
by 50%.
RNA substrate.23 The resulting RNA product was
analyzed after electrophoresis on 20% polyacrylamide/7
M urea gels, and the extent of the degradation was
measured with a phosphorimager. 5′-Monophosphoryl
2-5A tetramer, pA2′p5′A2′p5′A2′p5′A, was used as a
positive control, as human RNase L can be activated
with 2-5A with 5′-monophosphate, contrary to mouse
RNase L which requires 2-5A with 5′-di- or triphos-
phate for its activation.21 The results of the RNA
degradation with the activated RNase L are shown in
Figure 2. The cleaved RNA product was C12[32P]pCp,
because RNase L cleaves RNA between the UN
bond preferentially.24,25 The 5′-phosphoryl trimer and
tetramer 2-5A analogues of 8-azidoadenosine,
pApApA-N3 and pApApApA-N3, activates RNase L
significantly although <10-fold less effectively compared
with the unmodified 2-5A tetramer (Figure 2A). The
corresponding trimer and tetramer core, pApApA-N3
and ApApApA-N3, showed 100- to 1000-fold decrease in
activating RNase L. Therefore, the core analogues had
low level of RNase L activating ability in contrast to
the parent 2-5A core which is devoid of RNase L
activating ability. The dimers, pApA-N3 and ApA-N3,
at concentrations as high as 1 µM concentration did not
activate RNase L. The 8-aminoadenosine 2-5A ana-
logues exhibited no RNase L activating ability, although
the 5′-phosphoryl tetramer, pApApApA-NH2, was an
activator of RNase L at 1 µM concentration, a 1000-
fold reduction of 2-5A (Figure 2B). The loss of RNase
L activating ability was related to a loss of capability
to bind to RNase L.
The CD studies suggest that the 8-azidoadensoine
ananlogue of 2-5A adopts syn-conformation around the
base-glycosidyl bond in the free state in contrast to the
unsubstituted 2-5A which prefers anti-conformation.14
8-Bromo- or 8-methyladenosine analogues of 2-5A
which possess syn-conformation are also effective in the
binding and activating of RNase L.7,8,12 These bulky
8-substitutent groups in the adenosine residue make the
syn-conformation; however, they are not sufficient to
block the rotation around the base-glycosyl bond of the
2-5A analogues.26 It is reported that the 8-brom-
adenosine analogue of NAD changes its conformation
from syn in the free-state to anti when it is bound to
horse liver alcohol dehydrogenase.27 The 8-substitued
adenosine analogues of 2-5A may alter the conforma-
Exp er im en ta l Section
Ma ter ia ls a n d Sp ectr om etr ic Meth od . Adenosine-5′-
phosphate and nuclease P1 were purchased from Seikagaku
Kogyo. Venom phosphodiesterase and alkaline phosphatase
were from Worthington Biochemicals, RNase T2 from Sigma,
and T4 RNA ligase from Invitrogen. [5′-32P]-pCp (3000
Ci/mmol) was from Perkin-Elmer. DEAE-TOYOPEARL was
from Toso Co., and DEAE-Sephadex A-25 from Pharmacia.
Other reagents were obtained commercially. N-Ethylmorpho-
line was distilled before use. 8-Bromo-, 8-azido-, and 8-ami-
noadenosine 5′-monophosphate were prepared by the method
described previously.28
UV measurements were carried out on a Hitachi 3200
instrument or a Shimazu UV-1200 spectrophotometer. Con-
centrations of 2′,5′-oligonucleotides were determined spectro-
photometrically using the molar coefficients of adenosine-,
8-bromoadenosine-, 8-azidoadenosine-, and 8-aminoadenosine-
5′-monophosphate, ꢀ260 ) 15.3 × 103, ꢀ265 ) 16.7 × 103,
ꢀ281 ) 14.9 × 103, and ꢀ273 ) 16.4 × 103, respectively, after
correcting for the hypochromicity of each oligonucleotide.
Hypochromicity of the oligonucleotides was determined by the
ratio of absorbance before and after alkaline hydrolysis. CD
spectra were recorded with a J ASCO J -720 spectrometer in
10 mM phosphate buffer (pH 6.7). The CD results are
expressed in terms of molecular ellipticity per mole of nucle-
otide residue, [θ]. 1H NMR spectra were taken on a Varian
XL200 spectrometer with sodium 3-(trimethylsilyl)[2,2,3,3-2H4]-
propionate as an internal standard. ESI-mass spectra were
taken with a Perkin-Elmer Sciex API-100 instrument in a
negative mode.
HP LC. HPLC was taken with a Hitachi 638 or a J ASCO
PU880/UV870 apparatus on an RPC-5 or an ODS-silica gel
(Wakosil 5C18) column (40 × 250 mm). HPLC on an ODS-

4930 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23
Sawai et al.
F igu r e 2. RNase L activating ability of 8-azidoadenosine and 8-aminoadenosine analogues of 2-5A. (A) 8-Azidoadenosine
analogues of 2-5A. (B) 8-aminoadenosine analogues of 2-5A. 2-5A and analogues at different concentrations (as indicated)
were incubated with human recombinant RNase L on ice for 30 min followed by incubation with C7U2C12-[32P]pCp at 30 °C for 30
min. An autoradiogram of the dried gel is shown. Arrows indicate the intact and cleaved RNA. The graph represents percent
RNA cleavage as determined by phosphorimager analysis.
silica gel was carried out with a linear gradient elution of
methanol (2.4-51.2%) buffered with 10 mM triethylammo-
nium acetate (pH 7.3) in 60 min at flow rate of 1.0 mL/min.
HPLC on an RPC-5 was conducted with a linear gradient of
NaClO4 solution (0-50 mM) buffered with 2 mM Tris-acetate
(pH 7.5) and 0.1 mM EDTA in 60 min at a flow rate of 1.0
mL/min. Yields were calculated from the peak integrals of the
oligonucleotides on the HPLC chromatogram, and no correc-
tion for the hypochromicity of each oligonucleotide was made
on the yield. Preparative HPLC was carried out on an ODS-
silica gel (10 × 250 mm) with a linear gradient elution of
methanol (12-20%) buffered with 10 mM triethylammonium
acetate (pH 7.3) in 60 min at flow rate of 2.0 mL/min. Paper
chromatography was performed on Whatman 3MM paper by
a descending technique using 1-propanol-concentrated am-
monia-water (55:10:35).
several times to remove a trace of triethylammonium hydrogen
carbonate and water and then dissolved in dry DMF. Imidazole
(0.68 g, 10 mmol), triethylamine (1.0 mL), trioctylamine (0.5
mL), triphenylphosphine (0.79 g, 3.0 mmol), and 2,2′-dipyridyl
disulfide (0.66 g, 3.0 mmol) were added to the solution. The
reaction mixture was stirred for 2 h at room temperature. The
completion of the reaction was checked by cellulose-TLC
(solvent: 2-propanol-concentrated ammonia-water ) 7:2:1).
The reaction mixture was poured into a solution of acetone
(200 mL)-ether (100 mL)-triethylamine (1 mL)-acetone
saturated with NaClO4 (1 mL). The resulting white precipitate
was collected in a glass filter with a slow stream of dry
nitrogen, washed with dry acetone then dry ether several
times, and dried in a desiccator. 8-Aminoadenosine-5′-phos-
phorimidazolide was obtained in 58% yield (0.95 × 104 ODU273
,
0.58 mmol) as a sodium salt. Purity was checked by TLC and
HPLC.
P r ep a r a tion of 8-Am in oa d en osin e-5′-p h osp h or im id a -
zolid e (Im p A-NH2). 8-Aminoadenosine 5′-phoshorimidazolide
was prepared in the same way as adenosine 5′-phosphorimi-
P r ep a r a tion of 8-Azid oa d en osin e 5′-p h osp h or im id a -
zolid e (Im p A-N3). We cannot apply the conventional method
of the synthesis of adenosine 5′-phosphorimidazolide as de-
scribed above, because triphenylphosphine reacts easily with
dazolide.16 8-Aminoadenosine-5′-phosphate (1.64 × 104 ODU273
,
1.0 mmol) was coevaporated with water and then pyridine

8-Substituted 2′,5′-Oligoadenylates
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23 4931
an azido group of 8-azidoadenosine-5′-phosphate with conver-
sion into the 8-aminoadenosine-5′-phosphate. Thus we em-
ployed another method for the 5′-phosphorimidazolide syn-
thesis.17 After coevaporation of 8-azidoadenosine-5′-phosphate
(1.49 × 104 ODU281, 1.0 mmol) with water and then pyridine
several times to remove a trace of triethylammonium hydrogen
carbonate and water, dry DMF (10 mL), triethylamine (1.0
mL), trioctylamine (0.5 mL), and phenyl boric acid (200 mg,
1.6 mmol) were added and the solution was kept for 30 min
at 40 °C for blocking of the 2′,3′-dihydroxyl group of the ribose
ring by a boronate ester formation. Then, N,N-carbonyldiimi-
dazole (648 mg, 4.0 mmol) and imidazole (289 mg, 4.0 mmol)
were added to the reaction mixture, and the solution was kept
for 2 h with stirring at room temperature. The completion of
the reaction was checked by cellulose-TLC (solvent: 2-pro-
panol-concentrated ammonia-water ) 7:2:1). After the reac-
tion, a methanol solution (16.6 mL) containing ethylene glycol
(0. 5 mL) and 1,3-propandiol (1.0 mL) was added to the solu-
tion and stirred for 5 min at room temperature, then over-
night at 4 °C to remove the phenyl boronate from the
2′,3′-diol. Methanol was evaporated under reduced pressure,
and the mixture was poured into a solution of acetone (200
mL)-ether (100 mL)-triethylamin (1 mL)-acetone saturated
with NaClO4 (1 mL). The resulting white precipitate was
collected in a glass filter with a slow stream of dry nitrogen,
washed with dry acetone and then dry ether several times,
and dried in a desiccator. 8-Azidoadenosine-5′-phosphorimi-
dazolide was obtained in 61% yield (0.91 × 104 ODU281, 0.61
mmol) as a sodium salt.
Large-scale synthesis of oligomers of 8-aminoadenylates by
polymerization of ImpA-NH2 was carried out in a reaction
mixture (6.7 mL) containing 50 mM ImpA-NH2 (5490 ODU273),
25 mM uranyl nitrate in 0.2 M N-ethylmorpholine-HNO3
buffer (pH 7.0) at 25 °C for 1 d with stirring. The reaction
mixture was passed through a Chelex 100 column (Na+ form,
16 × 150 mm) to remove the uranyl ion and was concentrated
under vacuum to 2.0 mL to which acetate buffer (0.2 M, pH
5.5, 4.0 mL) and nuclease P1 solution (0.3 mg in 0.3 mL) were
added, and the reaction mixture was kept overnight at 37 °C
to degrade 3′,5′ linkage. The solution was subjected to QAE-
-
Sephadez A-25 anion column chromatography (HCO3 form,
30 × 800 mm). The elution was carried out with a linear
gradient [0.23 M (3 l) to 0.8 M (3 l)] of triethylammonium
hydrogen carbonate buffer (pH7.5) at a flow rate of 1.35 mL/
min. The fractions containing the dimer and trimer were
collected and concentrated. A portion of the dimer and trimer
fractions were further purified by paper chromatography on
Whatman 3MM paper with n-propanol:concentrated ammonia:
water (55:10:35) as an eluent using a descending technique.
The 2′,5′-linked dimer and trimer were obtained in 150 and
24 ODU273, respectively, in a purified form. The 2′,5′-linked
tetramer of 8-aminoadenylate was prepared by hydrogenation
of the tetramer of 8-azidoadenylate. The tetramer of 8-azi-
doadenylate (10 ODU281) was reduced on Pd-black (5 mg) in
0.15 mL of aqueous solution under an atmospheric pressure
of hydrogen for 2 d at room temperature. The reaction mixture
was filtered, and the filtrate was concentrated under reduced
pressure. The product was purified by preparative HPLC on
ODS-silica gel and lyophilised several times with water. The
tetramer was obtained in 58% yield (6.3 ODU273).
Cores of 2′,5′-linked oligomers of 8-azido- and 8-amino-
adenylates with a 5′-hydroxyl group were prepared by dephos-
phorylation of the 5′-phosphates of the corresponding oligomers
(2-10 ODU260) with bacterial alkaline phosphatase in nearly
quantitative yield and purified by HPLC on ODS-silica gel.
Ch a r a cter iza tion of 2′,5′-Lin k ed Oligom er s of 8-Azid o-
a n d 8-Am in oa d en yla tes. Identification of the products was
determined from the 1ESI-mass spectrum and NMR spectrum
and by enzyme digestion with snake venom phosphodiesterase,
nuclease P1, and alkaline phosphatase.
En zym e Digestion . Snake venom phosphodiesterase de-
grades both 2′,5′- and 3′-5′-linked oligonucleotides. Digestion
with venom phosphodiesterase was carried out at 37 °C for
2.5 h in a mixture (50 µL) containing the substrate (1-2
ODU260), 0.01-0.5 unit of the enzyme, 0.01 M MgCl2 in 0.02
M Tris-acetate (pH 8.8). An aliquot of the solution was
analyzed by HPLC.
Nuclease P1 degrades 3′-5′-linked oligonucleotides but is
inactive against the 2′,5′-linked oligonucleotides. The reaction
with nuclease P1 digestion was carried out at 37 °C for 1 d in
a mixture (50 µL) containing the substrate (1-2 ODU260), 1
mg of the enzyme, 2 mM EDTA, and 1/35 M Veronal-acetate
(pH 4.5). The sample was analyzed by HPLC.
Bacterial alkaline phosphatase (BAP) cleaves 5′-phosphate
of the oligonucleotides. The reaction with BAP was carried out
at 37 °C for 2.5 h in a mixture (50 µL) containing the substrate
(1-2 ODU260), 0.01 unit of the enzyme, 0.001 M MgCl2, and
0.01 M Tris-HCl (pH 8.05). The product was analyzed by HPLC
on ODS-silica gel.
RNa se L Bin d in g a n d Activa tin g Activities of 2′,5′-
Lin k ed Oligom er s of 8-Azid o- a n d 8-Am in oa d en yla tes.
Binding of the 2′,5′-linked oligomers of 8-azido- and 8-ami-
noadenylates with RNase L from mouse liver extract was
evaluated by radiobinding assay described previously using
125I-labeled 2-5A as a probe.20 Binding activity of the oliogmer
was expressed by IC50, the oligomer concentration which
requires 50% inhibition of binding of 125I-labeled 2-5A.
Assay for activation of RNase L was done using 32P-labeled
C7UUC12 as an RNA substrate.23 The synthetic RNA, C7UUC12,
which was prepared on an ABI model 380 DNA synthesizer,
was labeled at its 3′-terminus with [5′-32P]pCp (3000 Ci/mmol)
with T4 RNA ligase. Different 2-5A analogues at 1 to 1000
nM, compared with 0.1 to 1000 nM of pA4 [pA(2′p5′A)3] as a
Syn th esis of 2′,5′-Lin k ed Oligon u cleotid es of 8-Azid o-
a n d 8-Am in oa d en osin e by Ur a n yl Ion -Ca ta lyzed Oligo-
m er iza tion of Im p A-N3 a n d Im p A-NH2. Reactions were
carried out in an Eppendorf tube. The reaction mixture was
prepared on an ice bath by addition of compounds in the
following order: buffer solution, ImpA-N3 solution in water,
and uranyl nitrate solution. Samples of the mixture were
agitated vigorously and kept at 25 °C for various periods of
times. A typical reaction mixture (50 µL) contained ImpA-N3
(50 mM) and a catalytic amount of uranyl nitrate (1-10 mM)
in N-ethylmorpholine-HNO3 buffer (0.2 M, pH 7.0). The
reactions were stopped by adding 10 mL of 0.25 M EDTA
solution, and the solutions were analyzed by HPLC on an
RPC-5 column. Polymerization of ImpA-NH2 was carried out
in the same way as described above for ImpA-N3.
Large-scale synthesis of oligomers of 8-azidoadenylate by
polymerization of ImpA-N3 was carried out in a reaction
mixture (11.8 mL) containing 50 mM ImpA-N3 (8790 ODU281),
2.5 mM uranyl nitrate in 0.2 M N-ethylmorpholine-HNO3
buffer (pH 7.0) at 25 °C for 1 d with stirring. The reaction
mixture was passed through a Chelex 100 column (Na+ form,
16 × 150 mm) to remove the uranyl ion and was concentrated
under vacuum to 4.2 mL to which acetate buffer (0.2 M, pH
5.5, 4.3 mL) and nuclease P1 solution (0.3 mg in 0.3 mL) were
added, and the reaction mixture was kept overnight at 37 °C
to degrade 3′,5′ linkage. After confirming formation of a series
of 2′,5′-linked 8-azidoadenylate oligomers by HPLC, the solu-
tion was subjected to DEAE-TOYOPEARL 650 M anion
-
column chromatography (HCO3 form, 30 × 600 mm). The
elution was carried out with a linear gradient [0.15 M (3 l) to
1.0 M (3 l)] of triethylammonium hydrogen carbonate buffer
(pH7.5) containing 80% methanol at a flow rate of 1.35 mL/
min. The fractions containing UV absorption were collected.
The first fractions (3150 ODU281) contained 2′,5′-linked oligo-
mers of 8-azidoadenylate from dimer to pentamer. The second
fractions (2400 ODU281) contained the oligomers from trimer
to heptamer. HPLC analysis showed that the separation of
the oligonucleotides was not satisfactory. Thus one-fifth por-
tion of the first fraction and a quarter portion of the second
fraction were further separated by preparative HPLC on an
ODS-silica gel (10 × 250 mm), each several times. The 2′,5′-
linked dimer, trimer, tetramer, and pentamer were isolated
in 95, 161, 150, and 64 ODU281, respectively, in a purified form.
The purified oligomers were lyophilized several times with
water to remove triethylammonium acetate.

4932 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 23
Sawai et al.
(11) Charubala, R.; Pfleiderer, W.; Sobol J r, R. W.; Li, S. W.;
Suhalodonik, R. J . Chemical Synthesis of Adenylyl-(2′-5′)-
adenylyl-(2′-5′)-8-azidoadenosine, and Activation and Photo-
-affinity Labellinbg of RNase L by [32P]p5′A2′p5′A2′p5′N38A.
Helv. Chim. Acta 1989, 72, 1354-1361.
(12) Kitade, Y.; Nakata, Y.; Hirota, K.; Maki, Y.; Pabuccuoglu, A.;
Torrence, P. F. 8-Methyladenosien-substitued analogues of
2-5A: synthesis and their biological activities. Nucleic Acid Res.
1991, 15, 4103-4108.
positive control, were incubated on ice for 30 min with 0.2
µg/reaction of RNase L expressed from a baculovirus vector
in insect cells.21,22 Reaction mixtures were further incubated
with the RNA substrate, 80 nM of C7UUC12[32P]pCp, for 30
min at 30 °C. RNA was analyzed in 20% acrylamide/7 M urea
gels, and the extent of degradation was measured with a
phosphorimager (Amersham Biosciences).
(13) Nagai, K.; Kanbara, K.; Nakashima, H.; Yamamoto, N.; Tor-
rence, P. F.; Suhalodonik, R. J .; Takaku, H. Characterization
and biological activity of 8-substituted analogs of 2′,5′-oligoad-
enylates. Biochim. Biophys. Acta 1993, 1156, 321-326.
(14) van den Hoogen, Y. T.; Hilgerson, C. M. A.; Brozda, D.; Lesiak,
K.; Torrence, P. F.; Altona, C. Conformational analysis of
brominated pA2′-5′A2′-5′A analogs An NMR and model-building
study. Eur. J . Biochem. 1989, 182, 629-637.
Ack n ow led gm en t . This investigation was sup-
ported in part by a Grant-in-Aid for Scientific Research
from The Ministry of Education, Science and Technol-
ogy, J apan, to H.S., and by United States Public Health
Service Grants CA44059 from the Department of Health
and Human Services, National Cancer Institute to
R.H.S. We wish to thank Hagen Cramer (Cleveland) for
comments made during preparation of the manuscript.
(15) Shetzline, S. E.; Suhaldonik, R. J . Characterization of a 2′,5′-
Oligoadenylate (2-5A)-dependent 37-kDa RNase L. J . Biol.
Chem. 2001, 276, 23707-23711.
(16) Sawai, H.; Shibata, T.; Ohon, M. Preparation of Oligoadenylates
with 2′-5′ Linkage Using Pb2+ ion Catalyst. Tetrahedron 1981,
37, 481-485.
(17) Cramer, F.; Schatter, H.; Staab, A. Darstellung von imidazoliden
der phosphorsaure. Chem. Ber. 1961, 94, 1612-1616.
(18) Sawai, H.; Kuroda, K.; Hojo, T. Uranyl Ion as a Highly Effective
Catalyst for Internucleotide Bond Formation. Bull. Chem. Soc.
J pn. 1989, 62, 2018-2023.
(19) Sawai, H.; Kuroda, K.; Seki, J .; Ozaki, H. Conformational and
Stacking Properties of 3′-5′ and 2′-5′ Linked Oligoribonucleotides
Studied by CD. Biopolymers 1996, 39, 173-182.
(20) Sawai, H.; Taira, H.; Ishibashi, K.; Itoh, M. Synthesis and
Biological Activities of â-Alanyltyrosine Derivative of 2′,5′-
Oligoadenylate, and Its Use in Radiobinding Assay for 2′,5′-
Oligoadenylate. J . Biochem. 1987, 101, 339-346.
(21) Dong, B.; Xu, L.; Zhou, A.; Hassel, B. A.; Lee, X.; Torrence, P.
F.; Silverman, R. H. Intrinsic molecular activities of the interferon-
induced 2-5A-dependent RNase. J . Biol. Chem. 1994, 269,
14153-14158.
(22) Dong, B.; Silverman, R. H. A bipartite model of 2-5A-dependent
RNase L. J . Biol. Chem. 1997, 272, 22236-22242.
(23) Carroll, S. S.; Chen, E.; Viscount, T.; Geib, J .; Sardana, M. K.;
Gehman, J .; Kuo, L.C. Cleavage of oligoribonucleotides by the
2′,5′-oligoadenylate-dependent ribonuclease L. J . Biol. Chem.
1996, 271, 4988-4992.
(24) Wreschner, D. H.; McCauley, J . W.; Skehel, J . J .; Kerr, I. M.
Interferon action-sequence specificity of the ppp(A2′p)nA-de-
pendent ribonuclease. Nature 1981, 289, 414-417.
(25) Floyd-Smith, G.; Slattery, E.; Lengyel, P. Interferon action: RNA
cleavage pattern of a (2′-5′)oligoadenylate-dependent endonu-
clease. Science 1981, 212, 1020-1032.
(26) Saenger, W. Syn-anti conformation. In Principles in Nucleic Acid
Structure; Springer-Verlag: New York, 1984; pp 69-78.
(27) Abdallah, M. A.; Biellmann, J . F.; Nordstroem, B.; Braenden,
C. I. The conformation of adenosine diphosphoribose and 8-bro-
moadenosine diphosphoribose when bound to liver alcohol de-
hydrogenase. Eur. J . Biochem. 1975, 50, 475-481.
(28) Freist, W.; Cramer, F. Synthesis of AMP and ATP analogues.
In Nucleic Acid Chemistry, Part 2; Townsend, L. W., Tipson, R.
S., Eds.; J ohn Wiley & Sons: New York, 1978; pp 827-841.
Refer en ces
(1) Kerr, I. M.; Brown, R. E. pppA2′p5′A2′p5′A: an inhibitor of
protein synthesis synthesized with an enzyme fraction from
interferon-treated cell. Proc. Natl. Acad. Sci. U.S.A. 1978, 75,
256-260.
(2) Lengyel, P. Biochemistry of interferons and their actions. Annu.
Rev. Biochem. 1982, 51, 251-282.
(3) J ohnston, M. I.; Torrence, P. F. The role of interferon-induced
proteins, double-stranded RNA and 2′,5′-oligoadenylate in the
interferon-mediated inhibition of viral translation. In Interferon;
Friedman, R. M., Ed. Elsevier: Amsterdam/New York, 1985; Vol.
3, pp 189-298.
(4) Torrence, P. F.; Xiao, W.; Li, G.; Khamnei, S. Development of
2′,5′-oligonucleotides as potential therapeutic agent. Curr. Med.
Chem. 1994, 1, 176-191.
(5) Stark, G. R.; Kerr, I. M.; Williams, B. R. G.; Silverman, R. H.;
Schreiber, R. D. How cells respond to interferons. Annu. Rev.
Biochem. 1998, 67 227-264.
(6) J ustesen, J .; Hartmann, R.; Kjeldgaard, N. O. Gene Structure
and function of the 2′-5′-oligoadenylate synthetase family. CMLS
Cell. Mol. Life Sci. 2000, 57, 1593-1612.
(7) Lesiak, K.; Torrence, P. F. Synthesis and Biological Activities
of Oligo(8-bromoadenylates) as Analogues of 5′-O-Triphospho-
adenylyl(2′-5′)adenylyl(2′-5′)adenosine. J . Med. Chem. 1986, 29,
1015-1022.
(8) Lesiak, K.; Torrence, P. F. Purine 8-Bromination Modulates the
Ribonuclease L Binding and Activation Abilities of 2′,5′-Oligoad-
enylates. J . Biol. Chem. 1987, 262, 1961-1965.
(9) Kanou, M.; Ohomori, H.; Takaku, H.; Yokoyama, S.; Kawai, G.;
Suhaldonik, R. J .; Sobol, R W., J r. Chemical synthesis and
biological activities of analogues of 2′,5′-oligoadenylates contain-
ing 8-substittued adenosine derivatives. Nucleic Acids Res. 1990,
18, 4439-4446.
(10) Suhaldonik, R. J .; Kariko, K.; Sobol, R. W., J r.; Li, S. W.;
Reichenbach, N. L. Haley, B. E. 2- and 8-Azido Photoaffinity
Probes. Enzymatic Synthesis, Characterization, and Biological
Properties of 2- and 8-Azido Photoprobes of 2-5A and Photo-
labeling of 2-5A Binding Protein. Biochemistry 1988, 27, 8840-
8846.
J M030035K