Oligodeoxynucleoside phosphoramidates (P-NH
2
): synthesis and thermal stability of duplexes with DNA and RNA targets
Oligodeoxynucleoside phosphoramidates (P-NH 2 ): synthesis and thermal stability of duplexes with DNA and RNA targets
Suzanne
Peyrottes
,
Jean-Jacques
Vasseur*
,
Jean-Louis
Imbach
and
Bernard
Rayner
Laboratoire de Chimie Bio-Organique, UMR 5625 du CNRS, CC 008, Université Montpellier II, Place Eugène Bataillon, 34095
Montpellier
Cedex 5,
France
Received February 16, 1996;
Revised and Accepted April 1, 1996
ABSTRACT
Syntheses of non ionic oligodeoxynucleoside phosphoramidates (P-NH
2
) and mixed phosphoramidate-phosphodiester oligomers were accomplished on automated solid supported DNA synthesizer using both H-phosphonate and phosphoramidite chemistries, in combination with
t
-butylphenoxyacetyl for N-protection of nucleoside bases, an oxalyl anchored solid support and a final treatment with methanolic ammonia. Thermal stabilities of the hybrids formed between these new
analogues and their DNA and RNA complementary strands were determined and compared with those of the corresponding unmodified oligonucleotides, as well as of the phosphorothioate and methylphosphonate derivatives. Dodecathymidines containing P-NH
2
links form less stable duplexes with DNA targets, d(C
2
A
12
C
2
) (
[Delta]
Tm/modification -1.4
o
C) and poly dA (
[Delta]
Tm/modification -1.1
o
C) than the corresponding phosphodiester and methylphosphonate analogues, but the hybrids are slightly more stable than the one obtained with phosphorothioate derivative. The destabilization is more pronounced with poly rA as the target (
[Delta]
Tm/modification -3
o
C) and could be compared with that found with the dodecathymidine methylphosphonate. The modification is less destabilizing in an heteropolymer-RNA duplex (
[Delta]
Tm/modification -2
o
C). As expected, the P-NH
2
modifications are highly resistant towards the action of various nucleases. It is also demonstrated that an all P-NH
2
oligothymidine does not elicit
Escherichia coli
RNase H hydrolysis of the poly rA target but that the modification may be
exploited in chimeric oligonucleotides combining P-NH
2
sections with a central phosphodiester section.
INTRODUCTION
The use of natural oligonucleotides as therapeutic agents in an antisense
approach in which the target is RNA or DNA, suffers from drawbacks such as
their inherent instability towards degradation by extra and intra-cellular nucleases and their relatively poor cellular uptake (
1
). Attempts to produce nuclease resistant oligonucleotides and to improve their cellular uptake while keeping the specificity of binding
to their targets, has resulted in the preparation of several phosphodiester
backbone modified analogues (
2
,
3
). Of these, phosphorothioates (
4
) are the most widely applied derivatives of oligodeoxynucleotides for an
antisense approach, but they remain somewhat susceptible to degradation by nucleases (
5
), are taken up by cells more slowly than phosphodiesters and present a strong
tendency to bind with numerous proteins (
6
). Replacement of the negatively charged oxygen atom in the phosphodiester
backbone by uncharged groups to produce non-ionic oligonucleotide analogues such as methylphosphonates (
7
) or phosphoramidates (
8
) would allow them to enter cells by an alternative mechanism and confer increased nuclease resistance. However, methylphosphonates suffer from the disadvantage of being relatively
insoluble in water, aqueous buffers and biological media (
9
). Furthermore, these modifications induce chirality and the binding properties
of the resulting diastereoisomers to a complementary target are clearly
dependent on the orientation of the substituted groups around the phosphorus
atom (
9
-
11
). Despite the effort directed towards the stereocontrolled synthesis of phosphorothioates (
12
) and methylphosphonates (
13
), mixtures of heterogeneous oligomers are largely used which lower the binding to the
target (
14
). To bypass the chirality problem, a number of non-phosphate, neutral and achiral linkages have been prepared (
15
,
16
). However, the syntheses of such backbones are in general difficult and
versatile. Reported preparations are not particularly suitable for automated
synthesizers to the level at which all linkages are modified. The most frequently used strategy is to prepare
a modified nucleoside dimer, to convert it in a phosphoramidite synthon and to
incorporate it into an oligonucleotide sequence where modified linkages are
alternated with normal phosphodiesters (
15
). In such oligomers, the polyanionic character is reduced, but not suppressed
and these modifications do not necessarily provide increased cellular uptake as
well as a total nuclease resistance. For that reason and although the automatic
synthesis of non-ionic phosphoro oligonucleosides cannot be rendered stereospecific for the
time being, these phosphoro analogues stay attractive because they can be
prepared in a relatively straightforward manner by minor modifications to the existing
automated DNA synthetic methods.
If it has been clearly demonstrated that the negative impact of stereoisomerism on hybridization is enhanced with bulky substituents (
8
). It is assumed that the influence of a modification on the stabilities of
duplexes also results from the disturbance of the spin of hydration around the
modified phosphate groups (
8
,
17
). The amides of phosphorus acid are known to have good solvating properties (
18
). Furthermore, in phosphoramidates (P-NH
2
), the fact that the NH
2
group is small, uncharged in aqueous solution and forms hydrogen bonds readily
(
19
), gives them favourable solubility characteristics required for hydration, with
a minimum steric hindrance around the phosphorus atom, and makes them attractive potential candidates for non-anionic analogues of oligonucleotides. The syntheses of
N
-alkyl (
8
,
20
) and
N
-alkoxy (
21
) phosphoramidate derivatives of DNA have been described and their biophysical
properties studied. Unfortunately, due to the reported decomposition of dithymidine phosphoramidate under the base conditions
necessary for removal of the common heterocyclic
N
-acyl protecting groups and release of an oligonucleotide from the regular succinyl anchored solid support (
20
) (concentrated ammonia, 55oC, 5-16 h), syntheses of phosphoramidate (P-NH
2
) oligonucleotides have been limited to a dimer level (
19
,
20
,
22
). However, recent developments in oligonucleotide chemistry on a more labile
oxalyl anchored solid support (
23
) combined with the use of monomers possessing highly labile protecting groups (
24
), of mild and rapid deprotection conditions (
25
) have enabled us to synthesise these base-sensitive oligonucleotides. In this regard, the use of pent-4-enoyl (
26
) as nucleobase protective group (
27
) appears very promising, and recently allowed the synthesis of various
dinucleosides containing a P-NH
2
linkage (
28
). We would like to report here, the first synthesis of partially or fully
modified oligonucleotide phosphoramidates (P-NH
2
) of length sufficient to explore the antisense potential of these analogues.
The nuclease resistance properties of the modified oligonucleotides and the
thermal stabilities of duplexes formed with their DNA or RNA complements have
been evaluated and compared with those of unmodified oligonucleotide, as well
as those of the phosphorothioate and methylphosphonate analogues.
MATERIALS AND METHODS
Oligonucleotide synthesis
Syntheses of the oligodeoxynucleotide analogues were carried out on 1 [mu]mol scale using a 381A Applied Biosystems DNA-synthesizer. First, nucleoside was linked to LCAA-CPG (500 Å, from Sigma) by means of an oxalyl linker (
23
). Analogues combining phosphodiester and phosphoramidate internucleoside links
were prepared by using both phosphoramidite and hydrogen phosphonate
chemistries. After each coupling step using commercially available nucleoside-cyanoethyl phosphoramidites and appropriate reagents (from Millipore), the intermediate phosphite triesters were
oxidized by a 1 min treatment with a 1.1 M solution of
t
-butylhydroperoxide in dichloromethane. H-phosphonate couplings were carried out using deoxynucleoside H-phosphonate derivatives (
24
) and standard reagents from Glen Research. At the end of the chain elongation,
oxidative amidation was performed manually by treating CPG linked
oligonucleotide with a saturated solution of ammonia in carbon
tetrachloride/dioxan (4/1, v/v) for 30 min at 0oC (
20
). After filtration, the oligomers were released from the solid support and
deprotected with a treatment with saturated methanolic ammonia at 20oC over 2 h for phosphodiester/phosphoramidate oligomers
1
,
2
,
3
and
7
, and 5 min for the fully modified P-NH
2
oligomers
4
and
6
(Table
1
).
HPLC methods
HPLC was performed on a Waters-Millipore instrument equipped with two M510 solvent delivery systems, a
M680 solvent programmer, a U6K injector and a M990 diode array UV detector.
Oligonucleotides were purified by reverse-phase HPLC on a Delta-pack preparative column (7.8 * 300 mm, C18, 5 [mu]m, Millipore) using a gradient of acetonitrile from 10 to
25% in 0.05 M triethylammonium acetate buffer (pH 7) in 40 min at a flow rate of 2 ml/min.
The desired fractions were combined, evaporated, dissolved in water and
lyophilised. Purity of the samples was checked by HPLC at 260 nm on a Nucleosil analytical column (4.6 * 150 mm, C18, 5 [mu]m, Macherey-Nagel) using a gradient of acetonitrile from 0 to 30% in 0.05 M
triethylammonium acetate buffer (pH 7) in 30 min at a flow rate of 1 ml/min.
Using this procedure, 7-20 A
260
units of purified oligomers were isolated in each case.
Mass spectrometry
Mass spectra were recorded on a SSQ 7000 quadrupole mass spectrometer (Finnigan
MAT, San Jose, USA) fitted with an electrospray interface. Samples (final
concentration 30 [mu]M) were dissolved in water-methanol (v/v, 1:1) containing triethylamine (1.5%) and introduced
into the mass spectrometer at a 10 [mu]l/min flow-rate with a Harvard Apparatus 22 pump model. The negative ion
electrospray mass spectra were transformed into real mass spectra, using the
DEC-Finnigan software.
NMR measurements
31
P-NMR spectra were recorded on a Brüker AC 250 spectrometer at 100 MHz. Samples were dissolved in D
2
O. Chemical shifts values are in p.p.m. relative to external 85% H
3
PO
4
.
Formic acid mediated degradation of d(ACACCCAATTCT) analogues (29)
One A
260
unit of oligonucleotide was dissolved in 90% aqueous formic acid (1 ml) and the
resulting solution was heated at 120oC for 12 h. After cooling, the solution was evaporated to dryness under
reduced pressure, and the residue was dissolved in 0.1 M triethylammonium
acetate buffer, pH 7 (0.2 ml) for HPLC analysis (Nucleosil analytical column,
4.6 * 150 mm, C18, 5 [mu]m, Macherey-Nagel, eluent: 0.05 M triethylammonium acetate, pH 6.9, flow rate of 1 ml/min). The base composition of the
oligonucleotides
6
,
7
and
8
was confirmed after applying this degradation, which cleanly liberated the
purine and pyrimidine bases (Cyt, 3.03 min; Thy, 7.59 min; Ade, 17.63 min) and
quantification of the peak areas at 260 nm, using an equimolar mixture of the
three bases as standard.
Melting temperatures
Melting curves were recorded on a UVIKON 931 spectrophotometer (Kontron). The temperature control was through a HUBER PD 415 temperature programmer connected to a refrigerated water bath (Huber
Ministat). Typical experiments were carried out with equimolar ratio of modified oligomers and their targets in a 10 mM sodium cacodylate buffer (pH 7 or pH 5.45) containing 10 mM, 100 mM or 1 M sodium chloride. The samples were preheated at 75oC, and the change in absorbance at 260 nm as a function of the temperature
was recorded. The cooling or heating rate was 0.5oC/min. The cell compartment was continuously flushed with dry nitrogen for
temperatures below room temperature. Tm values were determined from the maxima
of the first derivative plots of absorbance versus temperature.
Enzymatic hydrolysis experiments
Snake venom phosphodiesterase (SVPDE) (
Crotalus durissus
), Nuclease S1 (
Aspergillus oryzae
) and Calf spleen Phosphodiesterase II (CSPDE) were purchased from Boehringer. The oligothymidines (2 A
260
U) were incubated at 37oC in either: 10 [mu]l 50 mM sodium acetate buffer (pH 4.5), 300 mM sodium chloride, 100 mM
zinc acetate added with 70 [mu]l of water containing 2 U nuclease S1; 100 [mu]l of 100 mM Tris-HCl buffer (pH 9), 10 mM magnesium chloride added with 2 [mu]l of the commercial solution of SVPDE (3 U/ml); or 80 [mu]l of 125 mM ammonium acetate buffer (pH 6.8), 2.5 mM EDTA
containing 2 [mu]l of the commercial solution of CSPDE (0.5 U/ml). Aliquots were analysed
using the analytical conditions described in HPLC methods.
RNase H digestion of poly rA/oligothymidine analogues duplexes
Escherichia coli
RNase H was purchased from Pharmacia. These experiments were performed with
oligothymidylate analogues and poly rA (240 [mu]M in nucleotide concentration for each strand) in 10 mM Tris-HCl buffer (pH 7), 10 mM magnesium chloride and 100 mM sodium
chloride (500 [mu]l). The samples were preheated at 70oC, allowed to slowly cool down to 0oC, then the temperature was stabilised at 20oC before adding the enzyme (5 U). Digestion curves were obtained by plotting the absorbance at 260 nm as a function of time.
RESULTS AND DISCUSSION
Preparation of oligonucleosides containing phosphoramidate (P-NH
2
) linkages
Standard synthesis of oligonucleotides anchored by a succinyl linker to
controlled pore glass (CPG) support, requires an ultimate chemical step with
concentrated ammonium hydroxide at 55oC for 5-16 h to release the oligonucleotides from the solid support and completely remove nucleobase and phosphodiester protecting groups. Although a number of
N
-alkylphosphoramidate oligonucleotide derivatives (
8
) have been prepared using this current methodology, the synthesis of less
hindered P-NH
2
analogues has been limited to the preparation of dinucleotides (
19
,
20
,
22
) or trinucleotide (
28
). This is due to the fact that the P-NH
2
linkage could not survive the drastic ammonia treatment. Indeed, it has been
shown that a dithymidine phosphoramidate (P-NH
2
) was rapidly hydrolysed under such alkaline treatment (t
1/2
15 min) (
20
) affording a mixture of thymidine and corresponding 3'- and 5'-phosphoramidic acid monoesters (
22
,
30
). Fortunately, we found that a treatment with saturated methanolic ammonia gave appreciably less
degradation of this dimer (t
1/2
3.8 days). This mild treatment has been used to deprotect base-sensitive oligonucleotides such as RNA and methylphosphonates (
24
) as well as oligonucleoside
N
-alkoxyphosphoramidates (
21
). Under these conditions, the release of an oligonucleotide from an oxalyl
anchored solid support is quantitative after 5 min at room temperature (
23
), and the complete removal of cyanoethyl phosphate protecting groups is
achieved within 2 h at room temperature (
21
,
24
). We estimated that <1.5% of a P-NH
2
linkage will be degraded (after 2 h treatment with methanolic ammonia) on an
oligonucleotide containing both P-NH
2
and P-O
-
(initially protected with cyanoethyl) linkages, which is quite acceptable.
To examine the effect of the modification on binding capacities, several
analogues of dodecathymidylate
5
were prepared (Table
1
). Oligomer
1
was synthesized with one modification placed in the middle of the sequence. In
compound
2
, modified linkages were alternated with normal phosphodiesters, in such a way
that the oligomer had six phosphoramidate and five phosphodiester internucleoside bonds.
In the oligothymidine
3
, an internal part of five adjacent phosphodiesters was surrounded by two
terminal sections of three contiguous phosphoramidates. In dodecathymidine
4
, natural phosphodiester linkages were completely replaced by phosphoramidate
ones. In addition, two analogues of the mixed-base dodecanucleoside
8
complementary to the splice acceptor site of mRNA coding for HIV-1 tat protein (
31
) were synthesized (Table
1
). In the oligomer
7
, as for compound
3
, an internal part of five adjacent phosphodiesters was surrounded by two
terminal sections of three contiguous phosphoramidates. Finally,
oligonucleotide
6
was fully modified. In order to obtain phosphodiester and phosphoramidate linkages in the same oligomer, cyanoethyl
phosphoramidite and H-phosphonate chemistries were respectively employed. As it has been
demonstrated that
t
-butylhydroperoxide is able to selectively oxidize phosphite triester
intermediates into phosphate triesters without affecting H-phosphonate linkages (
32
), this reagent was used instead of the common iodine treatment after each
phosphoramidite coupling step. Then, the H-phosphonate diesters were oxidized into phosphoramidate linkages, by
treatment with a saturated solution of ammonia in dioxan/CCl
4
(
20
). As classical amino protecting groups are routinely removed under drastic
conditions (concentrated aqueous ammonia, 5-16 h at 55oC),
N
-
t
-butylphenoxyacetyl protected nucleoside phosphoramidites and H-phosphonates (
24
) were employed during the synthesis of heteropolymers
6
and
7
, in combination with
t
-butylphenoxyacetic anhydride as capping reagent. In our hands and in
contrast to what has been reported in the literature (
24
), the coupling efficiency of the H-phosphonate salts of deoxyadenosine and deoxycytidine was much lower (<90%) than the one of the corresponding phosphoramidites (>98%) and may
account for the low yields of purified compounds (Table
2
). Finally, the oligonucleotides were deprotected and removed from the solid
support by treatment with saturated methanolic ammonia at room temperature (2 h
for oligomers
1
,
2
,
3
and
7
; 5 min for compounds
4
and
6
).
The chromatograms of
3
and
4
(Fig.
1
A) are representative analyses of the products released from the solid support
after the ammonia treatment. The oligonucleotides were purified by reverse
phase HPLC. Figure
1
B depicts the chromatograms of compounds
3
and
4
obtained after purification. Owing to the absence of charge on phosphoramidate
linkages, retention times of those analogues on reverse phase HPLC increase
with the number of modifications (Table
2
).
Base-pairing of oligonucleotide analogues with DNA and RNA complementary
strands
Interactions between dodecathymidylate analogues with
9
, poly dA and poly rA were investigated by both thermal denaturation and
renaturation analyses using changes in absorbance at 260 nm versus temperature.
The experiments were carried out at an equal nucleotide concentration (60 [mu]M) of the thymidylate analogues and the target compounds. At the end of the
experiment, a sample was examined by HPLC to confirm that no significant degradation of the modified
oligonucleotide had occurred. Melting temperatures are presented in Table
3
. No differences were observed between thermal association and dissociation
curves. It can be seen from these results that the introduction of
phosphoramidate internucleoside linkages into the dodecathymidylate affects the thermal stability of the duplexes formed with either DNA or RNA complementary strands.
Thus, incorporation of one, six and eleven modifications leads to a progressive decrease in the melting temperature of the complexes compared to the `parent' unmodified duplexes. Tm value is almost a linear function of the number of modifications.
The average destabilisation is 1.4oC and 1.1oC per modification in duplexes obtained with
9
and poly dA, respectively. As generally observed for oligo (dT) and its
analogues (
32
), the affinity for duplex formation with poly rA is lower than that observed
with poly dA. However, in this case, the difference of stability between the
oligothymidine-poly dA and the oligothymidine-poly rA hybrids is enhanced (1.8oC) when compared with natural complexes (0.2oC). Moreover, the average decrease in Tm of ~2.9oC/modification compared with (dTp)
11
dT-poly rA hybrid, did not allow us to observe the binding of the uniformly
modified (dTpn)
11
dT with poly rA, indicating an appreciable distortion of the structure of the
modified oligothymidine-RNA duplex compared with the `parent'duplex. Nevertheless, at a four
times higher nucleotide concentration (240 [mu]M), Tm values of 16oC and of 38oC were observed for the modified,
4
, and the natural hybrids ([Delta]Tm/mod = -2oC), respectively. So, even if the modification is really
destabilizing, one can admit that it does not compromise Watson-Crick base pairing with a ribonucleotide complementary strand.
Furthermore, the Tm value (13.3oC) of the duplex formed between oligonucleotide
3
and poly rA supports this conclusion. Indeed, this Tm value cannot be
attributed only to the binding with poly rA of the internal non-modified section (dTp)
5
dT of the oligomer
3
under the same conditions, no transition above 0oC was observed with (dTp)
5
dT and poly rA.
.
Thermal stabilities (oC) of duplexes formed by oligonucleoside phosphoramidates
a
Oligomer
Complementary strand
9
poly dA
poly rA
Tm
[Delta]Tm/mod
Tm
[Delta]Tm/mod
Tm
[Delta]Tm/mod
5
27.0
-
31.5
-
29.5
-
1
25.2
-1.8
nd
26.4
-3.1
2
17.7
-1.6
24.2
-1.2
11.3
-3.0
3
18.9
-1.4
nd
13.3
b
-2.7
4
13.0
-1.3
20.5
-1.0
<0
c
<-2.7
average [Delta]Tm/mod
-1.4
-1.1
-2.9
a
Experiments were carried out in a buffer (pH 7) containing 10 mM sodium
cacodylate, 0.1 M sodium chloride, at 60 [mu]M nucleotide concentration in each strand.
b
In the same experimental conditions, no transition was observed for (dTp)
5
dT with poly rA (Tm <0oC).
c
In 10 mM Tris-HCl, 10 mM MgCl
2
, 0.1 M NaCl buffer solution (pH 7) at a nucleotide concentration of 240 [mu]M, a Tm value of 16oC was observed for fully modified oligomer
4
with poly rA target and a Tm value of 38oC was found for the natural duplex.
Experiments at different NaCl concentrations (Table
4
) show that the Tm value for the hybrid involving the fully modified
oligonucleotide d(Tpn)
11
dT
4
and
9
is weakly modified by changes in the ionic strength of the medium either at pH
7 or 5.45. This behaviour most likely reflects the absence of charge repulsion
between a non-ionic oligonucleotide and the negatively charged phosphodiester backbone
of the target strand as observed for other neutral backbone analogues (
8
). Consequently, the modification P-NH
2
is not protonated between pH 5.45 and 7. This result is in accord with the
known low basicity of phosphoric amide diesters (
18
,
33
,
34
). In contrast, the stability of the `parent' duplex d(Tp)
11
dT-
9
increases with increasing NaCl concentration as expected. Thus, under low salt
conditions (10 mM NaCl), the modified duplex is as stable as the `parent' one
(Table
4
).
.
Salt concentration dependence of the Tm (oC) of the fully modified phosphoramidate oligomer with
9
compared with the natural duplex
Oligomer
NaCl concentration
0.01 M
0.1 M
1 M
5
pH 7
15
27
42
4
pH 7
13
13
16
4
pH 5.45
14
14
16
Experiments were carried out in a 10 mM sodium cacodylate buffer, at 60 [mu]M nucleotide concentration for each strand.
The affinity of the phosphoramidate oligomer
4
was also compared with that of two other backbone modified dodecathymidines: a phosphorothioate and a methylphosphonate analogues. Data on the interactions with DNA and RNA targets are reported in Table
5
and some typical dissociation curves are shown in Figure
3
. It appears, when considering duplexes formed with single stranded DNA, that
phosphoramidate linkage is slightly less destabilizing ([Delta]Tm/mod = -1.3oC with
9
and -1oC with poly dA) than the phosphorothioate modification ([Delta]Tm/mod = -1.5oC). Comparatively, the introduction of
methylphosphonate linkages into the oligodeoxyribonucleotide does not perturb
its ability to bind to the DNA target. However, Figure
3
clearly shows broader melting transitions for neutral methylphosphonate and
phosphoramidate analogues than for ionic phosphorothioate and phosphodiester
oligomers. These broad curves are most likely due to significant differences in
the binding properties of the diastereoisomers within these oligomers (
35
). In contrast to the behaviour of the neutral phosphoramidate and
methylphosphonate oligomers with DNA targets, duplex formation between these
oligonucleotides and the RNA target was not detected under the conditions used
for these experiments (Fig.
3
). Only the upper portion of an apparent melting transition was observed with
the phosphoramidate analogue, and the transition with the methylphosphonate
compound was so broad that it was not possible to ascertain unambiguously that
the oligomer hybridized to the poly rA target. Finally, the ionic
phosphorothioate modification was the less destabilizing of the studied modifications ([Delta]Tm/mod = -1.7oC) when considering duplexes formed with the ribonucleotide
target.
Enzymatic degradation of oligonucleoside phosphoramidates
Stability of phosphodiester links present in oligonucleotide analogues
2
and
3
was investigated in comparison with that of unmodified (dTp)
11
dT towards the action of purified S1 nuclease, calf spleen phosphodiesterase
(CSPDE) or snake venom phosphodiesterase (SVPDE). Table
6
summarises the half-lives of the oligomers. As can be seen, the rate of hydrolysis by nuclease
S1 of the alternated oligonucleotide
2
(t
1/2
= 20.7 h) is greatly reduced compared with that of unmodified
5
(t
1/2
= 7 min). In contrast, the phosphodiester `window' of the oligomer
3
is not protected by P-NH
2
sections present at 3'- and 5'-ends. Towards the action of the 5'-exonuclease CSPDE, one modification at
the 5'-end of the oligomer
2
prevents vicinal phosphodiester link from the degradation (t
1/2
= 26 h). The stabilizing effect of three consecutive phosphoramidate
modifications at the 5'-end on a neighbouring phosphodiester bond (oligomer
3
, t
1/2
= 12 days) is more pronounced. When
2
or
3
was incubated with 3'-exonuclease SVPDE, hydrolysis of the phosphodiester bonds was lower when
compared with those obtained with
5
(t
1/2
= 14 min). No significant difference was observed between the two modified
oligomers (t
1/2
= 9 h for compound
2
, and 8.5 h for compound
3
). This could result, at least in part, from the chemical instability of the
phosphoramidate link in base medium (pH 9) used in these experiments (t
1/2
of ~20 h for
2
and
3
). Data reported in Table
6
clearly indicate a pH dependence for the hydrolysis of the phosphoramidate
links with a maximum of stability ~pH 7.
.
Half-lives of oligonucleotide analogues in the presence of purified enzymes
(dTp)
11
dT (
5
)
(dTpndTp)
5
dTpndT (
2
)
(dTpn)
3
(dTp)
5
(dTpn)
3
dT (
3
)
S1 nuclease
a
7 min
20.7 h (11 days)
7 min (14 days)
CSPDE
b
22 min
26 h (19 days)
12 days (20 days)
SVPDE
c
14 min
9 h (23 h)
8.5 h (20 h)
Half-lives were determined by measuring the disappearance of starting
oligonucleotide. Reactions were performed at 37oC in:
a
50 mM sodium acetate buffer (pH 4.5) containing 300 mM sodium chloride and 100
mM zinc acetate;
b
125 mM ammonium acetate buffer (pH 6.8) containing 2.5 mM EDTA;
c
100 mM Tris-HCl buffer (pH 9) containing 10 mM magnesium chloride.
Values indicated in brackets refer to half-lives in absence of nuclease.
Phosphoramidate (P-NH
2
) represents an interesting oligonucleotide backbone modification because the NH
2
group is small, uncharged in aqueous solution and has favourable solubility characteristics. We have shown that the combination of H-phosphonate and phosphoramidate chemistries, with the use of the easily cleaved
oxalyl anchor, labile exocyclic amino protecting groups, and a final
deprotection with methanol saturated with ammonia had enabled us to synthesize
partially or totally modified dodecadeoxynucleotides. These compounds are able
to form hybrids with complementary DNA or RNA strands. The resulting duplexes
are significantly less stable than those of the `wild-type' species, especially with RNA targets. We demonstrated that a
modified oligomer combining an internal phosphodiester section surrounded by
phosphoramidate sections at 5'- and 3'-ends was resistant to exonuclease hydrolysis and able to elicit
RNase H cleavage of the RNA target. These properties make the modification attractive for constructing potential candidates for the antisense approach.
ACKNOWLEDGEMENTS
This work was supported by ANRS (Agence Nationale de Recherche sur le Sida) and ARC (Association pour la Recherche contre le Cancer. One of us
(S.P.) thanks the Ministère de l'Enseignement Supérieur et de la Recherche for the award of a research
studentship. The authors are grateful to Dr W. Douglas for careful reading of
the manuscript.
REFERENCES
1 Uhlmann, E. and Peyman, A. (1990) Chem. Rev. 90, 543-584.
2 Milligan, J. F., Matteucci, M. D. and Martin, J. C. (1993) J. Med. Chem. 36, 1923-1937.MEDLINE Abstract
4 Cohen, J. S. (1993) In S. T. Crooke and B. Lebleu (ed.), Antisense Research and Applications; CRC Press: Boca Raton (Florida); pp 205-221.
5 Stein, C. A., Tonkinson, J. L. and Yakubov, L. (1991) Pharmac. Ther. 52, 365-384.
6 Crooke, R. M. (1993) In S. T. Crooke and B. Lebleu (ed.), Antisense Research and Applications; CRC Press: Boca Raton (Florida); pp 427-449.
7 Miller, P. S. (1989) In J. S. Cohen (ed.), Oligodeoxynucleotides. Antisense Inhibitors of Gene Expression, MacMillan Press, London, pp 79-95.
8 Froehler, B. C. and Matteucci, M. D. (1988) Nucleic Acids Res. 16, 4831-4839.MEDLINE Abstract
9 Fathi, R., Huang, Q., Syi, J. L., Delaney, W. and Cook, A. F. (1994) Bioconjugate Chem. 5, 47-57.
10 Ozaki, H., Kitamura, M., Yamana, K., Murakami, A. and Shimidzu, T. (1990) Bull. Chem. Soc. Jpn 63, 1929-1936.
11 Durand, M., Maurizot, J. C., Asseline, U., Barbier, C., Thuong, N. T. and Helene, C. (1989) Nucleic Acids Res. 17, 1823-1837.MEDLINE Abstract
12 Stec, W. J. and Wilk, A. (1994) Angew. Chem. 33, 709-722.
13 Vyazovkina, E. V., Savchenko, E. V., Lokhov, S. G., Engels, J. W., Wickstrom, E. and Lebedev, A. V. (1994) Nucleic Acids Res. 22, 2404-2409.MEDLINE Abstract
14 Nielsen, P., Kirpekar, F. and Wengel, J. (1994) Nucleic Acids Res. 22, 703-710.MEDLINE Abstract
15 Sanghvi, Y. S. and Cook, P. D. (1993) In C. K. Chu and Baker D.C. (ed.), Nucleosides and Nucleotides as Antitumor and Antiviral Agents,Plenum Press: New York; pp 311-323.
16 Milligan, J. F., Matteucci, M. D. and Martin, J. C. (1993) J. Med. Chem. 36, 1923-1937.MEDLINE Abstract
17 Goodchild, J. (1990) Bioconjugate Chem. 2, 165-187.
18 Matrosov, E. I., Kryuchkov, A. A., Nifantyev, E. E., Kozachenko, A. G. and Kabachnik, M. I. (1982) Phosphorus Sulfur 13, 69-78.
19 Letsinger, R. L., Bach, S. A. and Eadie, J. S. (1986) Nucleic Acids Res. 14, 3487-3499.MEDLINE Abstract
20 Froehler, B. C. (1986) Tetrahedron Lett. 27, 5575-5578.
21 Peyrottes, S., Vasseur, J. J., Imbach, J. L. and Rayner, B. (1994) Nucleosides Nucleotides 13, 2135-2149.
22 Tomasz, J. (1983) Nucleosides Nucleotides 2, 51-61.
23 Alul, R. H., Singman, C. N., Zhang, G. and Letsinger, R. L. (1991) Nucleic Acids Res. 19, 1527-1532.MEDLINE Abstract
24 Sinha, N. D., Davis, P., Usman, N., Perez, J., Hodge, R., Kremsky, J. and Casale, R. (1993) Biochimie 75, 13-23.MEDLINE Abstract
25 Polushin, N. N., Morocho, A. M., Chen, B. and Cohen, J. S. (1994) Nucleic Acids Res. 22, 639-645.MEDLINE Abstract
26 Madsen, R., Roberts, C. and Fraser-Reid, B. (1995) J. Org. Chem. 60, 7920-7926.
27 Iyer, R. P., Devlin, T., Habus, I., Ho, N. H., Yu, D. and Agrawal, S. (1996) Tetrahedron Lett. 37, 1539-1542.
28 Iyer, R. P., Devlin, T., Habus, I., Yu, D., Johnson, S. and Agrawal, S. (1996) Tetrahedron Lett. 37, 1543-1546.
29 Stec, W. J., Zon, G., Egan, W., Byrd, R. A., Philips, L. R. and Gallo, K. A. (1985) J. Org. Chem. 50, 3908-3913.
30 Tomasz, J. and Simoncsits, A. (1981) Tetrahedron Lett. 22, 3905-3908.
31 Morvan, F., Porumb, H., Degols, G., Lefebvre, I., Pompon, A., Sproat, B. S., Rayner, B., Malvy, C., Lebleu, B. and Imbach, J. L. (1993) J. Med. Chem. 36, 280-287.MEDLINE Abstract
32 Jung, P. M., Histand, G. and Letsinger, R. L. (1994) Nucleosides Nucleotides 13, 1597-1605.
33 Oney, I. and Caplow, M. (1967) J. Am. Chem. Soc. 89, 6972-6980.
34 Bollinger, J. C., Faure, R. and Yvernault, T. (1983) Canad. J. Chem. 61, 328-333.
35 Kean, J. M., Cushman, C. D., Kang, H. M., Leonard, T. E. and Miller, P. S. (1994) Nucleic Acids Res. 22, 4497-4503.MEDLINE Abstract
36 Stein, C. A., Subasinghe, C., Shinozuka, K. and Cohen, J. S. (1988) Nucleic Acids Res. 16, 3209-3221.MEDLINE Abstract
37 Giles, R. V. and Tidd, D. M. (1992) Anti-Cancer Drug Des. 7, 37-48.
38 Giles, R. V. and Tidd, D. M. (1992) Nucleic Acids Res. 20, 763-770.MEDLINE Abstract
39 Dagle, J. M., Walder, J. A. and Weeks, D. L. (1990) Nucleic Acids Res. 18, 4751-4757.MEDLINE Abstract
40 Potts, J. D., Dagle, J. M., Walder, J. A., Weeks, D. L. and Runyan, R. B. (1991) Proc. Natl. Acad. Sci. USA 88, 1516-1520.MEDLINE Abstract
41 Bloch, E., Lavignon, M., Bertrand, J. R., Pognan, F., Morvan, F., Malvy, C., Rayner, B., Imbach, J. L. and Paoletti, C. (1988) Gene 72, 349-360.MEDLINE Abstract
