Stereo-enriched phosphorothioate oligodeoxynucleotides: Synthesis, biophysical and biological properties

Article abstract of DOI:10.1016/S0968-0896(99)00275-8

Stereo-enriched [Rp] and [Sp]-phosphorothioate oligodeoxynucleotides are synthesized using oxazaphospholidine derivatized monomers. Three different designs of phosphorothioate oligodeoxynucleotides (PS-oligos), (i) stereo- enriched all-[Rp] or all-[Sp] PS

Full text of DOI:10.1016/S0968-0896(99)00275-8

Bioorganic & Medicinal Chemistry 8 (2000) 275±284
Stereo-Enriched Phosphorothioate Oligodeoxynucleotides:
Synthesis, Biophysical and Biological Properties
Dong Yu, a Ekambar R. Kandimalla, a Allysen Roskey, a Qiuyan Zhao, a Lihong Chen, b
Jiangdong Chen b and Sudhir Agrawal a,*
aHybridon, Inc., 155 Fortune Boulevard, Milford, MA 01757, USA
bDepartment of Microbiology, Louisiana State University Medical Center, New Orleans, LA 70112, USA
Received 22 July 1999; accepted 7 October 1999
AbstractÐStereo-enriched [Rp] and [Sp]-phosphorothioate oligodeoxynucleotides are synthesized using oxazaphospholidine deri-
vatized monomers. Three di€erent designs of phosphorothioate oligodeoxynucleotides (PS-oligos), (i) stereo-enriched all-[Rp] or
all-[Sp] PS-linkages, (ii) stereo-random mixture of PS-linkages, and (iii) segments containing certain number of stereo-enriched [Rp]
and [Sp] PS-linkages ([Sp-Rp-Sp] or [Rp-Sp-Rp]), have been studied. Thermal melting studies of these PS-oligos with RNA com-
plementary strands showed that the binding anities are in the order [Rp] > [Sp-Rp-Sp]=[Rp-Sp-Rp] > stereo-random > [Sp].
Circular dichroism (CD) studies suggest that the stereochemistry of the PS-oligo does not a€ect the global conformation of the
duplex. The in vitro nuclease stability of these PS-oligos is in the order [Sp] > [Sp-Rp-Sp] > stereo-random > [Rp]. The RNase H
activation is in the order [Rp] > stereo-random > [Rp-Sp-Rp] > [Sp] > [Sp-Rp-Sp]. Studies in a cancer cell line of PS-oligos
targeted to MDM2 mRNA showed that all oligos had similar biological activity under the experimental conditions employed.
Protein- and enzyme-binding studies showed insigni®cant stereo-dependent binding to proteins. The [Sp] and [Sp-Rp-Sp] chimeric
and stereo-random PS-oligos that contained a CpG motif showed higher cell proliferation than [Rp] PS-oligo of the same sequence.
# 2000 Elsevier Science Ltd. All rights reserved.
Introduction
polymerases and nucleases interact and exert action
on PS-oligos in a diastereomeric-selective fashion.3
Although there are no detailed comparative structural
studies on stereo-regular [Sp] and [Rp] phosphoro-
thioate containing duplexes, molecular modeling studies
of the heteroduplexes containing stereo-regular [Sp] and
[Rp] phosphorothioate strands suggested that in the [Sp]
stereomer sulfur atom is directed towards the minor
groove and in the [Rp] stereomer sulfur atom is pointed
into the major groove of the heteroduplex.8 In vivo
pharmacokinetic and safety pro®les of stereo-regular
PS-oligos are not known because of the lack of con-
venient synthetic methods to prepare diastereomerically
pure PS-oligos in sucient quantities required for these
studies.
Phosphorothioate oligodeoxynucleotides (PS-oligos) are
being evaluated in clinical trials for ecacy against viral
and in¯ammatory diseases and cancers.1,2 As a result of
sulfur substitution for a non-bridging oxygen on phos-
phorus, chirality is introduced at the phosphorus center
resulting in [Rp] and [Sp] diastereomers.3 Our previous
studies using enzymatically synthesized stereo-regular
[Rp] PS-oligo in comparison with a stereo-random PS-
oligo showed stereochemistry-dependent biophysical
and biochemical properties.4 Subsequent studies of
chemically synthesized stereo-regular [Rp] PS-oligos
further con®rmed higher binding anity to the target
RNA, better activation of RNase H enzyme, and faster
degradation by 30-exonucleases than stereo-regular [Sp]
and stereo-random PS-oligos.5,6 Stereo-regular PS-
oligos showed di€erent properties of triple helix formation
in Pu.Pu:Py and Py.Pu:Py motifs.7 In addition, certain
PS-oligos are obtained as a mixture of 2n diastereomers,
(n is number of phosphorothioate internucleoside lin-
kages) on an automated DNA synthesizer using b-cya-
noethylphosphoramidite chemistry. Diastereomerically
pure PS-oligos can be obtained by HPLC puri®cation
of random mixtures obtained by b-cyanoethylphos-
phoramidite chemical synthesis.9 Such a methodology,
however, is limited to pentamer or shorter PS-oligos
Keywords: anticancer; antisense; oligonucleotides; phosphorothioates;
stereo-enriched.
*Corresponding author. Tel.: +1-508-482-7000; fax: +1-508-482-
7692; e-mail: sagrawal@hybridon.com
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00275-8

276
D. Yu et al. / Bioorg. Med. Chem. 8 (2000) 275±284
and the yields of stereo-regular isomers obtained does
not provide quantities required for detailed studies. In
recent years e€orts have been focused on developing
methods for the synthesis of stereo-regular PS-oligos.6,10±14
Nucleoside oxathiaphospholane synthons have been
developed and used for stereoselective synthesis of
PS-oligos in recent years.6,11 This approach, however,
requires prior chromatographic separation of the
individual P-diastereomers of nucleoside oxathia-
phospholane synthons, which are formed in 55:45
ratio.11 Nucleoside oxathiaphospholane synthons
require longer coupling times than standard b-cya-
noethylphosphoramidite coupling cycles for ecient
synthesis of stereo-regular PS-oligos. In addition, the
presence of unprotected phosphorothioate groups gen-
erated during the coupling cycles may potentially lead
to the formation of side products during oligonucleotide
chain elongation.11
lengths and base compositions (Scheme 1). Stereo-
selectivity of coupling of each 50-DMTr-protected oxa-
zaphospholidine monomer (a±h) to a thymine nucleo-
side attached to a controlled pore glass (CPG) support
was determined by isolating and analyzing the dimers
on HPLC as reported earlier for a TT dimer.15 Stereo-
selectivity of each dimer is shown in Table 1. These data
suggest that the stereo-selectivity of the coupling is
dependent on the nucleoside base.
Both [Rp] and [Sp] PS-oligos were synthesized on CPG
solid support using appropriate 50-DMTr-protected
oxazaphospholidine monomers (Scheme 1). The stereo-
random PS-oligo was synthesized using b-cyanoethyl-
phosphoramidite chemistry producing a diastereomeric
mixture. The chimeric PS-oligos containing [Sp] seg-
ments at the 30 and 50-ends and [Rp] segments in the
middle (or vice versa) were synthesized using appro-
priate oxazaphospholidine monomers (Table 2). The
stereo-enriched PS-oligos were deprotected and puri®ed
by polyacrylamide gel electrophoresis (PAGE). The
purity of the oligonucleotides was determined by 31P
NMR, reverse-phase HPLC, capillary gel electrophoresis
(CGE) and denaturing PAGE. Diastereomeric purity
was assessed by nuclease (snake venom phosphodiester-
ase and nuclease P1) digestion assays as described ear-
lier.3 The oligonucleotide sequences synthesized are
shown in Table 2. We have used two di€erent sequences
that have been studied extensively in the past. The PKA
sequence shown in Table 2 is complementary to an 18-
base long site on the mRNA of regulatory subunit Ia
(RIa) of human protein kinase A (PKA).17,18 The
MDM2 oligo sequences (Table 2) are complementary to
a 20-base long sequence on the mRNA of human
MDM2 gene.19
In an e€ort to develop cost-e€ective and ecient meth-
odologies for the stereo-controlled synthesis of phos-
phorothioate oligonucleotides, we recently reported the
use of (S)-(+)-2-pyrrolidine-methanol (l-prolinol), and
(R)-( )-2-pyrrolidine-methanol (d-prolinol) for the
synthesis of 30-bicyclic oxazaphospholidine nucleo-
sides.15 Importantly, the reaction between the chloro-
phosphite reactant and the N-protected 50-O-DMTr
nucleoside gives a single diastereomer of bicyclic oxaza-
phospholidine nucleoside minimizing further separation
of the two diastereomers. Using oxazaphospholidine
derivatized nucleosides and employing automated solid
phase DNA synthesis, we obtained stereo-enriched PS-
dinucleotides.15 We have recently extended this chem-
istry for the synthesis of PS-20-O-methyl dinucleotides.16
In continuation, now we have extended the methodol-
ogy for the synthesis of longer PS-oligos (up to 20-mers)
containing stereo-enriched all [Rp] or all [Sp] linkages.
We have also synthesized PS-oligos containing a central
[Rp] segment and [Sp] segments on either side or vice
versa using appropriate oxazaphospholidine nucleoside
building blocks. Our goal is to explore the potential of
stereo-enriched PS-oligos as antisense agents. We report
here, the results of stereo-enriched PS-oligos for their
target binding anity by UV thermal melting studies, and
the e€ect on the global conformation of the duplexes of
the stereo-enriched PS-oligos with the RNA and DNA
target strands by circular dichroism (CD) studies. We
have also examined the nuclease stability towards snake
venom phosphodiesterase, human DNA polymerase I,
and human serum, RNase H activation, protein binding,
cell proliferation, and biological activity in cell cultures,
and compared the results with a stereo-random PS-oligo
of the same sequence and length synthesized by using
b-cyanoethylphosphoramidite chemistry.
UV thermal melting studies
UV thermal melting studies of the duplexes of PS-oligos
with their target RNA and DNA strands were carried
out. Thermal melting curves of the duplexes of PS-oligos
1±4 with the target RNA are shown in Figure 1A and
the Tms measured for each PS-oligo duplex are shown
in Table 3. These data suggest that [Rp], [Sp] and
stereo-random PS-oligos have di€erent binding anities
to the target RNA. The [Rp] PS-oligo 2 formed a duplex
with the RNA target that has a Tm of 67.7 ꢀC (Table 3).
The duplex of the [Sp] PS-oligo 3 and the RNA target
has a Tm 8.7 ꢀC lower than that obtained for the duplex
of 2 and the RNA target. These data suggest that [Rp]
con®guration is more favorable for binding to the RNA
than the [Sp] con®guration; these ®ndings are consistent
with earlier reports.3±6 The duplex of the stereo-random
PS-oligo 1 and the RNA target has a Tm that is inter-
mediate to the Tms of the duplexes of [Rp] (2) and [Sp]
(3) PS-oligos of the same sequence and the RNA target.
The thermal stability of the duplex of chimeric PS-oligo
4 and the RNA target suggests that the six [Sp] linkages
(three at each end) reduced Tm only by 1.3 ꢀC compared
to [Rp] PS-oligo 2 duplex with the same target strand.
Results and Discussion
Synthesis
We have examined whether the new chemistry could be
extended to synthesize [Rp] and [Sp] PS-oligos of varying
The duplexes of PS-oligos 1±4 with the DNA target
have similar Tms (60.3‹1.0 ꢀC) (Table 3). The thermal

D. Yu et al. / Bioorg. Med. Chem. 8 (2000) 275±284
277
Scheme 1. Synthesis of stereo-enriched PS-oligos. a±d and e±h are obtained from (S)-(+)-2-pyrrolidine methanol and (R)-( )-2-pyrrolidine metha-
nol, respectively. (1) The 50-O-DMTr group of the nucleobase attached to CPG solid support is cleaved by dichloroacetic acid (DCA) in dichloro-
methane (DCM). (2) Coupling of the incoming 50-O-DMTr bicyclic oxazaphospholidine nucleoside a±d or e±h, tetrazole is an activator; B stands for
base, a and e=T; b and f=Abz; c and g=Gibu; d and h=Cbz. (3) Sulfurization (oxidation) was carried out with 3H-1,2-benzodithiole-3-one-1,1-
dioxide. (4) Capping of the unreacted hydroxy groups. (5) At the end of the synthesis, PS-oligos were cleaved, deprotected and puri®ed as described
in Experimental.
Table 1. Stereo-selectivity of synthesis of dinucleotide phosphor-
othioates
and C bases, while the MDM2 sequence contains only
50% G and C bases.
Bicyclic oxazaphospholidine
nudeoside monomer
Dimera
Ratio of stereo-selectivityb
[Rp]:[Sp]
CD studies
a
b
c
d
e
f
50-dTT
50-dAT
50-dGT
50-dCT
50-dTT
50-dAT
50-dGT
50-dCT
12:88
14:86
10:90
13:87
91:9
93:7
92:8
92:9
CD experiments have been carried out in order to
examine whether the duplexes of the two diastereomers
with RNA and DNA targets exhibit any di€erences in
the global conformations. The CD spectra of the
duplexes of PS-oligos 1±4 with the RNA/DNA strands
are shown in Figure 1B and C. The duplexes of both the
[Rp]. [Sp] (2 and 3) and stereo-random (1) PS-oligos
with the DNA target sequence have similar CD spectral
characteristics with slight di€erences in band intensities
(Fig. 1B). The spectra of the duplexes of PS-oligos with
the DNA target strand exhibited B-type spectral char-
acteristics. We did not observe a C-type conformational
CD spectrum for the duplex of [Sp] PS-oligo and DNA
target strand as reported for a duplex of an [Sp] phos-
phorothioate pentamer and its DNA complementary
strand.9
g
h
aBase shown as plane letter indicates CPG attached, the incoming
nucleoside (a±h) is shown in bold.
bStereo-selectivity of coupling of each nucleoside synthon in dimer is
determined by analytical HPLC as described in Experimental.
melting studies of the duplexes of PS-oligos with the
DNA target strand suggest that binding to the DNA
complementary strand is less dependent on the stereo-
chemistry compared to the RNA complementary strand.
Similar thermal melting trends were observed for
duplexes of MDM2 PS-oligos 5±10 with the RNA target
(Table 3). The di€erence in the Tms of duplexes of [Rp]
and [Sp] MDM2 PS-oligos was not as pronounced as in
the case of PKA oligos. This could be as a result of the
di€erence in the base composition of the two oligonu-
cleotide sequences. The PKA sequence contains 72% G
The CD spectra of the duplexes of the PS-oligos and the
RNA target showed a larger positive CD band at
around 267 nm and two negative CD bands at around
235 (shallow) and 210 nm (deep) characteristic of A-
type conformation (Fig. 1C).20,21 The CD intensities
of the three duplexes of the PS-oligos with the RNA

278
D. Yu et al. / Bioorg. Med. Chem. 8 (2000) 275±284
Table 2. Oligonucleotide sequences
No.
PS-oligoa
Sequence
PKA oligos
1
2
3
4
SR
[Rp]
50-d(GCGTGCCTCCTCACTGGC)-30
50-d(GrCrGrTrGrCrCrTrCrCrTrCrArCrTrGrGrC)-30
50-d(GsCsGsTsGsCsCsTsCsCsTsCsAsCsTsGsGsC)-30
50-d(GsCsGsTrGrCrCrTrCrCrTrCrArCrTsGsGsC)-30
[Sp]
[Sp-Rp-Sp]-chimer
MDM2 oligos
5
6
7
8
SR
[Rp]
[Sp]
50-d(TGACACCTG-ITCTCACTCAC)-30
50-d(TrGrArCrArCrCrTrGrTrTrCrTrCrArCrTrCrArC)-30
50-d(TsGsAsCsAsCsCsTsGsTsTsCsTsCsAsCsTsCsAsC)-30
50-d(TrGrArCrArCrCrTsGsTsTsCsTrCrArCrTrCrArC)-30
50-d(TsGsAsCsArCrCrTrGrTrTrCrTrCrAsCsTsCsAsC)-30
50-d(TGACTCTTGTACTTACTCAC)-30
[Rp-Sp-Rp]-chimer
[Sp-Rp-Sp]-chimer
Sr control
9
10
Oligos containing CpG motif
11
12
13
14
SR
[Rp]
50-d(TCCATGACGTTCCTGATGC)-30
50-d(TrCrCrArTrGrArCrGrTrTrCrCrTrGrArTrGrC)-30
50-d(TsCsCsAsTsGsAsCsGsTsTsCsCsTsGsAsTsGsC)-30
50-d(TsCsCsArTrGrArCrGrTrTrCrCrTrGrAsTsGsC)-30
[Sp]
[Sp-Rp-Sp]-chimer
aStereo-random (SR) represents a PS-oligo synthesized by b-cyanoethylphosphoramidite chemistry, containing 2n diastereomers (n=number of
phosphorothioate internucleoside linkages). Chimer represents stereo-enriched PS-oligo that contains a speci®c number of [Sp] or [Rp] phosphor-
othioate linkages as represented with R for [Rp] and S for [Sp] linkages. The CpG motif in oligos 11±14 is shown in italics. Underlined bases in SR
control sequence represent mismatched bases.
Figure 1. (A) UV thermal melting curves of the duplexes of PS-oligos 1 (~), 2 (^), 3 (*), and [Sp-Rp-Sp] chimer 4 (&) with the target RNA. CD
spectra of the duplexes of PS-oligos 1±3 with the target DNA (B) and RNA (C) sequences. In both panels B and C: ^ (1), * (2) and ~ (3).
target strand are in the order [Rp] > stereo-random >
[Sp]. These results suggest that the stereochemistry of
the PS-oligo backbone does not alter the global
conformation of the duplexes formed with RNA/DNA
target strands.4,20±22
the lanes corresponding to oligo 8, which contained [Sp]
linkages in that region of the sequence. The PS-oligos 1±4
showed similar stability as observed with oligos 5±7
against SVPD as studied by CGE (data not shown).
Figure 2B shows digestion of oligos 5±9 in the presence
of DNA polymerase I. DNA polymerase I exhibits
30-exonuclease activity under the appropriate experi-
mental conditions. The digestion pro®les in Figure 2B
suggest that DNA polymerase has diastereomeric selec-
tivity similar to that of SVPD. The extent of degrada-
tion, however, is weaker than in the case of SVPD.
Nuclease stability of stereo-enriched PS-oligos
We have studied the stability of PS-oligos 5±9 against
DNA polymerase I and SVPD and in human serum.
Figure 2A shows SVPD digestion of oligos 5±9. From
the digestion pro®les, it is clear that SVPD digests [Rp]
diastereomer selectively over [Sp] diastereomer.3 The
[Sp] PS-oligo 7 is more stable than [Rp] PS-oligo 6. The
stereo-random PS-oligo 5 has a stability intermediate to
the stability of [Sp] and [Rp] PS-oligos. The extent of
degradation of PS-oligos 5±7 suggest percentage of [Rp]
(degraded bands) and [Sp] (intact oligo band) popula-
tion in each oligo. The chimeric PS-oligo 9 ([Sp-Rp-Sp])
with [Sp] linkages at the 30- and 50-ends is as stable as
the [Sp] PS-oligo. The PS-oligo 8 with [Rp] linkages at
the ends is digested as in the case of [Rp] PS-oligo 6.
Note the less intense bands in the middle of the gel in
The two nucleases used in these studies showed
diastereomeric-dependent activity. To examine if human
serum nucleases exhibit any selectivity for [Rp] and [Sp]
diastereomers and study the stability of PS-oligos 5±9
against serum nucleases, we incubated these oligos in
human serum and examined the degradation pattern of
each PS-oligo on denaturing polyacrylamide gels. Fig-
ure 3 shows the stability of PS-oligos in human serum at
di€erent time intervals. In human serum also, [Sp] PS-
oligo 7 was more stable than the [Rp] PS-oligo 6, as was

D. Yu et al. / Bioorg. Med. Chem. 8 (2000) 275±284
Table 3. Tms of the duplexes of PS-oligos with the RNA and DNA target strands
279
Tm (ꢀC)a with
PS-oligob
RNA (ÁTm)c
DNA (ÁTm)
PKA oligos
1
2
3
4
SR
[Rp]
63.3 (Ð)
67.7 (+4.4)
59.0 ( 4.3)
66.4 (+3.1)
59.2 (Ð)
61.3 (+2.1)
59.3 (+0.1)
60.6 (+1.4)
[Sp]
[Sp-Rp-Sp]-chimer
MDM2 oligos
5
6
7
8
SR
[Rp]
[Sp]
57.8 (Ð)
58.8 (+1.0)
56.0 ( 1.8)
57.5 ( 0.3)
57.3 ( 0.5)
23.3 ( 34.5)
58.6 (Ð)
ndd
nd
nd
[Rp-Sp-Rp]-chimer
[Sp-Rp-Sp]-chimer
SR control
9
10
nd
26.5 ( 32.1)
Target RNAs
For PKA oligos
For MDM2 oligos
50-ACCGCCGCCAGUGAGGAGGCACGCAGCCUU-30e
50-CUGUGAGUGAGAACAGGUGUCACCU-30
aSee Experimental for bu€er conditions and oligo concentrations.
bSequences of PS-oligos are shown in Table 1.
cThe numbers shown in parentheses (ÁTm) are the di€erences in the Tm and relative to the Tm of the duplex of stereo-random PS-oligo with the target.
dnd, not determined.
eThe complementary sites for the corresponding antisense oligos in the target RNA sequences are shown in bold face.
Figure 2. Autoradiograms showing (A) SVPD and (B) DNA polymerase I digestion of PS-oligos 5±9 at di€erent time points.
observed earlier. Importantly, the [Sp-Rp-Sp] chimeric
PS-oligo 9 was as stable as [Sp] PS-oligo in human
serum.
chirality at the phosphorus center. To assess the rate of
RNase H hydrolysis of RNA in the presence of each PS-
oligo, the intact RNA band in each lane was determined
and plotted as percent RNA hydrolyzed at the 1- and
5-min time points (Fig. 4B). The eciency of RNase
H cleavage of RNA in the presence of PS-oligos is in
the order: [Rp] > stereo-random > [Sp]. These results
con®rm earlier observations that [Rp] diastereomer
activates RNase H better than [Sp] diastereomer.
RNase H activity of stereo-enriched PS-oligos
Figure 4A shows RNase H hydrolysis of RNA in the
presence of PS-oligos 5±10. RNA is cleaved in the pre-
sence of all PS-oligos, except stereo-random mis-
matched PS-oligo 10, which has four mismatches. The
results in Figure 4A show that the rate and pattern of
RNase H cleavage is di€erent in the presence of each
PS-oligo. In general, no intact RNA was left at the 15-
min time point (except in the case of chimeric PS-oligo
9). The presence of stereo-random, [Rp], and [Sp] PS-
oligos yielded similar RNA cleavage patterns. The rate
of RNase H cleavage of RNA is dependent on the
Comparison of the RNA cleavage pattern in the pre-
sence of chimeric PS-oligos 8 and 9 suggest that clea-
vage is centered in the middle of the binding site with
oligo 9 where the [Rp] segment is located. In the case of
oligo 8, the cleavage bands in the middle of the binding
site are lighter, where the [Sp] segment is located, and
the bands at the bottom of the gel are darker, where the

280
D. Yu et al. / Bioorg. Med. Chem. 8 (2000) 275±284
Figure 3. Autoradiograms showing digestion of PS-oligos 5±9 at di€erent time points in human serum.
Figure 4. (A) An autoradiogram showing RNase H digestion pro®les of the complementary RNA in the presence of PS-oligos 5±9 as labeled on top
of the gel at di€erent time points. The 0 time point lanes represent control RNA-PS-oligo duplex in the absence of RNase H. (B) A plot showing %
of the target RNA hydrolyzed in the presence of each PS-oligo at 1 min (white bars) and 5 min (black bars) time points.
30-end [Rp] segment is present. Figure 4B suggests that
RNA cleavage is faster in the presence of chimeric oligo
8 than in the presence of oligo 9. Our earlier studies with
MBOs containing RNase H-activating segments either
in the center (end-modi®ed) or at the 30- and 50-ends
with a central non-RNase H activating segment (cen-
trally-modi®ed) showed that RNase H cleavage of RNA
is rapid in the presence of centrally modi®ed MBOs.23
In agreement with those results, the chimeric PS-oligo 8
containing more ecient RNase H activating [Rp]
segments on either end ([Rp-Sp-Rp]) showed higher
RNA hydrolysis.
Serum protein and enzyme binding of stereo-enriched
PS-oligos
The 50-end-labeled PS-oligos 5±9 were incubated with
DNA polymerase I, recombinant HIV-reverse tran-
scriptase, and human serum albumin (HSA), g-globu-
lins, and ®brinogen, and examined on non-denaturing

D. Yu et al. / Bioorg. Med. Chem. 8 (2000) 275±284
281
polyacrylamide gels for diastereomeric-dependent pro-
tein binding of stereo-enriched PS-oligos. All ®ve oligos
showed binding to proteins examined and the extent
and strength of binding of these oligos was dependent
on the protein used (data not shown) as reported
for other proteins.24,25 We did not observe signi®cant
di€erences in protein binding (for a given protein) of
oligos 5±9 (data not shown). These results suggest
that the protein binding of PS-oligos is not dependent
on the nature of the diastereomeric isomer as has
been observed with other stereo-regular PS-oligo
sequences.26
Inhibition of MDM2 by stereo-enriched PS-oligos in cell
cultures
PS-oligo 5 is complementary to MDM2 mRNA and has
been studied in detail previously.19 The earlier studies
showed that PS-oligo 5 e€ectively down regulates
MDM2 mRNA and its protein, which in turn stabilizes
p53 and induces p21. The PS-oligo 5 has been used in
several di€erent cell lines. To determine p53 transcrip-
tional activity, a p53-responsive luciferase reporter BP
100-luc containing the p53 binding site from the MDM2
gene was transfected into JAR cells with a neomycin-
resistant marker plasmid as used previously.19 The
JAR-BP 100-luc cells were treated with di€erent con-
centrations of PS-oligos and the luciferase activities
were determined.
Cell proliferation of stereo-enriched PS-oligos
PS-oligos containing CpG motif are known to induce
cytokines, which has been associated with non-sequence-
speci®c biological activity and side-e€ects.27±30 To
examine the role of [Rp] and [Sp] diastereomers of PS-
oligos on cell proliferation, we synthesized PS-oligos
(11±14) that contained a CpG motif (Table 2) and stu-
died for their cell proliferative activity in cell cultures.
The data are presented in Figure 5. The stereo-random
and [Sp] PS-oligos showed higher proliferative activity
than [Rp] PS-oligo. The PS-oligo 12 with [Rp] linkages
did not induce signi®cant cell proliferation under the
experimental conditions. In order to examine whether
the inability of [Rp] PS-oligo (12) to induce prolifera-
tion resulted from its lower stability against nucleases,
we synthesized a chimeric PS-oligo (14, Table 2) with
[Sp] linkages at both the ends and [Rp] linkages in the
middle of the sequence. The linkage between C and G of
the CpG motif was of the [Rp] type (see Table 2). This
kind of chimeric design showed stability similar to that
of [Sp] PS-oligo in nuclease stability studies. The data in
Figure 4 suggest that this chimeric PS-oligo has cell pro-
liferative activity comparable to that of [Sp] PS-oligo.
These results suggest that the higher cell proliferative
property of [Sp] diastereomer of [PS] oligos is probably
the result of its higher stability against nucleases.
The [Sp] and [Rp] PS-oligos (6 and 7), [Sp-Rp-Sp] and
[Rp-Sp-Rp] chimeric PS-oligos (8 and 9), and stereo-
random PS-oligo (5) exhibited similar levels of down-
regulation of MDM2, thereby activating p53 as detected
by luciferase activity (Fig. 6). The control stereo-random
PS-oligo (10) with four mismatches induced minimal
luciferase activity (Fig. 6). These results suggest that
under the experimental conditions used in cell cultures
both [Rp] and [Sp] PS-oligos have activity similar to
that of stereo-random PS-oligo.
There are contradicting reports in the literature about
the biological activity of stereo-regular PS-oligos. A 15-
mer-all [Rp] PS-oligo, which contained four contiguous
Gs, has been reported to be consistently more active in
inhibiting proliferation in cell culture experiments than
were stereo-random and all [Sp] PS-oligos of the same
sequence.31 It was concluded that the higher ecacy
observed with all-[Rp] PS-oligo was the result of
increased hybridization eciency to the target RNA
Figure 5. E€ect of p-chirality on cell proliferative activity of PS-oligos
(11±14) containing CpG motif. Each value is an average of two inde-
pendent measurements. The white and dark shaded bars represent the
concentrations of PS-oligos at 1.0 and 10.0 mM/mL, respectively.
Figure 6. Inhibitory activity of MDM2 PS-oligos 5±10 in transfected
JAR cells. Each value is an average of at least two independent
experiments.

282
D. Yu et al. / Bioorg. Med. Chem. 8 (2000) 275±284
and greater ability to activate RNase H.31 Higher e-
cacy of this [Rp] PS-oligo could also result from the
increased stability of the oligonucleotide in serum as a
result of higher-order structure formation by oligonu-
cleotides that contain G-rich sequences.32 A 16-mer all-
[Sp] PS-oligo complementary to human PAI-1 mRNA
has been shown to be more potent in cultures than
stereo-random and all-[Rp] counterparts as a result of its
higher stability against cellular and serum nucleases.33,34
This oligonucleotide also contains a GAGGG sequence,
however, which can potentially form a higher-order
structure, although it was claimed not to form. The
truncated and control oligonucleotides that were devoid
of the GAGGG sequence were signi®cantly less active,
suggesting that this speci®c sequence motif is required
for the observed biological activity.33,34
against nucleases. From these initial studies of chimeric
PS-oligos, we propose that it might be possible to mod-
ulate pharmacokinetic and tissue disposition pro®les of
PS-oligos using chimeric designs, while retaining bioac-
tivity similar to that of mixed-backbone oligonucleo-
tides.23 Further studies are ongoing to understand the
impact of stereo-enriched PS-oligos on pharmacokinetic
and stability, safety pro®les, and in vivo activity.
Experimental
Preparation of bicyclic oxazaphospholidine nucleoside
synthons
Synthesis of monomer d: to a cold solution of 50-O-
DMTr-20-deoxy-N-benzoyl-cytidine (2.5 g; 3.96 mmol)
in dry methylene chloride (ꢁ40 mL) and diisopropyl-
ethylamine (1 mL) was slowly added p-chloro-oxaza-
phospholidine (0.6 g; 3.96 mmol, derived from (S)-(+)-
2-pyrrolidine methanol15) and stirred under argon at
78 ꢀC for 30 min. The reaction mixture was allowed to
come to ambient temperature and stirred overnight. The
reaction was quenched by adding 100 mL of 5% ice-
cold sodium bicarbonate solution. The product was
extracted with methylene chloride (3Â50 mL) and the
combined organic extracts were dried over anhydrous
sodium sulfate. The organic layer was evaporated and
the solid d obtained was dried overnight in vacuo (yield
88%). The other monomers a±c were prepared similarly.
For monomers e±h, phosphitylation agent was prepared
from (R)-( )-2-pyrrolidine methanol.15 The yields were
90‹5%. The absolute con®gurations of a±h have not
been assigned yet. The 31P NMR spectra of monomers
a±h were recorded on a Varian 600 MHz NMR spec-
trometer in CDCl3 with 85% phosphoric acid as an
external 31P standard. 31P NMR data (d): a, 151.5; b,
153.9; c, 152.8; d, 152.6; e, 152.8; f, 152.4; g, 151.3; and
h, 153.2.
In spite of di€erences in the biophysical and biochem-
ical properties of the two stereo-enriched PS-oligos, the
oligonucleotides targeted to MDM2 mRNA did not
show signi®cant stereo-dependent di€erences in activity
in cell cultures. The oligonucleotide sequence used in the
present study does not contain motifs that promote
formation of higher-order or self-complementary struc-
tures. This leaves the question, how would a less nuclease-
resistant [Rp] diastereomer survive in cell culture
medium and intracellularly, and exert comparable
biological activity as that of [Sp] and stereo-random
PS-oligos? In general, cationic lipids are used to
facilitate transfection of oligonucleotides into cells in
cell culture experiments. In the present study, lipofectin
was used as a carrier for oligonucleotide delivery and
the cells were incubated with oligos for 24 h. Negatively
charged oligonucleotides form complexes with cationic
molecules. The association of negatively charged oligo-
nucleotides with lipids not only increases cellular uptake
but also protects oligonucleotides from nuclease diges-
tion.35 Perhaps the lipids used in the cell culture experi-
ments contributed to the longer half-life of [Rp] PS-
oligo, which in turn could result in activity comparable
to that of [Sp] and stereo-random PS-oligos.
Stereo-selectivity of the coupling of nucleoside mono-
mers a±h to a thymidine nucleoside attached to a CPG
support was determined by synthesizing and analyzing
the dimers on HPLC as reported earlier.15
Conclusions
We have shown that it is possible to synthesize stereo-
enriched PS-oligos of varying lengths and base compo-
sitions using 50-DMTr-protected oxazaphospholidine
monomers. The present study showed that [Rp] PS-
oligos have a higher binding anity to the target RNA
and greater RNase H activation, but lower stability
against nucleases, than the stereo-random and [Sp] PS-
oligos. The [Sp] PS-oligos have a lower binding anity
to the target RNA and lower RNase H activation, but
greater stability against nucleases, than the stereo-
random and [Rp] PS-oligos. In the present study we
designed chimeric PS-oligos [Sp-Rp-Sp] and [Rp-Sp-Rp]
in which advantageous properties of both the diaste-
reomers can be combined. The cell culture studies
showed that stereo-random, all-[Rp], all-[Sp], and chi-
meric PS-oligos have similar antisense activities. The
cell-proliferatory activity of PS-oligos could be the
result of the higher stability of the [Sp] diastereomer
Oligonucleotide synthesis and puri®cation
All oligodeoxynucleotides were synthesized on 1-mmol
scale on a PerSeptive Biosystem's Expedite 8909 auto-
mated DNA synthesizer. Both [Rp] and [Sp] PS-oligos
were synthesized on a CPG-solid support using appro-
priate 50-DMTr-protected oxazaphospholidine mono-
mers prepared as above. The stereo-random PS-oligo
was synthesized using b-cyanoethylphosphoramidite
chemistry producing a diastereomeric mixture. The chi-
meric PS-oligos containing [Sp] linkages at the 30- and
50-ends and [Rp] linkages in the middle (or vice versa)
were synthesized using appropriate oxazaphospholidine
monomers. The average coupling eciency, as mea-
sured automatically by the synthesizer by trityl color
determination at 498 nm at each detritylation step, was
about 98%. The protected PS-oligos on CPG-solid support
were treated with concentrated ammonium hydroxide at

D. Yu et al. / Bioorg. Med. Chem. 8 (2000) 275±284
283
65 ꢀC for 16 h. The stereo-random PS-oligo with a 50-
DMTr group was puri®ed by reverse phase HPLC on a
C18 column and was subsequently detritylated with 80%
acetic acid. Stereo-enriched PS-oligos were puri®ed by
preparative polyacrylamide gels. All the PS-oligos were
desalted by dialysis against double distilled water. The
concentrations were determined by UV absorbance
measurement at 260 nm using extinction coecients
calculated by the nearest neighbor method. The yields
of gel puri®ed stereo-enriched PS-oligonucleotides were
about 25% as further optimization of coupling condi-
tions required.
A small amount (30 pmol) of 50-32P-end-labeled PS-
oligo was incubated with 1.5 units of DNA polymerase
I (Worthington Biochemical) in 20 mL of 50 mM Tris,
pH 8.0 bu€er containing 5 mM MgCl2. An aliquot
(5 mL) was removed at various time points and the
reaction was stopped by adding 10 mL of formamide.
The samples were analyzed on polyacrylamide gels as
described above.
A small amount (30 pmol) of 50-32P-end-labeled PS-
oligo was incubated with 100 mL of human serum (Bio
Whittaker). An aliquot (25 mL) was removed at di€erent
time points and the sample was treated with 100 mg
proteinase K (Boehringer Mannheim), 40 mL DNA
extraction bu€er (0.5% SDS, 10 mM NaCl, 20 mM
Tris, pH 7.6, 10 mM EDTA) and 175 mL distilled water
for 2 h at 60 ꢀC. The oligos were extracted with 300 mL
phenol and 300 mL chloroform and ethanol precipitated.
The samples were dried, resuspended in 20 mL water and
analyzed on polyacrylamide gels as described above.
UV absorbance melting experiments
The PS-oligos were mixed with either RNA or DNA
target strands at a 1:1 molar ratio in a bu€er containing
100 mM NaCl, 0.1 mM Na2EDTA, and 10 mM sodium
phosphate, pH 7.4. The total oligonucleotide con-
centration was 2 mM in a ®nal volume of 1 mL. The
samples were heated for 3 min and cooled slowly to
room temperature prior to running the experiment.
Absorbance melting curves were recorded on a Perkin±
Elmer Lambda 20 UV±vis spectrometer equipped with
Peltier-e€ect 6-cell holder. The temperature of the cell
holder was controlled by an external PTP-6 Peltier Sys-
tem. The heating rate was 0.5 ꢀC/min. Data were col-
lected on a Dell OptiPlex GM5133 computer interfaced
with the spectrometer and processed using the software
supplied by Perkin±Elmer with the spectrometer. The
melting temperatures (Tms) of the duplexes were mea-
sured from ®rst derivative plots obtained from thermal
melting curves. Each experiment was repeated at least
twice and the average was taken for Tm determinations.
The reproducibility of the Tm values was within
‹0.5 ꢀC.
RNase H experiments
Puri®ed 25-mer RNA sequence (Table 2) com-
plementary to the MDM2 antisense oligos was labeled
with 32P at the 50-end using T4 polynucleotide kinase
(New England Biolabs) and g-32P-ATP (Amersham)
and used for RNase H assay as described earlier.36
Brie¯y, the RNase H assay was carried out by forming a
heteroduplex of labeled RNA (1 pmol) and the PS-oli-
gos (1.0 pmol) in 10 mL of bu€er containing 10 mM
KCl, 5 mM MgCl2, 0.1 mM DTT, 5% glycerol, 20 mM
Tris±HCl, pH 7.5 and 40 units of RNase hihibitor
(Boehringer Mannheim) was added and the total
volume was adjusted to 30 mL. An aliquot of 5 mL was
removed for 0 time point and 0.002 units of Escherichia
coli RNase H (Amersham) was added and incubated at
37 ꢀC. Aliquots (5 mL) were removed at di€erent time
points and the reaction was stopped by adding 5 mL
formamide and freezing at 70 ꢀC. The cleavage pro-
ducts were fractionated on polyacrylamide gels as
described above.
CD experiments
The same oligonucleotide sample solutions used for UV
thermal melting studies were also used for CD experi-
ments. The CD spectra were recorded on a Jasco J-710
spectropolarimeter with a 0.5 cm quartz cell attached to
a Peltier thermal controller. The samples were equili-
brated at 21 ꢀC for 15 min before recording the spectra.
Each spectrum was an average of 8 scans with the bu€er
blank subtracted, which was also an average of 8 scans
and obtained at the same scan speed (100 nm/min). All
the spectra were noise reduced using the software sup-
plied by Jasco, and the molar ellipticities were calcu-
lated using the same software.
In vitro cell proliferation assay
The cell proliferation assay was carried out as described
earlier.29 Brie¯y, spleen cell (4- to 5-week-old male CD1
mouse, 20±22 g; Charles River, Wilmington, MA) sus-
pensions were prepared and plated in 96-well dishes at a
density of 106 cells/mL in a ®nal volume of 100 mL. The
cells were incubated at 37 ꢀC after the addition of 10 mL
PS-oligo solution. After 44 h incubation, 1 mCi [3H]thy-
midine (Amersham) was added and the cells were pulse
labeled for another 4 h. The cells were harvested by an
automatic cell harvester and the ®lters were counted
using a scintillation counter. All experiments were car-
ried out in triplicate.
Nuclease digestion and PAGE
A small amount (85 pmol) of 50-32P-end-labeled PS-
oligo was incubated with 2 units of snake venom phos-
phodiesterase (Boehringer Mannheim) in 30 mL of 10
mM Tris, pH 8.0 bu€er containing 10 mM MgCl2. An
aliquot (7 mL) was removed at various time points and
the reaction was stopped by adding 10 mL of for-
mamide. The samples were analyzed on gels containing
20% acrylamide and 8 M urea. The extent of the nucle-
ase hydrolysis was visualized by autoradiography.
Treatment of JAR-BP 100-luc cells by antisense
oligonucleotides
The PS-oligos complementary to MDM2 mRNA were
studied for their inhibitory activity in transfected JAR

284
D. Yu et al. / Bioorg. Med. Chem. 8 (2000) 275±284
cells as reported earlier.19 The JAR cells were obtained
from American Type Culture Collection. The JAR cells
stably transfected with the p53-inducible BP 100-luci-
ferase plasmid (JAR-BP100-luc) were treated with anti-
sense oligonucleotides at di€erent concentrations for
24 h. Luciferase activities in the treated cells were
determined as described earlier.19 The cells were cul-
tured in DMEM with 10% fetal bovine serum (FBS).
Before addition of antisense oligonucleotides, cells were
supplemented with DMEM containing 1% FBS. Lipo-
fectin (GIBCO/BRL) was incubated with serum-free
DMEM medium at room temperature for 45 min and
then mixed with antisense oligonucleotides for 10 min
and added to the culture. The ®nal concentration of
lipofectin was 7 mg/mL and the ®nal concentration of
FBS was 0.75%. Controls were treated with lipofectin
alone. Luciferase activity was determined using a luci-
ferase assay kit (Tropix, Bedford, MA) and shown as
luciferase activity/unit protein.
13. Iyer, R. P.; Yu, D.; Ho, N.; Tan, W.; Agrawal, S. Tetra-
hedron: Asymmetry 1995, 6, 105.
14. Wang, J.-C.; Just, G. Tetrahedron Lett. 1997, 38, 3797.
15. Iyer, R. P.; Guo, M.-J.; Yu, D.; Agrawal, S. Tetrahedron
Lett. 1998, 39, 2491.
16. Guo, M. J.; Yu, D.; Iyer, R. P.; Agrawal, S. Bioorg. Med.
Chem. Lett. 1998, 8, 2539.
17. Yokuzaki, H.; Budillon, A.; Tortora, G.; Meissner, S.;
Beaucage, S. L.; Miki, K.; Cho-Chung, Y. S. Cancer Res.
1993, 53, 868.
18. Nesterova, M.; Cho-Chung, Y. S. Nature Med. 1995, 1,
528.
19. Chen, L.; Agrawal, S.; Zhou, W.; Zhang, R.; Chen, J.
Proc. Natl. Acad. Sci. USA 1998, 95, 195.
20. Bachelin, M.; Hessler, G.; Kurz, G.; Hacia, J. G.; Dervan,
P. B.; Kessler, H. Nature Struct. Biol. 1998, 5, 271.
21. Furrer, P.; Billeci, T. M.; Donati, A.; Kojima, C.; Kar-
wowski, B.; Sierzchala, A.; Stec, W.; James, T. L. J. Mol. Biol.
1999, 285, 1609.
22. Clark, C. L.; Cecil, P. K.; Singh, D.; Gray, D. M. Nucleic
Acids Res. 1997, 25, 4098.
23. Agrawal, S.; Jiang, Z.; Zhao, Q.; Shaw, D. R.; Cai, Q.;
Roskey, A.; Channavajjala, L.; Saxinger, C.; Zhang, R. Proc.
Natl. Acad. Sci. USA 1997, 94, 2620.
Acknowledgements
24. Gao, W.-Y.; Han, F.-S.; Storm, C.; Egan, W.; Cheng, Y.
C. Mol. Pharmacol. 1992, 41, 223.
25. Agrawal, S.; Zhang, X.; Cai, Q.; Kandimalla, E. R.; Man-
ning, A.; Jiang, Z.; Marcel, T.; Zhang, R. J. Drug Target.
1998, 5, 303.
The authors thank Dr. MaoJun Guo for his help in the
synthesis of some of the bicyclic oxazaphospholidine
nucleoside monomers.
26. Benimetskaya, L.; Tonkinson, J. L.; Koziolkiewicz, M.;
Karwowski, B.; Guga, R.; Zeltser, R.; Stec, W.; Stein, C. A.
Nucleic Acids Res. 1995, 23, 4239.
References
1. Akhtar, S.; Agrawal, S. Trends Pharm. Sci. 1997, 18, 12.
3. Eckstein, F. Annu. Rev. Biochem. 1985, 54, 367.
2. Kronenwett, R.; Haas, R. Ann. Hematol. 1998, 77, 1.
4. Tang, J.-Y.; Roskey, A.; Li, Y.; Agrawal, S. Nucleosides
Nucleotides 1995, 14, 985.
5. Koziolkiewicz, M.; Krakowiak, A.; Kwinkowski, M.;
Boczkowska, M.; Stec, W. J. Nucleic Acids Res. 1995, 23,
5000.
27. Pisetsky, D. S.; Reich, C. F. Life Sci. 1993, 54, 101.
28. Krieg, A. M.; Yi, A. K.; Matson, S.; Waldschmidt, T. J.;
Bishop, G. A.; Teasdale, R.; Koretzky, G. A.; Klinman, D. M.
Nature 1995, 374, 546.
29. Zhao, Q.; Temsamani, J.; Iadarola, P. L.; Jiang, Z.; Agra-
wal, S. Biochem. Pharmacol. 1996, 51, 173.
30. Agrawal, S.; Zhao, Q. Curr. Opin. Chem. Biol. 1998, 2, 519.
31. Fearon, K. L.; Hirschbein, B. L.; Chiu, C.-Y.; Quijano, M.
R.; Zon, G. Oligonucleotides as Therapeutic Agents, Ciba
Foundation Symposium 209; John Wiley & Sons, Chichester,
1997; pp. 19±37.
6. Koziolkiewicz, M.; Wojcik, M.; Kobylanska, A.; Kar-
wowski, B.; Rebowska, B.; Guga, P.; Stec, W. J. Antisense
Nucleic Acid Drug Dev. 1997, 7, 43.
7. Hacia, J. G.; Wold, B. J.; Dervan, P. B. Biochemistry 1994,
33, 5367.
32. Agrawal, S.; Tan, W.; Cai, Q.; Xie, X.; Zhang, R. Anti-
sense Nucleic Acid Drug Dev. 1997, 7, 245.
8. Hartmann, B.; Bertrand, H.-O.; Fermandjian, S. Nucleic
Acids Res. 1999, 27, 3342.
9. Tamura, Y.; Miyoshi, H.; Yokota, T.; Makino, K.; Mur-
akami, A. Nucleosides Nucleotides 1998, 17, 269.
10. Stec, W. J.; Wilk, A. Angew. Chem., Int. Ed. Engl. 1994,
33, 709.
11. Stec, W. J.; Grajkowski, A.; Kobylanska, A.; Darwowski,
B.; Koziolkiewicz, M.; Misiura, K.; Okruszek, A.; Wilk, A.;
Guga, P.; Boczkowska, M. J. Am. Chem. Soc. 1995, 117,
12020.
33. Buczko, W.; Cierniewski, C.; Kobylanska, A.; Koziolk-
iewicz, M.; Okruszek, A.; Pawlowska, Z.; Pluskota, E.; Stec,
W. J. Pharmacol. Ther. 1997, 76, 161.
34. Stec, W. J.; Cierniewski, C. S.; Okruszek, A.; Kobylanska,
A.; Pawlowska, Z.; Koziolkiewicz, M.; Pluskota, E.; Macias-
zek, A.; Rebowska, B.; Stasiak, M. Antisense Nucleic Acid
Drug Dev. 1997, 7, 567.
35. Capaccioli, S.; DiPasquale, G.; Mini, E.; Mazzei, T.;
Quattrone, A. Biochem. Biophys. Res. Commun. 1993, 197,
818.
12. Iyer, R. P.; Yu, D.; DevIin, T.; Ho, N.; Agrawal, S. J. Org.
Chem. 1995, 60, 5388.
36. Shen, L. X.; Kandimalla, E. R.; Agrawal, S. Bioorg. Med.
Chem. 1998, 6, 1695.