Oligodeoxynucleotides containing C-7 propyne analogs of 7-deaza-2
'
-deoxyguanosine and 7-deaza-2
'-deoxyadenosine
Oligodeoxynucleotides containing C-7 propyne analogs of 7-deaza-2 ' -deoxyguanosine and 7-deaza-2 '-deoxyadenosine
Chris A.
Buhr
,
Richard W.
Wagner
,
Deborah
Grant
and
Brian C.
Froehler*
Gilead Sciences, 353 Lakeside Drive,
Foster City
, CA 94404,
USA
Received March 29, 1996;
Revised and Accepted June 16, 1996
ABSTRACT
The synthesis, hybridization properties and antisense activities of
oligodeoxynucleotides (ODNs) containing 7-(1-propynyl)-7-deaza-2
'
-deoxyguanosine (pdG) and 7-(1-propynyl)-7-deaza-2
'
-deoxyadenosine (pdA) are described. The suitably protected nucleosides
were synthesized and incorporated into ODNs. Thermal denaturation (
T
m) of these ODNs hybridized to RNA demonstrates an increased stability relative
to 7-unsubstituted deazapurine and unmodified ODN controls. Antisense inhibition by these ODNs was determined in a controlled
microinjection assay and the results demonstrate that an ODN containing pdG is
~
6 times more active than the unmodified ODN. 7-Propyne-7-deaza-2'-deoxyguanosine
is a promising lead analog for the development of antisense
ODNs with increased potency.
INTRODUCTION
Phosphorothioate oligodeoxynucleotides (ODNs) containing the C-5 propyne-modified pyrimidine nucleosides 5-(1-propynyl)- 2'-deoxyuridine (pdU) and 5-(1-propynyl)-2'-deoxycytidine (pdC) show
high binding affinities for single-strand RNA (
1
) and significantly enhance the antisense activity of these ODNs (
2
). The increased potency of these ODNs is the result of a decreased rate of
intracellular dissociation of the double helix and is believed to be mediated
by RNase H (
3
). The increased potency observed with ODNs containing C-5 propyne-modified pyrimidine nucleosides prompted us to explore the use of a similar propyne
modification with purine nucleosides.
7-Deaza-2'-deoxyadenosine (7-deaza-dA) and 7-deaza-2'-deoxyguanosine (7-deaza-dG)
are known nucleosides in which the 7-nitrogen is replaced by C-H (
4
-
7
). Seela and co-workers have incorporated these nucleosides into ODNs and studied their
binding affinity towards DNA (
8
-
12
). Recently, a number of 7-deaza-2'-deoxypurines substituted at the 7-position with either a halogen or a methyl group
were incorporated into ODNs and their binding affinity towards DNA determined (
13
,
14
). The results indicate that 7-substituted deazapurines can increase the binding affinity of ODNs for
DNA, but no studies have been carried out against RNA of a mixed sequence. ODNs
containing unsubstituted 7-deaza-dG and 7-deaza-2'-deoxyxanthosine have been used in the
context of triple helix recognition (
15
) and 7-alkynylamino-deazapurine nucleosides have also been synthesized and used as flourescent chain terminators in DNA sequencing (
16
-
18
).
Molecular modeling studies suggested that a propyne attached at the 7-position of 7-deaza-dG and 7-deaza-dA occupies space similar to the propyne in pdU
and pdC upon formation of a DNA/RNA double helix. We have synthesized the
appropriately protected 7-deazapurine nucleosides
1-4
(Fig.
1
) and incorporated these analogs into ODNs. The
T
m
of the DNA/RNA double helix and the antisense activities of these ODNs were
then determined. Incorporation of propyne-modified nucleosides
1
and
2
into ODNs significantly enhances the stability of the double helix with single-strand RNA. The ODN containing 7-propynyl -2'-deoxyguanosine (pdG) also leads to enhanced
antisense potency in a microinjection assay.
MATERIALS AND METHODS
General procedures
Propyne gas was purchased from Farchan Laboratories (Gainesville, FL) and all other reagents were purchased from Aldrich Chemical Co.
1
H and
31
P NMR spectra were recorded using a General Electric QE 300 spectrophotomer.
Exact masses were determined from high resolution FAB mass spectra on a VG
Analytical ZAB2-EQ spectrometer (Mass Spectrometry Laboratory, University of California, Berkeley, CA).
Nucleoside synthesis
Pyrrolopyrimidines
5
and
6
(
4
,
5
) were used to prepare nucleosides
8
and
9
by the sodium salt glycosylation method (
6
,
7
). Subsequent alkynylation (
16
,
17
) and protection led to nucleosides
17
and
22
.
4-Chloro-7-(2-deoxy-3,5-di-
O
-
p
-toluoyl-
[beta]
-D-
erythro
-pentofuranosyl)-5-iodo-2-(methylthio)pyrrolo[2,3-
d
]pyrimidine (
10
)
. To a solution of 15.0 g (27.2 mmol)
8
(
6
) in 810 ml anhydrous DMF was added 44.4 g (197 mmol)
N
-iodosuccinimide (NIS), the mixture was heated at 96oC for 6 h and cooled to room temperature. After addition of 4 ml 10%
aqeous NaHCO
3
the mixture was concentrated to ~150 ml, diluted with 300 ml ethyl acetate (EtOAc), washed with H
2
O, 5% aqeous sodium hydrosulfite (Na
2
S
2
O
4
), saturated NaCl, dried over Na
2
SO
4
and evaporated. The residue was purified by silica gel chromatography using 75-100% CH
2
Cl
2
in hexane to afford 16.85 g (91.5% yield)
10
.
1
H NMR (CDCl
3
) [delta] 7.94 (4H, m, Ph), 7.38 (1H, s, C
6
-H), 7.28 (4H, m, Ph), 6.71 (1H, m, 1'-H), 5.72 (1H, m, 3'-H), 4.8-4.5 (3H, m, 5'-CH
2
and 4'-H), 2.76 (2H, m, 2'-H
a,b
), 2.62 (3H, s, SCH
3
), 2.44 (3H, s, CH
3
), 2.43 (3H, s, CH
3
).
7-(2-Deoxy-3,5-di-
O
-
p
-toluoyl-
[beta]
-D-
erythro
-pentofuranosyl)-5- iodo-2-(methylthio)pyrrolo[2,3-
d
]pyrimidin-4-one (
11
)
. To a solution of 7.13 g (10.5 mmol)
10
in 106 ml DMF and 72 ml dioxane was added 6.47 g (53.0 mmol)
syn
-2-pyridinealdoxime and 7.22 ml (57.5 mmol) 1,1,3,3-tetramethylguanidine. The mixture was stirred at room temperature for 24 h and
concentrated. The residue was taken up in 400 ml CH
2
Cl
2
, washed with 0.1 M aqeous citric acid (2 * 250 ml), H
2
O (250 ml), saturated aqueous NaHCO
3
(250 ml), dried (Na
2
SO
4
) and evaporated. The residue was purified by silica gel chromatography using 0-10% EtOAc in CH
2
Cl
2
to afford 6.60 g (95.1% yield)
11
.
1
H NMR (CDCl
3
) [delta] 11.23 (1H, br s, NH), 7.94 (4H, m, Ph), 7.28 (4H, m, Ph), 7.06 (1H, s, C
6
-H), 6.62 (1H, m, 1'-H), 5.70 (1H, m, 3'-H), 4.73-4.53 (3H, m, 5'-CH
2
and 4'-H), 2.73 (2H, m, 2'-H
a,b
), 2.64 (3H, s, SCH
3
), 2.44 (3H, s, CH
3
), 2.42 (3H, s, CH
3
).
2-Amino-7-(2-deoxy-3,5-di-
O
-
p
-toluoyl-
[beta]
-D-
erythro
-pentofuranosyl)-5-iodopyrrolo[2,3-
d
]pyrimidin-4-one (
13
)
. To 6.60 g (10.0 mmol)
11
in 275 ml CH
2
Cl
2
at 0oC was added 4.29 g (~13.6 mmol) 3-chloroperoxybenzoic acid (MCPBA, 50-60%). After 15 min the ice bath was removed and stirring
continued for an additional 105 min. The reaction mixture was diluted with CH
2
Cl
2
, washed with saturated aqueous NaHCO
3
and dried (Na
2
SO
4
). After removal of Na
2
SO
4
by filtration, MeOH was added to the filtrate to make the solution 4% MeOH.
This solution was directly chromatographed through a short column of silica
gel, eluted with 4% MeOH in CH
2
Cl
2
and concentrated. The sulfoxide
12
was suspended in 120 ml dioxane, transferred to a Parr Bomb (stainless steel)
and saturated with anhydrous ammonia (NH
3
) at 0oC. The reaction vessel was sealed, heated at 145oC for 8.5 h, cooled and concentrated. The residue was taken up in CH
2
Cl
2
, extracted with saturated aqueous NaHCO
3
, dried (Na
2
SO
4
) and evaporated. Silica gel chromatography using 5% MeOH in CH
2
Cl
2
yielded 6.0 g impure
13
.
1
H NMR (CDCl
3
) [delta] 10.77 (1H, br s, NH), 7.95 (4H, d, J = 7.28 Hz, Ph), 7.26 (4H, m, Ph),
6.86 (1H, s, C
6
-H), 6.48 (1H, m, 1'-H), 6.10 (2H, br s, NH
2
), 5.68 (1H, m, 3'-H), 4.76-4.50 (3H, m, 5'-CH
2
and 4'-H), 2.42 (3H, S, CH
3
), 2.41 (3H, s, CH
3
).
2-Amino-7-(2-deoxy-
[beta]
-D-
erythro
-pentofuranosyl)-5-iodopyrrolo [2,3-
d
]pyrimidin-4-one (
14
)
. A suspension of 6.0 g
13
(from the previous reaction) in 120 ml saturated NH
3
/MeOH was heated in a Parr bomb at 140oC for 2 h, cooled and concentrated. The residue was taken up in H
2
O, washed twice with diethyl ether (Et
2
O), filtered (to remove solid impurities) and concentrated to a solid. The solid
was briefly heated (with a heat gun for <1 min) in 15 ml H
2
O, cooled for 2 h at 10oC, collected by filtration and washed with cold H
2
O and finally Et
2
O. After drying in a vacuum dessicator, 2.61 g (67% yield from
11
)
14
was obtained.
1
H NMR (Me
2
SO-d
6
) [delta] 10.51 (1H, br s, NH), 7.12 (1H, s, C
6
-H), 6.35 (2H, br s, NH
2
), 6.26 (1H, m, 1'-H), 5.21 (1H, d, J = 3.4 Hz, 3'-OH), 4.91 (1H, t, J = 5.2 Hz, 5'-OH), 4.27 (1H, m, 3'-H), 3.75 (1H, m, 4'-H), 3.49 (2H,
m, 5'-CH
2
), 2.32 (1H, m, 2'-H
a
), 2.06 (1H, m, 2'-H
b
).
2-Amino-7-(2-deoxy-
[beta]
-D-
erythro
-pentofuranosyl)-5-(1-propynyl) pyrrolo[2,3-
d
]pyrimidin-4-one (pdG,
15
)
. A mixture of 2.60 g (6.63 mmol)
14
, 612 mg (0.53 mmol) tetrakis(triphenylphosphine)palladium(0) [(Ph
3
P)
4
Pd(0)], 202 mg (1.06 mmol) CuI and 1.85 ml (13.3 mmol) triethylamine (TEA) in 22
ml DMF was saturated with propyne at -5oC. The reaction vessel was sealed, sonicated briefly and stirred at
room temperature. After 22 h, TLC (NH
4
OH/H
2
O/CH
3
CN, 1:5:94) showed that the reaction was incomplete and [(Ph
3
)
4
Pd(0)] (612 mg), CuI (202 mg), TEA (1.85 ml) and propyne (-5oC) were added. After stirring for an additional 19 h, MeOH (22 ml),
CH
2
Cl
2
(22 ml), Dowex 1 X 8-100 (HCO
3
-
form, 2.35 g) and Chelex-100 (sodium form, 607 mg; BioRad) were added. After stirring for 40 min the resins were removed by filtration and washed with MeOH. The combined filtrates were concentrated to a solid, 200 ml CH
2
Cl
2
was added and the mixture sonicated and cooled for 2 h at -20oC. The solid product was removed by filtration, washed with Et
2
O and dried by vacuum dessication to yield 2.11 g impure
15
that was used directly in the next reaction. Silica gel chromatography of a
small sample using 0-15% MeOH in CH
2
Cl
2
afforded a pure sample of
15
.
1
H NMR (Me
2
SO-d
6
) [delta] 10.45 (1H, br s, NH), 7.14 (1H, s, C
6
-H), 6.35-6.20 (3H, contains br s, NH
2
; and m, 1'-H), 5.20 (1H, d, J = 3.6 Hz, 3'-OH), 4.90 (1H, t, J = 5.3 Hz, 5'-OH), 4.27 (1H, s, 3'-H), 3.74 (1H, s, 4'-H), 3.48 (2H,
m, 5'-CH
2
), 2.29 (1H, m, 2'-H
a
), 2.06 (1H, m, 2'-H
b
), 1.98 (s, 3H, CCH
3
). FAB MS:
m/z
calculated for C
14
H
17
N
4
O
4
(MH
+
) 305.12498, found 305.12434.
7-(2-deoxy-
[beta]
-D-
erythro
-pentofuranosyl)-2-{[(
N
,
N
-dimethylamino) methylidene]amino}-5-(1-propynyl)pyrrolo[2,3-
d
]pyrimidin-4-one (
16
)
. To a mixture of 1.95 g (6.41 mmol)
15
in 22 ml DMF was added 0.22 ml absolute EtOH and 1.21 ml (7.06 mmol)
N
,
N
-dimethylformamide diethylacetal. After stirring for 24 h at room
temperature H
2
O (0.64 ml) was added and the reaction mixture was concentrated to an oil. The
crude
16
was used directly in the next reaction.
7-[2-Deoxy-5-
O
-(4,4
'
-dimethoxytrityl)-
[beta]
-D-
erythro
-pentofuranosyl]-2-{[(
N
,
N
-dimethylamino)methylidene]amino}-5-(1-propynyl) pyrrolo[2,3-
d
]pyrimidin-4-one (
17
)
. Crude
16
(~6.4 mmol) was evaporated from anhydrous pyridine (2*), taken up in 32 ml pyridine and 2.39 g (7.05 mmol) 4,4'-dimethoxytritylchloride (DMT-Cl) was added with stirring. After 18 h, MeOH (2
ml) was added and the mixture concentrated to an oil (~10 ml). The oil was dissolved in CH
2
Cl
2
, washed with saturated aqueous NaHCO
3
(2*), dried (Na
2
SO
4
) and evaporated. Silica gel chromatography using 0-10% MeOH in EtOAc afforded 1.25 g (30% yield from
14
)
17
.
1
H NMR (CDCl
3
) [delta] 9.04 (1H, br s, NH), 8.54 (1H, s, N=CH), 7.50-7.15 (9H, m, Ph), 6.95-6.75 (m, 5H, Ph and C
6
-H), 6.60 (1H, m, 1'-H), 4.51 (1H, br s, 3'-H), 4.08 (1H, m, 4'-H), 3.78 (6H, s, OCH
3
), 3.38 (1H, m, 5'-H
a
), 3.23 (1H, m, 5'-H
b
), 3.07 (3H, s, NCH
3
), 3.00 (3H, S, NCH
3
), 2.36 (2H, m, 2'-H
a,b
), 2.04 (3H, s, CCH
3
).
7-[2-Deoxy-5-
O
-(4,4
'
-dimethoxytrityl)-
[beta]
-D-
erythro
-pentofuranosyl]-2-{[(
N
,
N
-dimethylamino)methylidene]amino}-5-(1-propynyl) pyrrolo[2,3-
d
]pyrimidin-4-one 3
'
-(triethylammonium phosphonate) (
1
)
. Phosphitylation of 1.08 g (1.63 mmol)
17
by published methods (
19
) and purification by silica gel chromatography using 1% TEA and 0-10% MeOH in CH
2
Cl
2
afforded
1
. The product was taken up in CH
2
Cl
2
, washed with 1 M aqeous triethylammonium bicarbonate (TEAB, pH 8.3), dried over Na
2
SO
4
and evaporated from CH
3
CN (2*) and CH
2
Cl
2
(1*). This afforded 1.09 g (80.7% yield)
1
as a foam.
1
H NMR (0.5% TEA in CDCl
3
) [delta] 8.85 (1H, brs, NH), 8.69 (1H, s, N=CH), 7.50-7.15 (9H, m, Ph), 6.97 (1H, s, C
6
-H), 6.92 (1H, d, J = 616 Hz, P-H), 6.81 (4H, m, Ph), 6.65 (1H, m, 1'-H), 4.95 (1H, m, 3'-H), 4.32 (1H, m, 4'-H), 3.79 (6H, s, OCH
3
), 3.30 (2H, m, 2'-H
a,b
), 3.20 (3H, s, NCH
3
), 3.08 (3H, s, NCH
3
), 2.76 (q, J = 7.2 Hz, NC
H
2
CH
3
), 2.50 (2H, m, 2'-H
a,b
), 2.04 (3H, s, CCH
3
), 1.15 (t, J = 7.2 Hz, NCH
2
C
H
3
).
31
P NMR (0.5% TEA in CDCl
3
) [delta] 3.71 (dd,
1
J
P,H
= 616 Hz,
3
J
P,H
= 9.2 Hz).
7-[2-Deoxy-5-
O
-(4,4
'
-dimethoxytrityl)-
[beta]
-D-
erythro
-pentofuranosyl]-2-{[(
N
,
N
-dimethylamino)methylidene]amino}pyrrolo[2,3-
d
] pyrimidin-4-one 3
'
-(triethylammonium phosphonate) (
3
)
. This compound was prepared in the same manner as
1
.
1
H NMR (0.5% TEA in CDCl
3
) [delta] 8.73 (1H, s, N=CH), 8.63 (1H, br s, NH), 7.50-7.15 and 6.87-6.67 (15 H, m, Ph, C
6
-H and 1'-H), 6.94 (1H, d, J = 615 Hz, P-H), 6.53 (1H, d, J = 3.6 Hz, C
5
-H), 5.01 (1H, m, 3'-H), 4.33 (1H, m, 4'-H), 3.31 (2H, m, 5'-CH
2
), 3.21 (3H, s, NCH
3
), 3.08 (3H, s, NCH
3
), 2.67 (q, J = 7.2 Hz, NC
H
2
CH
3
), 2.55 (2H, m, 2'-H
a,b
), 1.11 (t, J = 7.2 Hz, NCH
2
C
H
3
).
31
P NMR (0.5% TEA in CDCl
3
) [delta] 3.72 (dd,
1
J
P,H
= 615 Hz,
3
J
P,H
= 9.6 Hz).
4-Chloro-7-(2-deoxy-3,5-di-
O
-
p
-toluoyl-
[beta]
-D-
erythro
-pentofuranosyl)-5-iodopyrrolo[2,3-
d
]pyrimidine (
18
)
. To a solution of 6.1 g (12 mmol)
9
(
6
) in 300 ml CH
2
Cl
2
was added 5.1 g (48 mmol) sodium carbonate (Na
2
CO
3
) and 3.9 g (24 mmol) iodine monochloride (ICl). The reaction was stirred for 21
h, washed with 0.1 M aqeous sodium hydrosulfite (Na
2
S
2
O
4
, 250 ml), saturated aqueous NaHCO
3
, dried over Na
2
SO
4
and evaporated. Silica gel chromatography using CH
2
Cl
2
afforded 6.7 g (88% yield)
18
.
1
H NMR (CDCl
3
) [delta] 8.60 (1H, s, C
2
-H), 7.95 (4H, m, Ph), 7.60 (1H, s, C
6
-H), 7.32 (4H, m, Ph), 6.79 (1H, t, J = 7.0 Hz, 1'-H), 5.75 (1H, m, 3'-H), 4.80-4.58 (3H, 5'-CH
2
and 4'-H), 2.78 (2H, m, 2'-H
a,b
), 2.44 (3H, s, CH
3
), 2.43 (3H, s, CH
3
).
4-Amino-7-(2-deoxy-
[beta]
-D-
erythro
-pentofuranosyl)-5-iodopyrrolo [2,3-
d
]pyrimidine (
19
)
. A mixture of 6.7 g (10.6 mmol)
18
in 100 ml 2 M NH
3
/MeOH was heated in a Parr bomb at 150oC for 24 h, cooled and concentrated. The residue was dissolved in hot H
2
O and washed with Et
2
O. The organic layer was washed with H
2
O and the combined aqueous phases evaporated. The product was dissolved in hot H
2
O (~75 ml), cooled and maintained at 5oC for 18 h. The solid was filtered, washed with cold acetone and dried
by vacuum dessication to yield 925 mg (60%)
19
.
1
H NMR (Me
2
SO-d
6
) [delta] 8.10 (1H, s, C
2
-H), 7.66 (1H, s, C
6
-H), 6.68 (2H, br s, NH
2
), 6.49 (1H, m, 1'-H), 5.26 (1H, d, J = 3.9 Hz, 3'-OH), 5.04 (1H, t, J = 5.4 Hz, 5'-OH), 4.33 (1H, br s, 3'-H), 3.81 (1H, m, 4'-H), 3.53 (2H, m
5'-CH
2
), 2.46 (1H, m, 2'-H
a
), 2.15 (1H, m, 2'-H
b
).
4-Amino-7-(2-deoxy-
[beta]
-D-
erythro
-pentofuranosyl)-5-(1-propynyl) pyrrolo[2,3-
d
]pyrimidine (pdA,
20
)
. The procedure of Hobbs (
17
) was used to prepare
20
from propyne and
19
in DMF at room temperature. Silica gel chromatography using 2-10% MeOH in CH
2
Cl
2
afforded an 80% yield of
20
.
1
H NMR (CD
3
OD) [delta] 7.97 (1H, br s, C
2
-H), 7.38 (1H, s, C
6
-H), 6.36 (1H, dd, 1'-H), 4.40 (1H, m, 3'-H), 3.88 (1H, m, 4'-H), 3.65 (2H, m, 5'-CH
2
), 2.52 (1H, m, 2'-H
a
), 2.21 (1H, m, 2'-H
b
), 1.98 (3H, s, CCH
3
). FAB MS:
m/z
calculated for C
14
H
17
N
4
O
3
(MH
+
) 289.13007, found 289.13002.
7-(2-Deoxy-
[beta]
-D-
erythro
-pentofuranosyl)-4-(
N
,
N
-dibenzoylamino)- 5-(1-propynyl)pyrrolo[2,3-
d
]pyrimidine (
21
)
. This compound was prepared from
20
using the general method of Jones (
20
).
1
H NMR (CDCl
3
) [delta] 8.48 (1H, s, C
2
-H), 7.86 (4H, m, Ph), 7.58-7.32 (6H, m, Ph), 7.26 (1H, s, C
6
-H), 6.31 (1H, m, 1'-H), 4.70 (1H, d, J = 5.3 Hz, 3'-H), 4.14 (1H, d, J = 1.2 Hz, 4'-H), 3.91 (1H, dd, J = 12.5 Hz, J =
1.9 Hz, 5'-H
a
), 3.75 (1H, dd, J = 12.5 Hz, J = 1.9 Hz, 5'-H
b
), 2.96 (1H, m, 2'-H
a
), 2.28 (1H, m, 2'-H
b
), 1.35 (3H, s, CCH
3
).
7-[2-Deoxy-5-
O
-(4,4
'
-dimethoxytrityl)-
[beta]
-D-
erythro
-pentofuranosyl]-4-(
N
,
N
-dibenzoylamino)-5-(1-propynyl)pyrrolo[2,3-
d
]pyrimidine (
22
)
. This compound was prepared from
21
in the same manner used for the preparation of
17
.
1
H NMR (CDCl
3
) [delta] 8.47 (1H, s, C
2
-H), 7.86 (4H, m, Ph), 7.58 (1H, s, C
6
-H), 7.5-7.1 (15H, m, Ph), 6.82 (4H, m, Ph), 6.70 (1H, m, 1'-H), 4.60 (1H, m, 3'-H), 4.09 (1H, m, 4'-H), 3.77 (6H, s, OCH
3
), 3.38 (2H, m, 5'-CH
2
), 2.55 (1H, m, 2'-H
a
), 2.44 (1H, m, 2'-H
b
), 1.34 (3H, s, CCH
3
).
U, 5-propynyl-2'-deoxyuridine; C, 5-propynyl-2'-deoxycytidine.
Synthesis, purification and
T
m analysis of oligodeoxynucleotides
All ODNs were synthezised on a Milligen-Biosearch 8750 using the H-phosphonate method and converted to phosphorothioate linkage via S
8
oxidation (
19
). The ODNs were purified by PAGE and desalted using a NAP-25 column.
T
m
measurements were obtained at 280 nm in 140 mM KCl/5 mM Na
2
HPO
4
at pH 7.2 as previously described (
1
). 7-Deaza-2'-deoxyguanosine analogs were incorporated into ODNs of the sequence 5'-
G
U
G
-
G
CU-
GGG
-CU
G
-UUC-3'; these ODNs are complementary to the RNA sequence 5'-GAA-CAG-CCC-AGC-CAC-3' in the
SV40 T Antigen (TAg) coding region (Table
1
). 7-Deaza-dA analogs were incorporated into ODNs of the sequence 5'-UC
A
-UC
A
-G
A
G-G
AA
-U
A
U-UCC-3'; these ODNs are complementary to the RNA sequence 5'-GGA-AUA-UUC-CUC-UGA- UGA-3' in
the TAg coding region (Table
1
). The following extinction coefficients ([Sigma]
260
) were used to calculate the [Sigma]
260
of the ODNs: dA (15 200), dG (11 800), pdU (3200), pdC (5000), pdG (10 500),
pdA (7300), 7-deaza-dG (10 000) and 7-deaza-dA (10 600).
Antisense assays
Microinjection, immunofluorescence staining and fluorescence microscopy were
carried out as described (
2
,
3
). Antisense inhibition was determined by co-microinjection of two expression plasmids (SV40 TAg, derived from p5080, and
Escherichia coli
[beta]-galactosidase, pRSVZ) and ODN into CV1 cells (
2
,
3
). The cells were incubated for 4.5 h at 37oC, fixed, immunolabeled and scored visually by fluorescence microscopy for TAg and [beta]-galactosidase expression. Each experiment was repeated in triplicate
and IC
50
values are +-5%. Inhibition of TAg was gene specific (no inhibition of [beta]-galactosidase) at concentrations up to 5 [mu]M.
RESULTS AND DISCUSSION
Nucleoside synthesis
Deazapurine oligodeoxynucleotides
These nucleoside analogs were incorporated into two separate ODN sequences, the 2'-deoxyguanosine analogs were incorporated into a 15mer and the 2'-deoxyadenosine analogs were incorporated into an 18mer (Table
1
), both sequences complementary to the coding region of TAg RNA. These ODNs were
used to assess both binding affinity and antisense activity, therefore, the
nuclease-stable phosphorothioate and the high affinity pdU and pdC modifications
were used for all ODNs.
Biochemical characterization of modified oligodeoxynucleotides
The binding affinity of the phosphorothioate ODNs containing the 7-deazapurine analogs to the complementary RNA strand, as measured by
thermal denaturation (
T
m
), are shown in Table
1
. The results demonstrate that the 7-propynyl analogs bind with higher affinity to RNA than the control
purines. Both the deoxyadenosine and deoxyguanosine derivative increased the
T
m
by 0.8oC/substitution relative to the control ODNs. The increase in binding
affinity is likely due to increased stacking interactions leading to more
favorable enthalpy of binding. Previous studies with 7-substituted deazaadenosine analogs have shown that substitution leads to a
more unfavorable entropy of binding that can then be compensated for by
increased enthalpy of binding (
13
). The unsubstituted 7-deazapurine analogs bound with lower affinity to RNA than the control; 7-deazaguanosine-substituted ODN led to a decrease in
T
m
of 0.6oC/substitution and 7-deazaadenosine substitution led to a decrease of 0.25oC/substitution. Our results with RNA as the target strand are consistent with
previous reports with DNA as the target strand, indicating that there are no
stacking interactions to compensate for the more unfavorable entropy of binding
(
13
,
14
).
These phosphorothioate ODNs were also assessed for their ability to inhibit gene
expression via an antisense mechanism of action. Both sequences shown in Table
1
are complementary to the coding strand of TAg RNA and are downstream of the
start codon. Antisense inhibition of TAg was assessed via a microinjection
assay with
E.coli
[beta]-galactosidase as an internal control. The results shown in Table
1
indicate that the antisense activities of the deoxyguanosine (dG) and
deoxyadenosine (dA) analogs follow very different trends. Replacement of the 2'-deoxyguanosines in ODN
23
with 7-deaza-dG (ODN
24
) results in a decrease in IC
50
from 1.0 to 0.6 [mu]M (~2-fold increase in activity) and replacement with 7-propynyl-dG (ODN
25
) decreases the IC
50
further to 0.15 [mu]M (~6-fold increase in activity). Conversely, replacement of the 2'-deoxyadenosines in ODN
26
with 7-deaza-dA (ODN
27
) resulted in an increase in the IC
50
from 0.5 to 1.5 [mu]M, while replacement with 7-propynyl-dA (ODN
28
) led to complete loss of antisense activity at a concentration of up to 5 [mu]M.
It has been shown that propyne substitution of deoxyuridine and deoxycytidine
enhances the antisense activity of phophorothioate ODNs and it was expected
that a similar trend would be observed with these purine analogs. 7-propynyl-dG does enhance the antisense activity of ODNs, but 7-propynyl-dA decreases the activity. This effect cannot be
explained by binding affinity of the ODNs, as both analogs increase the
affinity for the target RNA by the same degree. The reason for this dichotomy,
at this time, is unclear. It is possible that sequence context may play a role
in the difference in activity of these two purine analogs. Within the sequence
used for the dG analogs, five of the seven dGs are adjacent to other dGs and
the remaining two dG analogs are adjacent to propynyl pyrimidines. This
sequence context should favor stacking interactions between adjacent propynyl
groups. Conversely, the sequence used for the dA analogs contains only one AA
pair, with three of the six remaining dA analogs adjacent to unmodified dG.
This context would minimize favorable stacking interactions. Additionally,
there may be a difference in ability to recruit RNase H cleavage of the
complementary RNA by the two different ODNs. Continued experimentation will be
required to determine if either of these two factors influence the antisense activity of these ODNs or if another unknown variable is at work.
These antisense experiments were carried out by nuclear microinjection of the
ODN into cells, due to the lack of cellular permeation observed with
phosphorothioate ODNs (
2
,
28
-
31
). It has been shown that ODNs require complexation with cationic lipids to be
efficiently delivered to the nuclear compartments of cells (
2
,
28
-
32
). Substitution of the ODNs with 7-propynyl-dG does not enhance cellular uptake of these ODNs, addition of 5 [mu]M ODN to the medium resulting in no antisense inhibition with
all ODNs tested (data not shown).
In summary, suitably protected derivatives of 7-propynyl-dG and 7-propynyl-dA have been prepared and incorporated into ODNs. Thermal
denaturation of the phosphorothioate ODN/RNA double helix demonstrates that both analogs
increase binding affinity relative to unmodified purines. The deoxyguanosine
analog dramatically enhances antisense activity of the ODN, whereas the
deoxyadenosine derivative decreases the activity of the ODN. The 7-propynyl-2'- deoxyguanosine analog may be a valuable lead for
increased biological potency of antisense ODNs.
ACKNOWLEDGEMENTS
We would like to thank Terry J.Terhorst, Teresa Huang and Jason G.Lewis for
expert technical assistance. This research was supported in part by a grant
from the Advanced Research Projects Agency (ARPA).
