2'-O-aminopropyl ribonucleotides: A zwitterionic modification that enhances the exonuclease resistance and biological activity of antisense oligonucleotides

Article abstract of DOI:10.1021/jm950937o

Oligonucleotides containing 2'-O-aminopropyl-substituted RNA have been synthesized. The 2'-O-(aminopropyl)adenosine (APA), 2'-O- (aminopropyl)cytidine (APC), 2'-O-(aminopropyl)-guanosine (APG), and 2'-O- (aminopropyl)uridine (APU) have been prepared in hi

Full text of DOI:10.1021/jm950937o

5100
J . Med. Chem. 1996, 39, 5100-5109
2′-O-Am in op r op yl Ribon u cleotid es: A Zw itter ion ic Mod ifica tion Th a t En h a n ces
th e Exon u clea se Resista n ce a n d Biologica l Activity of An tisen se
Oligon u cleotid es†
Richard H. Griffey,* Brett P. Monia, Lendall L. Cummins, Susan Freier, Michael J . Greig, Charles J . Guinosso,
Elena Lesnik, Sherilynn M. Manalili, Venkatraman Mohan, Steven Owens, Bruce R. Ross, Henri Sasmor,
Ed Wancewicz, Kurt Weiler, Patrick D. Wheeler, and P. Dan Cook
Isis Pharmaceuticals, 2280 Faraday Avenue, Carlsbad, California 92008
Received December 22, 1995X
Oligonucleotides containing 2′-O-aminopropyl-substituted RNA have been synthesized. The
2′-O-(aminopropyl)adenosine (APA), 2′-O-(aminopropyl)cytidine (APC), 2′-O-(aminopropyl)-
guanosine (APG), and 2′-O-(aminopropyl)uridine (APU) have been prepared in high yield from
the ribonucleoside, protected, and incorporated into an oligonucleotide using conventional
phosphoramidite chemistry. Molecular dynamics studies of a dinucleotide in water demon-
strates that a short alkylamine located off the 2′-oxygen of ribonucleotides alters the sugar
pucker of the nucleoside but does not form a tight ion pair with the proximate phosphate. A
5-mer with the sequence ACTUC has been characterized using NMR. As predicted from the
modeling results, the sugar pucker of the APU moiety is shifted toward a C3′-endo geometry.
In addition, the primary amine rotates freely and is not bound electrostatically to any phosphate
group, as evidenced by the different sign of the NOE between sugar proton resonances and
the signals from the propylamine chain. Incorporation of aminopropyl nucleoside residues into
point-substituted and fully modified oligomers does not decrease the affinity for complementary
RNA compared to 2′-O-alkyl substituents of the same length. However, two APU residues
placed at the 3′-terminus of an oligomer gives a 100-fold increase in resistance to exonuclease
degradation, which is greater than observed for phosphorothioate oligomers. These structural
and biophysical characteristics make the 2′-O-aminopropyl group a leading choice for incorpora-
tion into antisense therapeutics. A 20-mer phosphorothioate oligonucleotide capped with two
phosphodiester aminopropyl nucleotides targeted against C-raf mRNA has been transfected
into cells via electroporation. This oligonucleotide has 5-10-fold greater activity than the
control phosphorothioate for reducing the abundance of C-raf mRNA and protein.
In tr od u ction
oxynucleotides. However, recognition of duplexes with
complementary RNA by enzymes such as RNAse H is
lost.8-10
Antisense therapy has stimulated interest in prepa-
ration of modified DNA and RNA oligonucleotides which
address the limitations of the unmodified phosphodi-
ester backbone of naturally occurring nucleic acids. In
vivo, oligonucleotides are degraded by nucleases, have
low cellular permeation, and reduce affinity of short
(∼20-mer) strands for their RNA target.1-3 Modification
of the phosphodiester backbone via replacement of
equatorial oxygen(s) with sulfur generates the phos-
phorothioate and phosphorodithioate oligonucleotides,
with a dramatic enhancement in nuclease resistance.
More extensive atomic replacements yield the MMI,
PNA, guanidyl, and amide backbones.4-7 Oligonucle-
otides containing these modifications have reduced
negative charge wherever the achiral phosphodiester is
replaced by a neutral or positively charged linkage. Such
backbone incorporations favorably enhance the proper-
ties of the resulting molecules, producing higher affinity
for complementary DNA or RNA, greater specificity of
base pairing, and reduced enzymatic degradation com-
pared to phosphodiester or phosphorothioate oligode-
The total net charge on a nucleic acid in vivo also can
be reduced by incorporating C5-(aminohexyl)pyrim-
idines to form zwitterionic molecules.11,12 Placement of
alkylamines in the major groove does not alter the
stability of the DNA duplex formed with a complemen-
tary strand. Any potential stabilization of the duplex
through reduction of unfavorable electrostatic interac-
tions between strands apparently is offset by new
unfavorable steric or entropic interactions involving the
long alkyl side chains off the nucleobases. No data on
stability or uptake have been reported.
Ch em istr y
We have investigated an alternative form of zwitte-
rionic RNA where a cationic group is introduced into
the minor groove of the duplex by derivitization of the
ribose sugar at the C2′-position. The synthesis of
precursors for incorporation of alkylamine-substituted
nucleotides using conventional phosphoramidite chem-
istry is complicated by the need to protect the amine
with a functional group which is stable to the conditions
of oligomerization and can be removed quantitatively.
We have explored incorporation of 2′-aminopropyl
and 2′-aminopentyl units via phthalimide protection
as well as the more labile trifluoroacetyl-protected deri
† Abbreviations: ES, electrospray; AP, aminopropyl; PAGE, poly-
acrylamide gel electrophoresis; DMT, dimethoxytrityl; CE, capillary
electrophoresis; MMI, methylenemethylimino; PNA, peptide nucleic
acid; SVPD, snake venom phosphodiesterase.
* Address correspondence to this author: tel, (619)-603-2430; fax,
(619)-929-0036.
X Abstract published in Advance ACS Abstracts, December 1, 1996.
S0022-2623(95)00937-X CCC: $12.00 © 1996 American Chemical Society

2′-Aminopropyl Oligonucleotides
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 26 5101
Sch em e 1. Route for Synthesis of Protected 2′-O-Aminopropyl Nucleoside Phosphoramiditesa
a
(i) NaH/R-Br/DMF/5 °C/16 h; (ii) Bz-Cl/pyridine/20 °C/8 h; (iii) EtOH/H2NNH2/16 h; (iv) DMT-Cl/pyridine/20 °C/16 h; (v) CF3CO2Et/
MeOH/TEA/20 °C/16 h; (vi) C9H18N3OP/CH2Cl2/20 °C/16 h.
vatives of the amine. As shown in Scheme 1, the
synthetic route starts by taking advantage of the acidity
(pKa ∼ 12) of the 2′-hydroxy proton of the purine
nucleosides. The reaction of adenosine and 2-amino-
adenosine with N-(3-bromopropyl)phthalimide results
in the isolation of the 2′-substituted nucleosides. The
2′,3′-stannyl complex of uridine provides entry to the
2′-O-alkylated product.13 In the case of cytidine, a
mixture of the 2′/3′-isomers was obtained and the pure
2′-product was not isolated until the DMT stage. Con-
version of the alkylated 2-aminoadenosine into the
2′-O-(aminopropyl)guanosine is accomplished using ad-
enosine deaminase.14 The trifluoroacetyl-protected nu-
cleosides are generated from the 2′-O-phthalimido nu-
cleosides. Hydrazine readily cleaves the phthaloyl
group, and the resulting primary amine is acylated
directly with trifluoroethyl acetate. Tetrabutylammo-
nium fluoride has been employed in the transient
protection of the heterocycle instead of ammonium
hydroxide to assure the integrity of the trifluoroacetyl
group.15 Conventional benzoyl and isobutyryl groups
were used to protect the amino functionality of the
nucleosides, which were subsequently converted to the
5′-O-DMT derivatives and finally the bis(N,N-diisopro-
pyl)phosphoramidites.
digestion, and incorporation into cells. We demonstrate
that the AP nucleosides can be incorporated into oligo-
nucleotides in high yield via the phthalimide-protected
phosphoramidites. Molecular dynamics analyses of
the hydration and electrostatic interaction of the
amine with a phosphate linkage have been performed.
NMR studies of a pentanucleotide containing a single
AP uridine suggest that the amino group rotates freely
relative to the phosphodiester backbone and is not
bound via inner-sphere coordination. Hybridization
studies for oligomers substituted with aminopropyl
and pentylamino residues as point substitutions and
in fully modified sequences show that the alkylamine
side chains reduce the Tm of a duplex with RNA
complement by an amount comparable to an alkyl
chain of the same length. However, incorporation of
two AP units at the 3′-terminus of an oligomer dramati-
cally enhances the exonuclease stability of the strand
compared to phosphodiester or phosphorothioate-
substituted oligonucleotides. A 2′-O-aminopropyl-modi-
fied oligonucleotide has been targeted against C-
raf mRNA in cells using electroporation. Compared to
the control phosphorothioate, this compound demon-
strates enhanced potency for reducing mRNA levels
and protein levels in A549 cells. Hence, aminopropyl
oligonucleotides will be useful in the design of novel
antisense therapeutics with favorable properties in
vivo.
The properties of 2′-O-aminopropyl (AP) oligonucle-
otides have been analyzed using molecular dynamics,
NMR, hybridization with RNA complement, enzymatic

5102 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 26
Ta ble 1. Sequence and Structure of Oligonucleotidesa
Griffey et al.
oligomer
sequence
A
B
C
D
E
F
G
H
I
ACTUC
ACCGAGGATCATCTCGTACGC
GGACCGGAAGGTACGAG
CGACTATGCAAAAAC
CGACGATGCAAGTAC
CTCGTACCATTCCGGTCC
GAGCTCCCAGGC
GAGCUCCCAGGC
CGACUAUGCAAGUAC
J
GGACCGGAAGGUACGAG
UUCUCGCCCGCTCCUCCUCC (full PdS)
UUCUCGCCCGCTCCUCCUCC (PdS gap)
UUUUUTTTTTTTTTUUUUTb
UUUUUUUUUUUUUUUU
TTTTTTTTTTTTTTUUUUT
TCCCGCCTGTGACAUGCAUU
TCCCGCCTGTGACATGCATT
K
L
M
N
O
P
F igu r e 1. ES-MS spectrum for propylamine oligonucleotide
M prepared using TFA protection of the primary amine. The
[M - 5H]5- and [M - 4H]4- charge states are shown between
m/ z 1200 and 1650. The ions originating from acrylonitrile
adducts (+53 Da) are indicated as +AN. Molecules with 1-4
acrylonitrile adducts can be observed in the mass spectrum
for both charge states. No acrylonitrile adducts were observed
for oligonucleotides prepared using the phthaloyl protection/
deprotection strategy. The samples were desalted via precipi-
tation from 10 M ammonium acetate prior to analysis.
Q
a
Italics indicate a 2′-O-aminopropyl-substituted nucleotide in
the sequence. For H, italics indicate 2′-O-methyl nucleotide
substitutions. Underlined residues contain internucleotide phos-
phorothioate linkages. Italics indicate
substituted nucleotide in the sequence.
b
a 2′-O-pentylamino-
Resu lts
Syn th esis of 2′-O-Am in op r op yl Oligon u cleotid es.
Two protection strategies have been evaluated for
incorporation of aminopropyl nucleotides into the cor-
responding amidites, as shown in Scheme 1. Initial
attempts were performed using the trifluoroacetamide
of the amine. Utilization of this group has been
described previously by Switzer et al.11 The synthetic
route begins with 2′-O-alkylation of adenosine, 2-ami-
noadenosine, and cytidine using the 3-N-phthalimi-
dopropyl halide to yield compounds 2a -d . The phthal-
imido moiety is removed with hydrazine, and the
primary amine is quantitatively converted to the trif-
luoroacetamide. Alkylation of uridine proceeds through
the 2′,3′-stannyl oxide. The 2′-O-(3-N-phthalimidopro-
pyl)guanosine is generated by enzymatic conversion of
the 2-aminoadenosine derivative using adenosine deam-
inase. Subsequent protection of the heterocycles with
benzoyl or isobutyryl groups followed by addition of a
dimethoxytrityl group at O5′ gives compounds 5a -d .
Conversion to the phosphoramidite using the bis(N,N-
diisopropylamino)phosphine provided phosphoramidites
6a -d . Incorporation of the amidites into oligonucle-
otides listed in Table 1 proceeded smoothly.
into oligomers with >98.5% stepwise coupling efficiency
using two sequential coupling steps. Removal of the
phthalimide group from the product oligonucleotide via
ammonium hydroxide treatment was ineffective, leaving
partially ring-opened phthalamide functionality bound
to the amine group. A second deprotection strategy was
developed, using an initial ammonium hydroxide treat-
ment followed by reaction with 10% aqueous methy-
lamine.16 In the initial treatment, the cyanoethyl
groups could be removed from the phosphate to prevent
alkylation of the nascent primary alkylamine with
acrylonitrile, and the benzoyl groups could be removed
rapidly from the exocyclic amine of C nucleotides to
preclude substitution of the exocyclic amine of C with
methylamine.17 Further reaction of the oligomer with
10% aqueous methylamine cleanly deprotects the isobu-
tyryl groups from guanosine and the phthalimide groups
from the alkylamine side chains. PAGE and ES-MS
analyses showed that the aminopropyl and aminopentyl
groups could be incorporated cleanly into oligomers
under these conditions.
Molecu la r Dyn a m ics. The 2′-O-aminopropyl group
has the potential to interact with any of the electrone-
gative groups on surrounding nucleotides. We have
investigated the probability for such interactions using
a UT dinucleotide model system that could be modeled
in the presence of explicit solvent molecules. The
simulation was performed in a 20.0 Å3 box of pre-
equilibrated water molecules. Data from the first 20
ps of the calculation have been attributed to the
equilibration process and were not included in the final
analysis. The system remains energetically stable after
the initial equilibration phase for the next 150 ps. The
computed structural properties (radial distribution func-
tions; data not shown) of the solvent model used in the
present study are in close agreement with literature
results employing different water models.18 Over the
time course of the simulation, the distance from the
terminal nitrogen to the equatorial oxygens of the
phosphate varied randomly between 3.2 and 6.5 Å. No
binding of the amino group was detected to neighboring
base carbonyl or sugar O4′ atoms.
In our hands, the trifluoroacetyl group was removed
very quickly upon treatment with ammonia during
oligonucleotide deprotection. ES-MS analysis of the
oligonucleotide products demonstrated the presence of
multiple side products. As shown in Figure 1 for M, a
series of adducted ions corresponding to n*53.1 Da (n
) 1-4) could be detected at higher m/ z for each charge
state of the oligonucleotide. The 53.1 Da increase in
molecular mass corresponds to addition of C3H3N,
readily rationalized by Michael addition of the primary
amine to acrylonitrile generated during deprotection of
the cyanoethyl-protected phosphate triesters. Prolonged
deprotection (24 h) with ammonia at 55 °C did not
reduce the quantity of the adduct or change the observed
ion intensities and values of m/ z.
Observation of the acrylonitrile adducts in five oligo-
mers prepared using the trifluoroacetyl protecting group
led us to evaluate a different deprotection strategy for
the aminopropyl functionality. The phthalimide-pro-
tected phosphoramidites 9a -d could be incorporated

2′-Aminopropyl Oligonucleotides
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 26 5103
Ta ble 2. Proton NMR Chemical Shifts, Coupling Constants, and Pseudorotation Parameters for A
chemical shift (ppm) coupling constant (Hz)
residue
H8/H6/H5
H1′
H2′, H2′′
H3′
3J (H1H2)
3J (H2H3)
3J (H3H4)
PN
15
PS
%N
A
C
8.01
7.62
5.70
7.43
1.67
7.69
5.74
7.72
5.91
3.62
1.80
2.98
6.24
6.06
2.64, 2.57
2.37, 2.09
4.60
4.69
8.0, 6.0
7.9, 6.1
5.6
5.6
2.4
168
15.0
nda
T
U
C
6.03
5.82
6.11
2.33, 2.14
4.04
4.71
4.57
4.38
8.0, 6.0
5.2
5.6
4.8
4.8
nd
e4.4
2.6
1.5
4.0
170.5
164.5
40.0
14.0
2.26, 2.14
6.8, 6.8
propylamine
a
Could not be determined accurately due to spectral overlap.
Ta ble 3. Hybridization Data for Oligonucleotides with AP
Point Substitutions against RNA Complement
The pseudorotation of the 2′-O-aminopropylU sugar
was calculated from the internal dihedrals as a function
of time. The psuedorotation lies predominately in the
C3′-endo form but does transit to a C2′-endo pucker.
This interconversion is independent of the motion of the
propylamine side chain. The upper sugar tends toward
a C4′-exo conformation, although this may be an artifact
of having no constraints on the 5′-end of the dinucle-
otide.
sequence
substitution
B
C
D
E
F
G
H
DNA, PdS
2′-O-methyla
2′-O-propyla
-0.38 nd
-0.52 -0.80 -0.65 -0.73 nd
-0.01 +0.96 +0.91 +0.31 +0.03 +0.67 +2.13
-0.36 +0.24 +0.84 -1.02 -1.14 -0.10 +2.00
2′-O-propylamine -0.40 +0.27 +1.09 -1.08 -1.78 +1.13 +1.88
a
Values are listed as ∆Tm values calculated versus the unmodi-
fied DNA control. Data taken from ref 10; nd ) not determined.
NMR Sp ectr oscop y. The results obtained from
molecular modeling have been confirmed using NMR
in a larger model, 5-mer oligonucleotide A containing
an AP uridine. A pKa value of 9.4 has been determined
for the propylamine group in the oligomer from observa-
tion of the proton chemical shifts of the propyl protons
as the pH was titrated between 5.0 and 11.0. The 31P
NMR chemical shifts for the phosphates were invariant
as a function of pH between 5.0 and 9.0.
Ta ble 4. Hybridization Data for Fully Substituted
Oligonucleotides against RNA Complement
sequence
substitution
I
J
DNA, PdS
-0.52
+1.29
+0.99
+0.43
nd
2′-O-methyla
2-O-propyl
+1.49
+1.10
+0.39
2-O-propylamine
The coupling constants for the sugar protons have
been measured directly from a 1D spectrum or a TOCSY
spectrum obtained with 0.5 Hz resolution in the acquisi-
tion dimension and are presented in Table 2. The
pseudorotation parameters for the AP uridine have been
determined from the values of the H1′-H2′, H2′-H3′,
and H3′-H4′ scalar coupling constants. These values
have been used to generate a fit to a two-state model
for psuedorotation using a version of PSEUDOROT
modified for 2′-O-alkyl substituents.19 <Sen>The AP
uridine exists as 40% C3′-endo conformer (P ) 1.5°) and
60% C2′-endo (P ) 170.5°). Variations of (0.4 Hz in
the value of the H3′H4′ coupling constant used in the
calculation of the pseudorotation parameters shifted PN
by (17° and PS by (5° but did not alter the %N
conformation for the propylamine sugar. The 3′-flank-
ing DNA residue adopted a C2′-endo conformation (%N
) 14.0) with PN ) -4.0° and PS ) 164.5°. Predomi-
nantly C2′-endo conformations are observed for the
remaining DNA nucleotides.
A NOESY spectrum was acquired with a 500 ms
mixing time at 20 and 37 °C. At 20 °C, interproton NOE
cross-peaks from the H5 and H6 base protons of C and
U have the same sign as the diagonal, and few NOE
cross-peaks are observed from sugar protons, while the
cross-peaks among the protons of the propyl arm are
inverted relative to the diagonal. A ROESY spectrum
obtained with a 100 ms mixing time generates off-
diagonal peaks which are inverted relative to the
diagonal. Assuming that ω*τ(c) ) 1.0 and given the
proton resonance frequency of 400 MHz, the effective
correlation time τ(c) for the sugar protons can be
a
Values are listed as ∆Tm values calculated versus the unmodi-
fied DNA PdO control; nd ) not determined.
calculated to be ∼8.0 × 10-10 s at this temperature. In
NOESY spectra obtained at 37 °C, the signs of the cross-
peaks between the H5,H6 base protons and the sugar
protons all invert relative to the diagonal.
Hybr id iza tion Stu d ies. Hybridization studies have
been conducted for both point substitutions in sequences
with a DNA background and fully modified sequences
against RNA complement. Oligomer sequences are
listed in Table 1, while ∆Tm values are listed in Table
3. As a point modification in six deoxynucleotide
sequences, AP A is stabilizing compared to PdS DNA
in four cases and destabilizing in two cases. In sequence
C where AP A residues are adjacent, the modification
provides some stabilization. Duplex stabilization by the
AP A relative to dA and 2′-O-alkyl A is observed in
sequences D and G, where the AP A is flanked on the
3′-side by a purine. As point substitutions in a 2′-O-
methyl background (sequence H), the AP A modification
is stabilizing compared to DNA but slightly destabilizing
compared to the 2′-O-methyl and 2′-O-propyl modifica-
tions. When incorporated into fully modified sequences
I and J , the AP nucleotides are stabilizing compared to
DNA by +0.4 °C but less stabilizing than the 2′-O-
methyl and -propyl derivatives, listed in Table 4. The
effect of incorporating 2′-substituents in the wings of
phosphorothioate and phosphodiester chimeric oligo-
nucleotides has been investigated. As shown in Table
5, the aminopropyl group does not destabilize the fully

5104 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 26
Griffey et al.
Ta ble 5. Hybridization Data for Chimeric Oligonucleotides
Containing Modified Wings against RNA Complement
sequence
substitution
K
L
2′-O-methyla
2-O-propyl
2-O-propylamine
+0.95
-0.93
+0.01
nd
+0.00
-0.46
a
Values are listed as ∆Tm values calculated versus the unmodi-
fied DNA PdO control; nd ) not determined.
Ta ble 6. Rates of Degradation of Oligonucleotide O by SVPD
modification
sequence
T14U4T
t1/2 SVPD (N f N - 1; min)
PdO
PdS
4 ( 1
360
20
T
T
T
T
T
T
T
T
T
14U4T
17UT
2′-AP PdO
AP PdO
AP PdO
AP PdO
Pr PdO
1 AP, 3 Pr
2 AP, 2 Pr
3 AP, 1 Pr
16U2T
15U3T
14U4T
14U4T
14U4T
14U4T
14U4T
600
1440 ( 60
>1440
7
e15
800
1200
F igu r e 2. Scan of a PAGE gel showing time course of SVPD-
catalyzed degradation of oligonucleotide O with 1-4 2′-O-
(aminopropyl)uridine residues. The samples in each lane were
obtained at 0, 0.5, 1, 2, 4, 8, and 24 h time points.
PdS sequence K, in contrast to the oligomer containing
2′-O-propyl nucleotides. For comparison, the 2′-O-
methyl nucleotides provide some stabilization in se-
quence K. However, some destabilization from the AP
nucleotides is observed in oligomer L with phosphodi-
ester wings relative to the propyl derivative. Finally,
the salt dependence of the Tm value has been measured
for sequence M containing five 2′-O-pentylamino resi-
dues. As observed by Hashimoto et al., incorporation
of pendant amino groups reduces the slope of ∆Tm/∆-
[Na+] by 2.5%/amino substitution compared to a control
deoxynucleotide sequence.12
Nu clea se Digestion . The 3′-exonuclease resistance
of oligomers containing the propylamine modification
has been determined using snake venom phosphodi-
esterase (SVPD).20 All oligomers were 5′-end-labeled
with [32P]ATP and T4 polynucleotide kinase. Studies
were performed with sequence O, dT(14+N)U(4-N)dT,
where N ) 0-4 and U represents 2′-O-(aminopropyl)-
uridine. In all cases, a sequential degradation from the
3′-terminus was observed. Under the conditions used
in the assay, the t1/2 for a dT19 phosphodiester was ∼4
min, while the t1/2 value for a dT19 phosphorothioate was
∼360 min. The t1/2 values are presented in Table 6. In
the molecule with a single AP substitution, the 3′-
terminal dT residue is cleaved with a t1/2 of 20 min. A
gap is observed in the gel at the position where the
zwitterionic AP U residue is lost, as a result of the
increase in the charge/mass ratio. Surprisingly, incor-
poration of two sequential AP U nucleotides increases
the t1/2 for removal of the 3′-dT from 20 to 600 min.
Further incorporation of sequential AP U residues
generates more exonuclease resistance with t1/2 values
of e1440 and >1440 min for three and four incorpora-
tions, respectively. A gel showing the course of degra-
dation of the oligomers containing 1-4 AP U incorpo-
rations is presented in Figure 2.
U residues in a 2′-propyluridine background in oligomer
N. As listed in Table 6, the t1/2 values were similar to
those observed for the oligomers with dT substitutions.
The possibility that the on-rate for SVPD to the 3′-
terminus of the oligomer with multiple AP U incorpora-
tions was altered was explored in a series of oligomers
where the location of the AP U “block” was moved
internally. Normal kinetics were observed in all cases
for digestion of 3′-dT residues up to the first dT adjacent
to a AP-containing nucleotide. A kinetic analysis of the
competition for degradation between 3′-(p-nitrophenyl)-
phosphoryl dT demonstrated that oligomer O is not a
better competitive inhibitor of enzyme activity than a
comparable deoxyphosphodiester oligomer (data not
shown).
In h ibition of C-r a f m RNA a n d P r otein Exp r es-
sion . A 20-mer phosphorothioate oligonucleotide (Q)
previously demonstrated to reduce the levels of C-raf
mRNA and protein was prepared (P ) with nine AP
nucleotides at the 3′-terminus.27 Attempts to transfect
P into A549 cells with varying concentrations of cationic
lipids proved unsuccessful. However, both P and Q
could be transfected efficiently using electroporation
methods.28 As shown in Figure 3, a 250 nM concentra-
tion of P lowered the level of C-raf mRNA by 50% 24 h
following administration, while a 10-fold higher con-
centration of Q was required to effect the same reduc-
tion. The reduction in levels of C-raf protein was more
pronounced, where a ∼150 nM concentration of P
produced a 50% reduction in protein at 48 h, compared
to >300 µM Q.
Discu ssion
The origins of this increase in nuclease resistance was
explored in a second series of experiments. No change
in the rate of degradation of the 3′-dT was observed in
molecule O between pH 7.0 and 9.5. The possibility that
a conformational change or steric effect from incorporat-
ing 2′-substituted RNA into the oligomer instead of DNA
was ruled out by incorporating varying numbers of AP
The design of antisense oligonucleotides is compli-
cated by the need to improve properties such as resis-
tance to nuclease degradation while maintaining cellu-
lar activity. Generation of chimeric oligomers which
mix different chemistries has proven to be an effective
approach.21 This strategy keeps the favorable proper-
ties of a negatively charged deoxynucleotide “gap” which

2′-Aminopropyl Oligonucleotides
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 26 5105
to be incompatible with the conditions of DNA synthesis.
Early deprotection of the primary amine allows Michael
addition to acrylonitrile generated during removal of the
cyanoethyl protecting groups from the phosphates. A
differential deprotection strategy can be employed when
the amine is incorporated as an alkylphthalimido
moiety. The cyanoethyl and benzoyl protecting groups
are removed rapidly during a preliminary treatment
with ammonia. Subsequent treatment with methy-
lamine rapidly removes the phthalimido protecting
groups and the isobutyryl protecting groups on the
exocyclic amine of G.16
Molecu la r Mod elin g a n d NMR. In the MD study,
both the propylamine and the phosphate backbone are
coordinated to waters of hydration. The protonated
propylamine is not attracted toward the phosphate by
electrostatic forces during the 150 ps of the simulation
and does not form an inner-sphere coordination com-
plex. The lack of a strong electrostatic interaction of
the propylamine with anions is confirmed following
addition of sodium chloride to the simulation. The
position of the chloride anion equilibrates without strong
binding to the propylamine. These results suggest that
the propylamine side chain is of a correct length to
minimize the formation of tight ion pairs and hydrogen-
bonding interactions with neighboring functionality in
the oligomer.
The MD study also shows that the AP sugar adopts
a conventional northern conformation centered around
0°. During the course of the simulation, one movement
toward an O4′-endo conformation was observed, but the
conformation rapidly fell back to the more stable C3′-
endo sugar pucker. This sugar conformation minimizes
potential steric interactions between the propylamine
side chain and the H5′,H5′′ sugar protons on the
3′-nucleotide.
F igu r e 3. Inhibition of C-raf kinase mRNA and protein
expression in A549 cells. Upper: quantitation of C-raf mRNA
levels following electroporation with increasing concentrations
of 20-mer PS oligonucleotides P ([; 3′-AP2), Q (9; no AP), and
a scrambled control (b) phosphorothioate. Total RNA was
prepared 24 h after treatment and analyzed for C-raf and
G3PDH mRNA levels by Northern blot analysis. Lower:
quantitation of C-raf protein levels in A549 cells treated with
varying concentrations of P ([) and Q (9). Cells were treated
with ODN, protein extracts were prepared 48 h later, and C-raf
protein levels were determined as described in the Experi-
mental Section.
The solution dynamics of the propylamine side chain
1
have been confirmed in H NMR studies, which show a
greater mobility for the AP protons compared to the
protons of the sugar-phosphate backbone. The com-
parison of cross-peak intensities from the ROESY and
NOESY spectra for a pentanucleotide containing a
single AP substitution suggests that the propylamine
group has a 2-4-fold shorter correlation time than the
oligonucleotide backbone.24 This short correlation time
is consistent with uncorrelated motion between the
phosphate and the AP chain and confirms the modeling
results on the longer NMR time scale. In addition, the
31P NMR chemical shift is invariant over a pH range
where the amino group titrates, suggesting that the
amine does not bind via inner-sphere coordination to
the flanking phosphate. Formation of a tight ion pair
between the amine and a desolvated phosphate would
be undesirable, since hydration of the phosphates may
be critical for solubility.
can provide a binding domain for enzymes such as
RNAse H, while the surrounding “wings” can have
reduced net charge or other characteristics which
provide improved uptake and pharmacokinetics, alter
affinity for specific proteins, or increase nuclease resis-
tance.22 In the wings, positioning of positively charged
groups in the minor groove rather than the major groove
decreases their proximity to the density of negative
charge of the phosphate backbone and is straightfor-
ward synthetically. A short 2-3-carbon alkyl chain will
have less entropy and less disruption of hydration of a
duplex and should have superior properties compared
to a longer alkyl linker. The distance from the amino
group to hydrogen-bonding/electrostatic functionality of
bases on the other strand is longer with a short alkyl
chain, which should minimize nonspecific steric interac-
tions. Hence, a short linker chain positioned in the
minor groove might be expected to overcome the prob-
lems associated with an amine situated in a major
groove location off C5 of pyrimidine nucleotides.
The sugar pucker for the AP unit is altered from
conventional RNA as evidenced from the large H1′-H2′
coupling constant compared to RNA or 2′-O-Me deriva-
tives.9 The coupling constants can be fit to a two-state
exchange model between C3′-endo and C2′-endo sugar
puckers, with a 40% proportion of a C2′-endo conforma-
tion. This conformation can be accommodated in the
single-stranded oligomer, but the C2′-endo sugar pucker
for an AP nucleotide would be difficult to fit into an ‘A’-
form duplex with RNA complement, where the alkyl
Syn th esis. A convenient route to the preparation of
oligonucleotides containing 2′-O-alkylamino moieties via
phthalimido protection and differential deprotection has
been developed. Amidites of all four bases can be
prepared using trifluoroacetyl or phthaloyl protecting
groups.23 In our hands, however, the TFA amide proved

5106 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 26
Griffey et al.
chain would have bad steric contacts with the proximate
3′-residue. The destabilization of the C3′-endo confor-
mation by the side chain, combined with the incompat-
ibility of a C2′-endo conformation in a duplex, may
account in part for the reduced Tm of the AP-containing
oligomers. No NOE is detected from the side chain to
protons of flanking sugars, showing that the AP does
not bind to O4 of the 3′-sugar ring. In a duplex with
RNA, the AP side chain will traverse a broad region in
the vicinity of the flanking phosphate. The pH deter-
mination for the AP side chains shows that the amine
moiety will stay protonated in a zwitterionic form under
physiological conditions.
quence when transfected into cultured cells by elec-
troporation. These results are both sequence-specific
and target-specific, consistent with an antisense mech-
anism of action. Oligonucleotide P contained nine
phosphodiester linkages, and these results demonstrate
that the greatly enhanced nuclease stability observed
for AP oligonucleotides in cell-free systems using SVPD
translates into enhanced stability and antisense activity
in cells. Thus, antisense oligonucleotides containing 2′-
O-AP modifications are attractive alternatives for the
design of effective, nuclease-stable antisense molecules
with reduced phosphorothioate content. Since the hu-
man C-raf antisense sequence employed in this study
has previously been shown to be a potent antitumor
agent in animals when synthesized as a phosphorothio-
ate ODN, future studies examining the antitumor
properties of AP analogs of this and additional se-
quences are highly warranted.
Hybr id iza tion . Substitution of AP nucleotides into
DNA oligomers produces stabilization of a duplex with
complementary RNA roughly equivalent to alkyl chains
of the same length. Duplexes with uniformly AP-
modified sequences are less stable than 2′-O-methyl or
2′-O-propyl sequences but still superior to DNA se-
quences. In a 2′-O-methyl sequence, substitution of AP
residues generates no destabilization. In a winged
phosphorothioate oligonucleotide, the AP nucleotides
stabilize superior to a 2′-O-propyl wing but less than a
2′-O-methyl wing. In an oligomer with phosphodiester
wings, substitution with propylamine residues is de-
stabilizing relative to the other sugar alkyls. Incorpora-
tion of five pentylamine residues reduces the salt
dependence of the Tm similar to C5-substituted pyrim-
idines with amino groups, suggesting that binding of
cations and anions to the strands of the propylamine-
substituted strand:RNA duplex during the dissociation
process generates an unfavorable entropic effect. Sub-
stitution of aminopropyl residues into the sequence
provides improved affinity compared to phosphorothio-
ate oligonucleotides.
Con clu sion s
The 2′-O-propylamine ribonucleotides are incorpo-
rated readily into antisense oligonucleotides using
conventional phosphoramidite monomers and a phthal-
imido protecting group for the amine. A differential
deprotection strategy with an initial ammonia treat-
ment followed by reaction with methylamine cleanly and
rapidly removes all of the protecting groups. The
propylamine side chain is mobile and does not form a
tight ion pair with the flanking 3′-phosphate in molec-
ular dynamics simulations or an NMR analysis of chain
motion. Oligonucleotides incorporating propylamine
groups have reduced net charge but maintain favorable
solubility properties.
No dramatic increase in affinity of oligomers contain-
ing propylamine sugars for RNA complement is ob-
served compared to simple alkyl side chains. This result
suggests that no cratic release of cations from the
phosphate backbone is produced, which would provide
stabilization. The shift in the sugar pucker of 2′-
substituted nucleotides toward C3′-endo geometries may
limit the location of the propylamine nucleotides to the
wings of gapmers to retain RNAse H activity.
Nu clea se Resista n ce. Capping of a phosphodiester
oligonucleotide with two propylamine units provides
nearly complete resistance to 3′-exonuclease degrada-
tion. These capped oligomers are ∼1000-fold more
resistant than a phosphodiester and have greater
resistance than a PdS linkage. This resistance to
exonuclease degradation is independent of pH between
pH 5 and 9. The increase in resistance is not just a
steric effect, since a propyl or longer pentyl side chain
confers only a mild increase in resistance.10 The sugar
conformation does not play a role, since incorporation
of propyl-substituted residues does not produce a dra-
matic increase in nuclease resistance. The net charge
of the oligo is not critical, since the digestion is rapid
up to the first residue 3′ from the AP nucleotide. These
data suggest that location of a positive charge in the
general proximity of the phosphate is sufficient to
inhibit enzymatic activity, presumably through intro-
duction of an unfavorable electrostatic interaction with
a hydrated metal ion or positively charged amino acid
side chain which catalyzes the cleavage process. While
SVPD can cleave through one AP residue, a two-AP
nucleotide cap provides an oligomer with nearly com-
plete 3′-exonuclease resistance.
The enhanced exonuclease resistance of propylamine-
containing oligomers may be related to an electrostatic
repulsion between the side chain and functional groups
at the catalytic site of the enzyme, which must dehy-
drate the phosphate and bring in a positively charged
catalyst such as a hydrated metal ion or charged amino
acid side chain prior to cleavage. The inability of SVPD
to cleave the propylamine-capped phosphodiester is not
just a result of steric hindrance or an unfavorable
nonspecific electrostatic interaction.
The potency of PS oligonucleotides is enhanced through
incorporation of two phosphodiester 2′-AP nucleotides
at the 3′-terminus. These capped oligonucleotides re-
duced levels of mRNA in cells more efficiently and
dramatically reduced protein expression relative to the
PS ODN of the same sequence. The increase in nuclease
resistance without the penalty in affinity for RNA
complement induced by phosphorothioates, and dem-
onstration of improved cellular potency, makes oligo-
mers containing phosphodiester-linked 2′-O-aminopro-
pyl nucleotides leading candidates as antisense thera-
peutic agents in vivo.
In h ibition of C-r a f m RNA a n d P r otein . An an-
tisense oligomer (P ) targeted against human C-raf
kinase containing nine 2′-O-aminopropyl modifications
has been found to be at least 10-fold more potent in
reducing C-raf mRNA and protein expression as com-
pared with a phosphorothioate ODN of the same se-

2′-Aminopropyl Oligonucleotides
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 26 5107
ytrityl chloride (95%) (3.6 g, 10 mmol) was added to a solution
of 4a (3.4 g, 5.4 mmol) in dry pyridine (100 mL) at room
temperature and stirred for 16 h. The solution was concen-
trated in vacuo and purified by chromatography on silica gel
(99/1 EtOAc/TEA) to afford 3.8 g of the product as a foam
(75%): 1H NMR (DMSO) δ 1.74 (m, 2H, CH2), 3.23 (m, 2H,
CH2NHCO), 3.60-3.73 (m, 10H, ArOCH3, C5′CH2, OCH2), 4.09
(m, 1H, C4′H), 4.51 (m, 1H, C3′H) 4.66 (m, 1H, C2′H), 5.34 (d,
1H, C3′OH), 6.19 (d, 1H, C1′H), 6.80-7.78 (m, 23H, ArH), 8.64
(s, 1H, C2H), 8.78 (s, 1H, C8H), 9.35 (s, 1H, NHCO); 19F NMR
(DMSO) δ -75.48. Anal. (C50H45F3N6O9) C, H, N.
5′-O-(Dim eth oxytr ityl)-6-N,N-d iben zoyl-2′-O-[[N-(tr if-
lu or oa cet yl)a m in o]p r op yl]a d en osin e-3′-O-(2-cya n oet h -
yl N,N-d iisop r op ylp h osp h or a m id ite) (6a ). A solution of
5a (6.5 g, 7 mmol) in dry methylene chloride was treated with
bis(N,N-diisopropylamino)cyanoethyl phosphite (1.1 equiv) and
(N,N-diisopropylamino)tetrazolide (catalytic amount) at room
temperature for 16 h. The reaction was concentrated in vacuo
and subjected to chromatography on silica gel (6/4 EtOAc/
hexane, 1% TEA) to afford 6.7 g of the product as a foam
(78%): 1H NMR (CDCl3) 1.0-1.31 (m, 12H, CH3), 1.85 (m, 2H,
CH2), 2.36 (t, 2H, CH2NH), 2.61 (t, 2H, CH2CH), 3.21-3.91
(m, 14H, AROCH3, C4, H, C5, CH2, CHN, OCH2, OCH2), 4.38
(m, 1H, C3, H), 4.62 (m, 1H, C2, H), 6.15 (d, 1H, C1, H), 6.77-
7.82 (m, 23H, ARH) 8.28 (S, 1H, C2H), 8.57 (S, 1H, C8H); 31P
NMR (CD3CN) δ 150.87; (CDCl3) δ 150.55, 150.83; 19F NMR
(CD3CN) δ -77.1; (CDCl3) δ -81.6. Anal. (C59H62F3N8O10P)
C,H,N.
6-N-Ben zoyl-2′-O-(p r op ylp h th a lim id o)a d en osin e (7a ).
Compound 1a (3.0 g, 6.6 mmol) was converted to 7a using
benzoyl chloride via the transient protection procedure de-
scribed above. After workup, concentration of the reaction in
vacuo followed by chromatography on silica gel (9/1 EtOAc/
MeOH) afforded 2.74 g of the product as a foam (74%): 1H
NMR (DMSO) δ 1.81 (m, 2H, CH2), 3.45-3.7 (m, 6H, C5′,
OCH2, PhthCH2), 3.93 (m, 1H, C4′H), 4.3 (m, 1H, C3′H), 4.5
(q, 1H, C2′H), 5.12 (t, 1H, C5′OH), 5.20 (d, 1H, C3′OH), 6.1
(d, 1H, C1′H), 7.5-8.06 (m, 9H, Ar), 8.68 (s, 1H, C2H), 8.71
(s, 1H, C8H), 11.2 (s, 1H, NH). Anal. (C28H26N6O) C,H,N.
5′-O-(Dim et h oxyt r it yl)-6-N-b en zoyl-2′-(p r op ylp h t h a l-
im id o)a d en osin e (8a ). Compound 8a was prepared from 7
(2.7 g, 5.0 mmol) following the procedure described for
compound 5a . Chromatography on silica gel (99/1 EtOAc/TEA)
afforded 3.1 g of the product (75%): 1H NMR (DMSO) δ 1.87
(m, 2H, CH2), 3.19-3.70 (m, 6H, C5′CH2, OCH2, PhthCH2),
4.07 (m, 1H, C4′H), 4.43 (m, 1H, C3′H), 4.63 (m, 1H, C2′H),
5.25 (d, 1H, C3′OH), 6.13 (d, 1H, C1′H), 6.8-8.08 (m, 22H,
Ar), 8.60 (s, 1H, C2H), 8.61 (s, 1H, C8H), 11.1 (s, 1H, NH).
Anal. (C49H44N6O9) C,H,N.
N -Be n zoyl-5′-O-(4,4′-d im e t h oxyt r it yl)-2′-O-(p r op yl-
p h t h a lim id o)a d en osin e-3′-O-(2-cya n oet h yl N,N′-d iiso-
p r op ylp h osp h or a m id ite) (9a ). Compound 9a was prepared
from 8a (2.7 g, 3.2 mmol) according to the procedure for
compound 6a . Chromatography on silica gel (1/1 EtOAc/
hexane, 0.5% TEA) afforded 2.7 g of the product (80%): 1H
NMR (CDCI3) δ 1.0-1.34 (m, 12H, CH3), 1.85 (m, 2H, CH2),
2.63 (t, 2H, CH2CN), 3.24-3.89 (m, 17H, AROCH3, C4, H, C5,
CH2, OCH2, OCH2, CHN, CH2NPHTH), 4.35 (m, 1H, C3, H),
4.60 (m, 1H, C2, H), 6.1 (d, 1H, C1, H), 6.72-7.91 (m, 22H,
ARH), 8.21 (S, 1H, C2H), 8.59 (S, 1H, C8, H); 31P NMR (CD3-
CN) δ 150.88, 151.22. Anal. (C58H61N8O10P) C,H,N.
Exp er im en ta l Section
Gen er a l Meth od s. Bulk chemicals were purchased from
Sigma or Aldrich and were of the highest available purity.
Ultraviolet-visible spectra were recorded on
a Hewlett-
Packard 8452A diode-array spectrophotometer with a Peltier
temperature controller accessory. Melting points (uncorrected)
were determined in an open glass capillary on a Thomas-
Hoover apparatus. Kieselgel 60 F254 plates from E. Merck
were used for TLC, with detection by UV light or H2SO4/MeOH
and heat. Flash chromatography was performed using silica
gel (40 µm; J . T. Baker). Elemental analyses were performed
by Quantitative Technologies (Bound Brook, NJ ). 1H NMR
spectra were recorded on a Gemini 200 (200 MHz) spectrom-
eter (Varian Associates, Palo Alto, CA) using tetramethylsilane
as an internal standard. 31P and 19F NMR studies utilized
phosphoric acid and trifluoroacetic acid as internal references,
respectively.
2′-O-(P r op ylp h th a lim id o)a d en osin e (1a ). Sodium hy-
dride (60%) (4.5 g, 112 mmol) was added to a solution of
adenosine (20.0 g, 75 mmol) in dry dimethylformamide (550
mL) at 5 °C under N2 atmosphere. After 1 h, N-(3-bromopro-
pyl)phthalimide (23.6 g, 86 mmol) was added and the temper-
ature raised to 30 °C and held for 16 h. After cooling, ice was
added and the reaction mixture concentrated to a gum. The
gum was partitioned between water and ethyl acetate and
extracted thoroughly (4 × 300 mL). The organic extracts were
dried and concentrated in vacuo, and the resulting gum was
subjected to chromatography on silica gel (95/5 CH2CL2/MeOH)
to afford a white solid. Recrystallization from MeOH afforded
5.7 g of the 2′-isomer (17%): mp 123-124 °C; 1H NMR (DMSO)
δ 1.70 (m, 2H, CH2), 3.4-3.7 (m, 6H, C5′CH2, OCH2, PhthCH2),
3.95 (q, 1H, C4′H), 4.30 (q, 1H, C3′H), 4.46 (t, 1H, C2′H), 5.15
(d, 1H, C3′OH), 5.41 (t, 1H, C5′OH), 5.95 (d, 1H, C1′H), 7.35
(s, 2H, NH2), 7.8 (brs, 4H, Ar), 8.08 (s, 1H, C2H), 8.37 (s, 1H,
C8H). Anal. (C21H22N6O6) C,H,N.
2′-O-(Am in op r op yl)a d en osin e (2a ). A solution of 1a (8.8
g, 19 mmol), 95% ethanol (400 mL), and hydrazine (10 mL, 32
mmol) was stirred for 16 h at room temperature, whereupon
the reaction mixture was filtered and the filtrate concentrated
in vacuo. Water (150 mL) was added, and the pH was adjusted
to 5.0 with HOAc. The aqueous layer was extracted with ethyl
acetate (2 × 30 mL) and concentrated in vacuo to afford 7.1 g
of the product as the HOAc salt (95%): 1H NMR (DMSO) δ
1.70 (m, 2H, CH2), 2.76 (m, 2H, CH2NH2), 3.55-3.67 (m, 4H,
C5′-CH2, OCH2), 4.0 (q, 1H, C4′H), 4.30 (q, 1H, C3′H), 4.47 (t,
1H, C2′H), 6.0 (d, 1H, C1′H), 7.39 (s, 2H, NH2), 8.16 (s, 1H,
C2H), 8.41 (s, 1H, C8H). Anal. (C13H20N6O4‚HOAc) C,H,N.
2′-O-[[N-(Tr iflu or oacetyl)am in o]pr opyl]aden osin e (3a).
A solution of 2a (6.0 g, 16 mmol) in methanol (50 mL) and
triethylamine (15 mL, 107 mmol) was treated with ethyl
trifluoroacetate (18 mL, 151 mmol). After 16 h, the reaction
mixture was concentrated in vacuo and the resultant gum
subjected to chromatography on silica gel (9/1 EtOAc/MeOH)
to afford 3.3 g of the product (54%): mp 200-201 °C; 1H NMR
(DMSO) δ 1.7 (m, 2H, CH2), 3.20 (m, 2H, CH2NHCO), 3.37-
3.67 (m, 4H, C5′CH2, OCH2), 4.00 (q, 1H, C4′H), 4.35 (q, 1H,
C3′H), 4.49 (t, 1H, C2′H), 5.26 (d, 1H, C3′OH), 5.43 (t, 1H,
C5′OH), 6.02 (d, 1H, C1′H), 7.38 (s, 2H, NH2), 8.16 (s, 1H,
C2H), 8.40 (s, 1H, C8H), 9.34 (s, 1H, NHCO); 19F NMR (DMSO)
δ -75.48. Anal. (C15H19F3N6O5) C,H,N.
6-N,N-Diben zoyl-2′-O-[[N-(tr iflu or oa cetyl)a m in o]p r o-
p yl]a d en osin e (4a ). Compound 3a (12.6 g, 30 mmol) was
converted to 4a using a modification of the procedure of Ti et
al.15 Tetrabutylammonium fluoride was utilized instead of
ammonium hydroxide in the workup. Concentration of the
reaction mixture followed by chromatography on silica gel (1/1
EtOAc/MeOH) afforded 6.8 g of the product as a foam (36%):
1H NMR (DMSO) δ 1.72 (m, 2H, CH2), 3.20 (m, 2H, CH2-
NHCO), 3.45-3.68 (m, 4H, C5′CH2, OCH2), 4.03 (q, 1H, C4′H),
4.37 (q, 1H, C3′H), 4.53 (t, 1H, C2′H), 5.18 (t, 1H, C5′OH),
5.34 (d, 1H, C3′OH), 6.17 (d, 1H, C1′H), 7.45-7.83 (m, 10H,
Ar), 8.73 (s, 1H, C2H), 8.90 (s, 1H, C8H), 9.37 (s, 1H, NHCO);
19F NMR (DMSO) δ -75.48. Anal. (C28H27F3N6O7) C, H, N.
5′-O-(Dim eth oxytr ityl)-6-N,N-d iben zoyl-2′-O-[[N-(tr if-
lu or oa cetyl)a m in o]p r op yl]a d en osin e (5a ). 4,4′-Dimethox-
5′-O-(4,4′-Dim eth oxytr ityl)-2′-O-(p r op ylp h th a lim id o)-
u r id in e-3′-O-(2-cya n oeth yl N,N-d iisop r op ylp h osp h or a -
m id ite) (9b). The 2′-(propylphthalimido)uridine nucleoside
was prepared as described for 7a , except that the alkylation
proceeded through the stannyl derivative. The nucleoside was
protected as the 5′-DMT derivative via reaction with DMT-Cl
in pyridine and purified by chromatography on silica gel. This
material was converted to the phosphoramidite using bis(N,N-
diisopropyamino)cyanoethyl phosphite.15 Following aqueous
extraction with bicarbonate, a solution of 9b was triturated
into petroleum ether. The resulting precipitate was collected
and evaporated twice from acetonitrile: 1H NMR (CD3CN) δ
0.99-1.21 (m, 12H, CH3), 1.82 (m, 2H, CH2), 2.51 (t, 2H, CH2-
NH), 2.64 (t, 2H, CH2CN), 3.2-4.19 (m, 14H, OCH2, C5, CH2,
CHN, AROCH3, C4, H, C3, H), 4.35-4.61 (m, 1H, C2, H), 5.23

5108 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 26
Griffey et al.
(d, 1H, CHd), 5.82 (d, 1H, C1, H), 6.82-7.52 (m, 14H, ARH),
7.61 (brs, 1H, NHCO), 7.75 (d, 1H, CHd), 8.85 (brs, 1H, ArCH);
31P NMR (CD3CN) δ 154.87, 155.66.
approximately 0.2 M was mimicked by the addition of one Na+
ion and one Cl- ion by randomly displacing two water
molecules. An all-atom CFF91 force field was employed in our
calculations (Biosym Technologies, San Diego, CA). A three-
point flexible water model was used, in which each atom
possessed a partial point charge (O, -0.82 eu; H, +0.41 eu).
Electrostatic interactions were computed via Coulomb’s law,
using a dielectric constant of 1.0. The cutoff used for the
intermolecular interactions was 9.0 Å. Periodic boundary
conditions were applied in all directions. The simulation was
carried out with a time step size of 1 fs, for a total of 150 ps at
constant volume and constant temperature. The simulation
was initiated by 1000 steps of steepest descent minimization.
Initial velocities were assigned from a Maxwell-Boltzmann
distribution at 300 K.
N-Ben zoyl-5′-O-(4,4′-d im et h oxyt r it yl)-2′-(3-N-p h t h a l-
im id op r op yl)cytid in e-3′-O-(2-cya n oeth yl N,N-d iisop r o-
p ylp h osp h or a m id ite) (9c). The 2′-(3-N-phthalimidopropyl)-
cytidine nucleoside was prepared following the protocol of
Guinosso.15,23 The nucleoside was protected as the 5′-DMT
derivative via reaction with DMT-Cl in pyridine and purified
by chromatography on silica gel. This material was converted
to the N-benzoyl derivative and the phosphoramidite using bis-
(N,N-diisopropylamino)cyanoethyl phosphite. Following aque-
ous extraction with bicarbonate, a solution of 9c was triturated
into petroleum ether. The resulting precipitate was collected
and evaporated twice from acetonitrile: 1H NMR (CDCl3) δ
1.11-1.32 (m, 12H, CH3), 1.99 (m, 2H, CH2), 2.76 (t, 2H, CH2-
CN), 3.42-4.21 (m, 16H, OCH2, OCH2, CHN, C5, CH2, CH2-
NPHTH, AROCH3, C4, H, C3, H), 4.52 (m, 1H, C2, H), 6.01 (S,
1H, C1, H), 6.90 (d, 1H, CHd), 7.0-7.92 (m, 19H, ARH, ArNH),
8.63 (d, 1H, CHd); 31P (CDCI3) δ 150.35.
N-Isobu tyr yl-5′-O-(4,4′-d im eth oxytr ityl)-2′-O-(3-N-p h -
th a lim id op r op yl)gu a n osin e-3′-O-(2-cya n oeth yl N,N-d i-
isop r op ylp h osp h or a m id ite) (9d ). The 2′-(3-N-phthalimi-
dopropyl)guanosine nucleoside was prepared following the
protocol of Guinosso.15,23 The nucleoside was protected as the
5′-DMT derivative via reaction with DMT-Cl in pyridine,
converted to the N-isobutyryl derivative, and purified by
chromatography on silica gel. This material was converted
to the phosphoramidite using bis(N,N-diisopropylamino)cya-
noethyl phosphite. Following aqueous extraction with bicar-
bonate, a solution of 9d was triturated into petroleum ether.
The resulting precipitate was collected and evaporated twice
from acetonitrile: 1H NMR (CDCl3) δ 0.90-1.3 (m, 18H, CH3),
1.98 (m, 2H, CH2), 2.16 (m, 1H, CHCO), 2.65 (t, 2H, CH2CN),
3.29-3.97 (m, 16H, OCH2, OCH2, CH2NPHTH, CHN, C51CH2,
AROCH3), 4.31-4.54 (m, 3H, C2, H, C3, H, C4, H), 5.90 (d, 1H,
C1, H), 6.79-7.5 (m, 17H, ARH), 7.85 (S, 1H, C8H); 31P NMR
(CDCI3) δ 150.496, 150.885.
DNA Syn th esis. DNA phosphoramidites were purchased
from Millipore (Boston, MA). Oligonucleotides were synthe-
sized on controlled pore glass solid support using conventional
phosphite triester chemistry with an Applied Biosystems
(Foster City, CA) 380B synthesizer. Phosphorothioate back-
bones were prepared using Beaucage reagent during the
oxidation step.25 Cleavage from support and deprotection were
accomplished by treatment with concentrated ammonium
hydroxide for 15 h at 55 °C unless otherwise stated. The
purity of oligonucleotides was monitored using capillary elec-
trophoresis (model 270A; ABI, Foster City, CA) or gel chro-
matography and confirmed using electrospray mass spectrom-
etry.
Hybr id iza tion Sta bility. Absorbance versus temperature
curves of duplexes were measured at a 4 µM strand concentra-
tion in a 10 mM phosphate buffer at pH 7.4 containing 100
mM Na+ and 0.1 mM EDTA. The Tm was determined from
fits of data to a two-state model with linear sloping baselines.10
The salt dependence of the Tm for duplexes were determined
using Na+ concentrations ranging from 0.005 to 0.6 M.
ES-MS. All experiments were performed in negative ion-
ization mode using a 5998 quadrupole MS instrument with a
59987A electrospray unit (Hewlett-Packard, Palo Alto, CA).
Samples were desalted via ethanol precipitation from 10 M
ammonium acetate at -20 °C.26 Typically, a 0.1 A260 unit
sample of oligo was dissolved in 100 µL of a solution containing
1/1 2-propanol/water. The sample was then infused at a rate
of 7.0 µL/min. Spectra were obtained over 5 min and signal-
averaged without filtering. At least three ions with differing
values of m/ z were identified, and the HP software deconvo-
lution algorithm was employed to calculate an average mo-
lecular mass for the molecule.
Nu clear Magn etic Reson an ce. All two-dimensional NMR
experiments were performed using a 400 MHz Unity (Varian
Associates, Palo Alto, CA) spectrometer. 1H NMR spectra were
referenced relative to internal TMS, while 31P NMR spectra
were acquired at 161.4 MHz using H3PO4 as an external
standard. Two-dimensional NOESY, ROESY, and TOCSY
spectra were acquired using conventional pulse sequences.
Nu clea se Digestion . Oligonucleotide derivatives of se-
quence O were purified by polyacrylamide gel electrophoresis
followed by desalting with Poly-pak cartridges (Glen Research,
Sterling, VA). Oligomers were 5′-end-labeled using [γ-32P]ATP
and T4 polynucleotide kinase. Stability assays were performed
using 5 × 10-2 or 5 × 10-3 U/mL SVPD (USB, Cleveland, OH)
and 1 µM solutions of oligonucleotide at 37 °C in a buffer of
50 mM Tris-HCl, pH 8.5, with 72 mM sodium chloride and 14
mM magnesium chloride. The enzyme was active for >24 h
under these conditions. Aliquots of the reaction mixture were
removed at the indicated times, quenched by addition to an
equal volume of 80% formamide gel loading buffer, and heated
for 2 min at 95 °C. Quantitation was performed by reading
intensities from
a phosphorimager (Molecular Dynamics,
Sunnyvale, CA). The listed values of t1/2 were determined for
loss of the 3′-terminal nucleotide.
Oligon u cleotid e Tr ea tm en t of Cells. The human blad-
der carcinoma cell line T24 (American Type Tissue Collection,
Bethesda) grown in T175 flasks to 75% confluency were
trypsinized and resuspended in complete tissue culture media
(McCoys 5A media supplemented with 10% fetal bovine serum)
at a density of 1 × 107 cells/mL. A total of 3.6 × 106 cells were
employed for each treatment by combining 360 µL of cell
suspension with oligonucleotide at the indicated concentrations
to reach a final volume of 400 µL. Cells were then transferred
to an electroporation cuvette and electroporated using an
Electrocell Manipulator 600 instrument (Biotechnologies and
Experimental Research Inc.) employing 350 V, 100 µF, at 13
Ω. Electroporated cells were then transferred to conical tubes
containing 5 mL of complete media, mixed by inversion, and
plated onto 10 cm culture dishes. Analysis of mRNA levels
by Northern blot was initiated 20 h following electroporation,
while analysis of protein levels by Western blot was initiated
48 h following electroporation.
Nor th er n Blot a n d Wester n Blot An a lyses. For deter-
mination of mRNA levels by Northern blot, total RNA was
prepared from cells by the guanidinium isothiocyanate pro-
cedure.27 Total RNA was isolated by centrifugation from cell
lysates. RNA was quantitated and normalized to levels of
glycerol-3-phosphate dehydrogenase mRNA using a Molecular
Dynamics Phosphoimager as described previously.27
For determination of protein levels by Western blot, cellular
extracts were prepared using 250 µL of RIPA extraction buffer/
10 cm dish. A 25 µg lot of protein was then separated by
electrophoresis on a 10% SDS-polyacrylamide minigel (Bio-
Rad). Once transferred, membranes were treated for 2 h with
a monoclonal antibody that specifically recognizes C-raf kinase
protein (Transduction Laboratories) at a dilution of 1:1000
followed by incubation with 5 µCi of 125I-labeled goat anti-
mouse antibody (ICN Radiochemicals) for 1 h. Labeled
proteins were visualized and quantitated by PhosphorImage
analysis.
Molecu la r Dyn a m ics. The following protocol has been
used in the present molecular dynamics study. A dinucleotide
was located in a cubical box of length 20.0 Å. Solvating this
solute with pre-equilibrated water molecules and removing bad
contacts to obtain a density of approximately 1 g/cm3 left a
total of 232 waters in the central box. A salt concentration of

2′-Aminopropyl Oligonucleotides
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 26 5109
(14) McGee, D. P.; Cook, P. D.; Guinosso, C. J . Novel 2′-O-alkyl-
nucleosides and phosphroamidites: processes for the preparation
and uses thereof. PCT WO 94/02501, 1994.
Ack n ow led gm en t. We wish to thank Stan Crooke
for helpful suggestions.
(15) Ti, G. S.; Gaffney, B. L.; J ones, R. A. Transient protection:
efficient one-flask syntheses of protected deoxynucleosides. J .
Am. Chem. Soc. 1982, 104, 1316-1319.
(16) Wolfe, S.; Hasan, S. K. Five-membered rings. II. Inter and
intramolecular reactions of simple amines with N-substituted
Refer en ces
(1) Crooke, S. T. Oligonucleotide therapeutics. In Burger’s Medicinal
Chemistry and Drug Discovery, Vol. 1: Principles and Practice;
Wolff, M. E., Ed.; J ohn Wiley & Sons, Inc.: New York, 1995; pp
863-900.
(2) Crooke, S. T. Oligonucleotide therapy. Curr. Opin. Biotechnol.
1992, 3, 656-661.
phthalimides. Methylamine as
a reagent for removal of a
phthaloyl group from nitrogen. Can. J . Chem. 1970, 48, 3572-
3579.
(3) Uhlmann, E.; Peyman, A. Antisense oligonucleotides: a new
therapeutic principle. Chem. Rev. 1990, 90, 544-579.
(4) Sanghvi, Y. S.; Cook, P. D. Towards second-generation synthetic
backbones for antisense oligonucleosides. In Nucleosides and
Nucleotides as Antitumor and Antiviral Agents; Chu, C.K.,
Baker, D., Eds.; Plenum Press: New York, 1993; pp 311-324.
(5) DeMesmaeker, A.; Waldner, A.; Lebreton, J .; Hoffmann, P.;
Fritsch, V.; Wolf, R. M.; Freier, S. M. Amides as a new type of
backbone modification in oligonucleotides. Angew. Chem., Int.
Ed. Engl. 1994, 33, 226-229.
(17) Reddy, M. P.; Hanna, N. B.; Farooqui, F. Fast cleavage and
deprotection of oligonucleotides. Tetrahedron Lett. 1994, 35,
4311-4314.
(18) Mohan, V.; Griffey, R. H.; Davis, D. R. Structure and dynmaics
of MMI linked nucleotides. Tetrahedron 1995, 51, 6855-6868.
(19) De Leeuw, F. A. A. M.; Altona, C. Computer-assisted pseudoro-
tation analysis of five-membered rings by means of proton spin-
spin coupling constants: program PSEUDOROT. J . Comput.
Chem. 1983, 4, 428-437.
(20) Cummins, L. L.; Owens, S. R.; Risen, L. M.; Lesnik, E. A.; Freier,
S. M.; McGee, D.; Guinosso, C. J .; Cook, P. D. Characterization
of fully 2′-modified oligoribonucleotide hetero- and homoduplex
hybridization and nuclease sensitivity. Nucleic Acid Res. 1995,
23, 2019-2024.
(21) Kiely, J . S. Recent advances in antisense technology. Annu. Rep.
Med. Chem. 1994, 29, 297-306.
(22) Crooke, S. T.; Leminidis, K. M.; Neilson, L.; Griffey, R. H.;
Lesnik, E. A.; Monia, B. P. Kinetic characteristics of Escherichia
coli Rnase H1: cleavage of various antisense oligonucleotide-
RNA duplexes. Biochem. J . 1995, 312, in press.
(23) C. Guinosso, private communication.
(24) Nogle, J . H.; Schirmer, R. E. The Nuclear Overhauser Effect;
Academic Press: New York, 1971.
(25) Iyer, P. R.; Phillips, L. R.; Egan, W.; Regan, J . B.; Beaucage, S.
The automated synthesis of sulfur-containing oligodeoxy-
ribonucleotides using 3H-1,2-benzodithiol-3-one 1,1-dioxide
as a sulfur-transfer reagent. J . Org. Chem. 1990, 55, 4693-
4698.
(26) Greig, M.; Griffey, R. H. Utility of organic bases for improved
electrospray mass spectrometry of oligonucleotides. Rapid Com-
mun. Mass Spectrom. 1995, 5, 97-102.
(27) Monia, B. P.; J ohnston, J . F.; Geiger, T.; Muller, M.; Fabbro, D.
Antitumor activity of a phsophorothioate antisense oligonucle-
otide targeted against C-raf kinase. Nature Med. 1996, 2, 668-
675.
(28) Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D.,
Seidman, J . G., Smith, J . A., Struhl, K., Eds. Introduction of
DNA into mammalian cells. In Current protocols in molecular
biology; Green Publishing Associates and Wiley Interscience:
New York, 1987.
(6) Nielsen, P. E.; Eghom, M.; Berg, R. H.; Buchardt, O. Peptide
Nucleic Acids (PNA). Potential anti-sense and anti-gene agents.
Anticancer Drug Des. 1993, 8 (1), 53-63.
(7) Browne, K. A.; Dempcy, R. O.; Bruice, T. C. Binding studies of
cation thymidyl deoxyribonucleic guanidine to RNA homopoly-
nucleotides. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 7051-7055.
(8) Egholm, M.; Buchardt, O.; Christensen, L.; Behrens, C.; Freier,
S. M.; Driver, D. A.; Berg, R. H.; Kim, S. K.; Norden, B.; Nielsen,
P. E. PNA hybridizes to complementary oligonucleotides obeying
the Watson-Crick hydrogen bonding rules. Nature 1993, 365,
566-568.
(9) Sanghvi, Y. S.; Cook, P. D. Carbohydrates: Synthetic methods
and applications in antisense therapeutics: An overview. In
Carbohydrate Modifications in Antisense Research; Sanghvi, Y.
S., Cook, P. D., Eds.; ACS Symposium Series No. 580; American
Chemical Society: Washington, DC, 1994; pp 1-22.
(10) Lesnik, E. A.; Guinosso, C. J .; Kawasaki, A. M.; Sasmor, H.;
Zounes, M.; Cummins, L. L.; Ecker, D. J .; Cook, P. D.; Freier, S.
M. Oligodeoxynucleotides containing 2′-O-modified adenosine:
Synthesis and effects on stability of DNA:RNA duplexes. Bio-
chemistry 1993, 32 (30), 7832-7838. Sanghvi, Y. S.; Bellon, L.;
Morvan, F.; Hoshiko, T.; Swayze, E.; Cummins, L.; Freier, S.;
Dean, N.; Monia, B.; Griffey, R. H.; Cook, P. D. Synthesis,
biophysical and biological evaluations of novel antisense oligo-
nucleosides containing dephosphono-internucleosidic linkages.
Nucleosides Nucleotides 1995, 14, 1087-1090.
(11) Hashimoto, H.; Nelson, M. G.; Switzer, C. Formation of chimeric
duplexes between zwitterionic and natural DNA. J . Org. Chem.
1993, 58, 4194-4195. Huang, C.-Y.; Cushman, C. D.; Miller, P.
S. Triplex formation by an oligonucleotide containing N4-(3-
acetamidopropyl)cytidine. J . Org. Chem. 1993, 58, 5048-5049.
(12) Hashimoto, H.; Nelson, M. G.; Switzer, C. Zitterionic DNA. J .
Am. Chem. Soc. 1993, 115, 7128-7134.
(13) Wagner, D.; Verheyden, J . P. H.; Moffat, J . G. Preparation and
synthetic utility of some organotin derivatives of nucleosides.
J . Org. Chem. 1974, 39, 24-30.
J M950937O