Solid-phase synthesis, thermal denaturation studies, nuclease resistance, and cellular uptake of (oligodeoxyribonucleoside)methylborane phosphine-DNA chimeras

Article abstract of DOI:10.1021/ja201314q

The major hurdle associated with utilizing oligodeoxyribonucleotides for therapeutic purposes is their poor delivery into cells coupled with high nuclease susceptibility. In an attempt to combine the nonionic nature and high nuclease stability of the P-C bond of methylphosphonates with the high membrane permeability, low toxicity, and improved gene silencing ability of borane phosphonates, we have focused our research on the relatively unexplored methylborane phosphine (Me-P-BH₃) modification. This Article describes the automated solid-phase synthesis of mixed-backbone oligodeoxynucleotides (ODNs) consisting of methylborane phosphine and phosphate or thiophosphate linkages (16-mers). Nuclease stability assays show that methylborane phosphine ODNs are highly resistant to 5′ and 3′ exonucleases. When hybridized to a complementary strand, the ODN:RNA duplex was more stable than its corresponding ODN:DNA duplex. The binding affinity of ODN:RNA duplex increased at lower salt concentration and approached that of a native DNA:RNA duplex under conditions close to physiological saline, indicating that the Me-P-BH₃ linkage is positively charged. Cellular uptake measurements indicate that these ODNs are efficiently taken up by cells even when the strand is 13% modified. Treatment of HeLa cells and WM-239A cells with fluorescently labeled ODNs shows significant cytoplasmic fluorescence when viewed under a microscope. Our results suggest that methylborane phosphine ODNs may prove very valuable as potential candidates in antisense research and RNAi.

Full text of DOI:10.1021/ja201314q

ARTICLE
pubs.acs.org/JACS
Solid-Phase Synthesis, Thermal Denaturation Studies,
Nuclease Resistance, and Cellular Uptake of
(Oligodeoxyribonucleoside)methylborane PhosphineꢀDNA
Chimeras
Heera Krishna and Marvin H. Caruthers*
Department of Chemistry and Biochemistry, University of Colorado at Boulder, Boulder, Colorado 80309, United States
S Supporting Information
b
ABSTRACT: The major hurdle associated with utilizing oligo-
deoxyribonucleotides for therapeutic purposes is their poor
delivery into cells coupled with high nuclease susceptibility. In
an attempt to combine the nonionic nature and high nuclease
stability of the PꢀC bond of methylphosphonates with the high
membrane permeability, low toxicity, and improved gene
silencing ability of borane phosphonates, we have focused our
research on the relatively unexplored methylborane phosphine
(MeꢀPꢀBH3) modification. This Article describes the auto-
mated solid-phase synthesis of mixed-backbone oligodeoxynu-
cleotides (ODNs) consisting of methylborane phosphine and phosphate or thiophosphate linkages (16-mers). Nuclease stability
assays show that methylborane phosphine ODNs are highly resistant to 50 and 30 exonucleases. When hybridized to a
complementary strand, the ODN:RNA duplex was more stable than its corresponding ODN:DNA duplex. The binding affinity
of ODN:RNA duplex increased at lower salt concentration and approached that of a native DNA:RNA duplex under conditions
close to physiological saline, indicating that the MeꢀPꢀBH3 linkage is positively charged. Cellular uptake measurements indicate
that these ODNs are efficiently taken up by cells even when the strand is 13% modified. Treatment of HeLa cells and WM-239A cells
with fluorescently labeled ODNs shows significant cytoplasmic fluorescence when viewed under a microscope. Our results suggest
that methylborane phosphine ODNs may prove very valuable as potential candidates in antisense research and RNAi.
’ INTRODUCTION
stability of the PꢀC bond with the high water solubility, low
toxicity, and improved gene silencing ability of borane phospho-
nates to generate the methylborane phosphine internucleotide
linkage.
The ideal oligodeoxynucleotide drug molecule would be one
that possesses enhanced intracellular delivery, high target-bind-
ing affinity, and biological activity, while maintaining low nucle-
ase susceptibility, nonspecific binding, and toxicity. The most
prevalent approach toward this goal has been the construction of
chimeric oligodeoxynucleotides (ODNs) containing segments of
different chemistries, each of which confers a certain advantage
to the oligomer.1,2 Taking a step further, we have now optimized
the chemistry behind the automated synthesis of a relatively
unexplored3 “hybrid” modification, which has overlapping
structural motifs of two well-known antisense molecules, the
methylphosphonate4 and borane phosphonate.5 Methylpho-
sphonate oligodeoxynucleotides contain nonionic internucleo-
tide linkages that are highly lipophilic and resist enzymatic
degradation. However, due to their lipophilicity, they have low
water solubility and are often bound to the cell membrane and
not available for biological activity in cells.6 Borane phosphonate
oligodeoxynucleotides, on the other hand, have high water solu-
bility, possess favorable cell uptake properties, and show siRNA
activity even as single strands.7 In this Article, we report our
research on combining the nonionic nature and high nuclease
In terms of size, the substitution of a nonbridging oxygen of
the phosphate with a methyl group to generate a methylphos-
phonate represents the smallest possible structural change that
eliminates the negative charge on the backbone. A borane
phosphonate modification is isosteric with a methylphosphonate
and also isoelectronic with the native phosphate. It also imparts
higher lipophilicity to the molecule as compared to a regular
phosphate while also retaining its negative charge. In the
methylborane phosphine linkage described here, the nonbridg-
ing oxygens of a phosphodiester are replaced by a borane and a
methyl group, giving rise to a P(IV) center. In the absence of
nonbridging oxygens, the electronegativities of the substituents
follow the order: C (2.5) > P (2.2) > B (2.0). As a result, the
biophysical properties of the resultant methylborane phosphine
(MeꢀPꢀBH3) linkage are expected to be different from those of
Received: February 11, 2011
Published: May 17, 2011
r
2011 American Chemical Society
9844
dx.doi.org/10.1021/ja201314q J. Am. Chem. Soc. 2011, 133, 9844–9854
|

Journal of the American Chemical Society
ARTICLE
Scheme 1. Synthesis of 50-O-Silyl-N-trimethoxytrityl-20-deoxynucleoside-30-O-Methylphosphinoamidites (A, G, C) and
50-O-Silyl-20-deoxythymidine-30-O-methylphosphinoamiditea
a Reagents and conditions: (i) 4 equiv of Me3SiCl, pyridine, 2 h; (ii) 1.2 equiv of 4,40,400-trimethoxytritylchloride, pyridine, 16 h; (iii) 1.1 equiv of
[benzhydryloxy-bis(trimethylsilyloxy)silyl]chloride, 3 equiv of imidazole, DMF, 6 h; (iv) 1.2 equiv of MeP[N(i-C3H7)2]2, CH2Cl2, 4,5-dicyano
imidazole, 40 min.
the methylphosphonate and the borane phosphonate analogues.
We hypothesize that the resultant new class of compounds might
embody the best characteristics of both precursor molecules, the
high nuclease stability of the PꢀC bond of methylphosphonates
combined with the high aqueous solubility, membrane perme-
ability, low toxicity, and improved gene silencing ability of borane
phosphonates.
literature report,13 decomposition of the PꢀBH3 linkage was
negligible when 80% aqueous acetic acid was used for detrityla-
tion. Drawing upon these studies, we have now developed an
efficient synthetic route to methylborane phosphine ODNs. In
this Article, we describe the automated solid-phase synthesis,
hybridization experiments (Tm), nuclease susceptibilities, and
preliminary in vivo studies on a series of mixed-sequence
methylborane phosphine 20-deoxyoligonucleotide analogues.
Some years ago, our group developed a phosphoramidite-
based method for the high yielding synthesis of mixed-sequence
borane phosphonate DNA.8 These and further research efforts
have led us to replace the conventional 50-O-dimethoxytrityl
protecting group of the phosphoramidite synthons with a fluoride
labile 50-O-silyl ether (50-O-[benzhydryloxy-bis(trimethylsilyloxy)-
silyl], Bzh). The exocyclic amines of adenine, guanine, and
cytosine are protected with the mild acid-labile trimethoxytrityl
group. Because phosphoramidite chemistry is being used, thy-
mine is not protected.9 This new base protection strategy had to
be employed because the borane reagents reduce the commonly
used amide protecting groups to N-alkyl or aryl exocyclic amines
or form stable, irreversible borane adducts with these bases.10
Apart from conferring complete orthogonality to the fluoride
labile 50-O-silyl chemistry, the trimethoxytrityl protecting group
cannot be reduced by borane and, perhaps because of its large
size, prevents the formation of irreversible borane adducts. It
should also be noted that the trimethoxytrityl cation generated
during the conventional detritylation step (3% trichloroacetic
acid in CH2Cl2) of DNA synthesis is incompatible with the
borane phosphonate linkage.11,12 We observed that the methyl-
borane phosphine linkage was also prone to decomposition
under similar conditions. However, in agreement with a previous
’ RESULTS AND DISCUSSION
Chemical Synthesis. Our primary objective was to develop
the methylphosphinoamidite synthons for all four bases that are
needed to generate the methylborane phosphine linkages
(Scheme 1). The phosphitylating reagent, bis(N,N-diisopropyl-
amino)methylphosphine, was synthesized in high yield (90%)
via the Grignard reaction, starting from bis(N,N-diisopropyl-
amino)chlorophosphine and methylmagnesium bromide.14 This
diamidite was allowed to react with appropriately N-protected 50-
O-silyl-20-deoxynucleosides to yield the 50-O-silyl-20-deoxynu-
cleoside-30-O-methylphosphinoamidites (amidites 13, 14, and
15). The thymidine synthon (16) does not require base protec-
tion. A detailed procedure for the synthesis of amidites 13ꢀ16
can be found in the Experimental Section. To synthesize
chimeric ODNs having the methylborane phosphine and phos-
phate (or thiophosphate) linkages, N-trimethoxytrityl-50-O-silyl-
20-deoxynucleoside-30-O-methylphosphoramidite synthons (17ꢀ19)
and 50-O-silyl-20-deoxythymidine-30-O-methylphosphoramidite (20)
were prepared from compounds (8ꢀ11) and bis(N,N-diisopro-
pylamino)methoxyphosphine [Supporting Information, S8ꢀS9].
9845
dx.doi.org/10.1021/ja201314q |J. Am. Chem. Soc. 2011, 133, 9844–9854

Journal of the American Chemical Society
ARTICLE
Scheme 2. Solid-Phase Synthesis of Methylborane PhosphineꢀDNA Chimeras
All amidite synthons (13ꢀ20) have been characterized by NMR
and mass spectral analysis. The compounds were obtained in high
purity after column chromatography with average yields ranging
from 40% to 60%.
Solid-Phase Synthesis. Once these synthons were available,
our next goal was to optimize the solid-phase synthesis cycle for
preparing methylborane phosphine-modified ODNs. A sche-
matic of this cycle is depicted in Scheme 2. Prior to synthesis,
the 50-DMT group on the 20-deoxythymidine bound to the
polystyrene support15 was removed with 3% trichloroacetic acid
in dichloromethane. The unprotected 20-deoxynucleoside was
allowed to react with the appropriate 50-O-silyl-20-deoxynucleo-
side-30-O-methylphosphinoamidite in anhydrous acetonitrile in
the presence of 5-ethylthio-1H-tetrazole to generate a family of
dimers having a methylphosphite diester internucleotide linkage.
The coupling wait time was 500 s for the methylphosphinoami-
dites 13ꢀ16; it was decreased to 180 s for the phosphoramidites
17ꢀ20. This condensation step was followed by capping with
9846
dx.doi.org/10.1021/ja201314q |J. Am. Chem. Soc. 2011, 133, 9844–9854

Journal of the American Chemical Society
ARTICLE
Table 1. Synthesis Cycle for Methylborane Phosphine ODNs
reaction
coupling
wash/reagents/solvents
time (s)
0.1 M phosphinoamidites (13ꢀ16) and 0.25 M ethylthio tetrazole in acetonitrile (1:1)
wait: 500 s
or
0.1 M phosphoramidites (17ꢀ20) and 0.25 M ethylthio tetrazole in acetonitrile (1:1)
acetonitrile
wait: 180 s
40 s
wash
capping
0.1 M Unicap Phosphoramidite and 0.25 M ethylthio tetrazole in acetonitrile (1:1) at port 5a
wait: 40 s
boronation
25 mM BH3 THF in THF at port 12
12 to column, 60 s; wait, 120 s
3
or
oxidation/sulfurization
wash
1 M anhydrous tert-butyl hydroperoxide or Beaucage reagent in acetonitrile at port 15
15 to column, 15 s; wait, 45 s
acetonitrile,
40 s
dichloromethane
45 s
dimethylformamide
45 s
50-deprotection
wash
1.1 M HF/1.1 triethylamine/0.2 M NMDEAb in dimethylformamide (pH 9.2)
10 to column, 25 s; wait, 45 s
dimethylformamide,
45 s
55 s
acetonitrile
a The amidite capping was performed as per the manufacturer’s instructions.23 b NMDEA = N-methyl diethanolamine.
Unicap Phosphoramidite16 to block failure sequences. The
methylphosphite diester was boronated using a freshly prepared
Table 2. MALDI-TOF Analysis of Mixed-Backbone ODNs
mol. wtb
solution of 25 mM BH3 THF in THF to generate a methylbor-
3
ane phosphine (PꢀIV) linkage. Boronation wait time of 120 s
was essential to prevent incomplete boronation, which leads to
the generation of methylphosphonate linkages during a subse-
quent oxidation step. After boronation, the support was washed
successively with acetonitrile (40 s), dichloromethane (45 s), and
dimethylformamide (45 s) to complete one synthesis cycle. After
each condensation step followed by capping and boronation, 50-
O-silyl deprotection was carried out using a freshly prepared
solution of 1.1 M HF/1.1 M Et3N/0.2 M N-methyl diethanol-
amine in DMF (pH = 9.2) to generate a family of dinucleotides
having a combination of any of the four bases. These dimers
could then be extended using the same repetitive cycle to
generate multiple methylborane phosphine linkages. To intro-
duce a phosphate or a thiophosphate linkage, the growing
20-deoxyoligonucleotide chain was allowed to react with the appro-
priate N-trimethoxytrityl protected 50-O-silyl-20-deoxynucleo-
side-30-O-methylphosphoramidite using 5-ethylthio-1H-tetrazole
as activator, followed by capping with Unicap Phosphoramidite.
The resultant phosphite triester was treated with anhydrous tert-
butyl hydroperoxide17,18 or Beaucage reagent19 to generate the
phosphate or the thiophosphate triester linkage, respectively.
The 50-O-silyl deprotection using triethylammoniumꢀhydrogen
fluoride mixture as described above completed the cycle. In this
manner, either synthesis cycle was repeated until the ODN of
appropriate length, sequence, and internucleotide linkage was
generated. The protecting groups were removed sequentially as
reported for the synthesis of borane phosphonate DNA8 (see
Experimental Section). After deprotection, the ODN was cleaved
from the polystyrene support using ethylenediamine in methanolꢀ
waterꢀacetonitrile.20 Table 1 summarizes the synthetic cycle.
Numerous mixed backbone ODNs (16-mers) bearing methyl-
borane phosphine and phosphate or thiophosphate internucleo-
tide linkages have been generated using this synthesis strategy.
For all oligomers, similar profiles were obtained during RP-
HPLC purification, with each having one major, relatively broad
product peak and capped failure sequences, which appeared as
earlier eluting products. Broad product peaks were anticipated
due to the presence of multiple chiral phosphorus centers. The
no.
ODNa
calcd
obs
1
2
3
4
5
6
7
TbApApCpApCpGpApTpApCpGpCpGpAbT
TbAbApCpApCpGpApTpApCpGpCpGbAbT
TbAbAbCpApCpGpApTpApCpGpCbGbAbT
TbASASCSASCSGSASTSASCSGSCSGSAbT
TbAbASCSACSGSASTSASCSGSCSGSAbT
TbAbASCSASCSGSASTSASCSGSCSGbAbT
TbASASCSAbCSGbASTSASCSGSCSGSAbT
4862.01
4856.10
4848.22
5071.69
5051.77
5031.85
5031.85
4863.11
4853.44
4845.92
5068.33
5052.62
5030.92
5028.73
a p, phosphate; b, methylborane phosphine; s, thiophosphate. b Ob-
served masses represent MALDI-TOF data acquired on a Voyager DE-
STR instrument.
average yields (95% purity) for a 0.2 μM synthesis ranged from
50 to 80 OD.21 Aliquots of these ODNs were characterized by
analytical RP-HPLC. Figures 8ꢀ11 in the Supporting Informa-
tion show analytical RP-HPLC profiles for the ODNs 1, 4, 6, and
7, respectively. The mass data for ODNs 1ꢀ7 having combina-
tions of methylborane phosphine modified segments and inter-
nucleotide phosphate or thiophosphate linkages are listed in
Table 2. The observed masses for all ODNs correspond to those
as calculated. The 31P NMR analyses of the oligomers display a
broad signal at 146ꢀ150 ppm (methylborane phosphine) and a
sharp peak at ꢀ1.2 ppm when phosphate is part of the backbone
(Figure 1 and Supporting Information, S9, Figure 6), or at
57 ppm for a thiophosphate (Supporting Information, S10,
Figure 7). Very similar chemical shifts have been reported for
small molecules containing MeꢀPꢀBH3 linkages.22
Polyacrylamide-gel electrophoresis (14% gel, 8 M urea) of
chimeric methylborane phosphineꢀphosphate ODNs was car-
ried out to determine their electrophoretic mobility relative to an
unmodified phosphate control (oligo 8) and the tracking dye
bromophenol blue. The electrophoretic mobility followed the
order: oligo 8 > ODN 2 > bromophenol blue ≈ ODN 3. When
the ODN was 53% modified (ODN 10), the mobility decreased
further (Supporting Information, Table 1, S18). Thus, the electro-
phoretic mobility of the methylborane phosphine modified ODNs
9847
dx.doi.org/10.1021/ja201314q |J. Am. Chem. Soc. 2011, 133, 9844–9854

Journal of the American Chemical Society
ARTICLE
Figure 2. Nuclease resistance of ODNs against SVPDE. Half-lives are
defined as time taken for 50% degradation of the starting oligodeox-
ynucleotide; 8 = phosphate DNA; 9 = thiophosphate DNA. *Legends
1ꢀ9 correspond to ODNs 1ꢀ9 respectively, as listed in Tables 2 and 3.
Figure 1. 31P NMR spectrum (300 MHz, D2O, 13 064 scans) of ODN
2 (50- TbAbApCpApCpGpApTpApCpGpCpGbAbT-30).
decreased with increasing number of modifications. A similar
reduction in gel mobility has been reported for oligodeoxynu-
cleotides containing neutral methylphosphonate linkages.4 Such
an effect was also found in zwitterionic oligomers bearing
positively charged ω-aminohexyl groups24 and oligoarginineꢀ
oligonucleotide conjugates.25
Enzymatic Studies. It is known that both the methylphos-
phonate and the borane phosphonate modifications are highly
resistant to exonuclease activity.26ꢀ28 To evaluate the relative
exonuclease susceptibility of the methylborane phosphine-mod-
ified ODNs, we synthesized a series of 16-mers having the
random sequence 50-TAA CAC GAT ACG CGA T-30 with an
increasing number of methylborane phosphine caps and phos-
phate or phosphorothioate gaps. Unmodified phosphate and
phosphorothioate oligomers having the same sequence were
used as controls. The oligomers were tested for stability against
snake venom phosphodiesterase (SVPDE, 30-exonuclease) and
calf spleen phosphodiesterase (CSPDE, 50-exonuclease). For the
SVPDE assay, a sample of the oligomer (1 OD) was incubated at
37 °C with 5 μg of the enzyme in Tris-HCl buffer (pH 9)
containing 10 mM MgCl2. The CSPDE assay was carried out
using 1 OD of the oligomer with 0.1 U enzyme in ammonium
acetate buffer (pH 6.8) containing 2.5 mM EDTA at 37 °C.
Aliquots were removed at different time points, denatured, and
analyzed by RP-HPLC as described in the Supporting Informa-
tion. The half-lives were determined by calculating the time taken
for 50% degradation of the starting oligodeoxynucleotide as
observed by RP-HPLC. ODN 3, with three methylborane
phosphine “caps” on each end and a phosphate “gap”, was found
to have a half-life of ∼2.5 h in the presence of SVPDE. A rapid
degradation of the internal phosphodiester linkages was observed
as soon as the methylborane phosphine cap at the 30 end was
degraded by the enzyme. The all-phosphate DNA control was
fully degraded in less than 5 min under the same experimental
conditions. ODN 2 and ODN 1 with two and one methylborane
phosphine caps on each end respectively were degraded at a
relatively faster rate (2, t1/2 = ∼90 min; 1, t1/2 = ∼ 60 min) by the
enzyme (Figure 2). The stability of oligomers increased at least
6-fold when the internal phosphate linkages were substituted by
thiophosphates. Only a slight increase in the rate of degradation
of ODNs 4 and 6 was observed even when the enzyme
concentration was increased 10-fold. With CSPDE, even a single
modification at each end of the oligomer (ODN 1) prevented the
Figure 3. Nuclease resistance of ODNs against CSPDE. Half-lives are
defined as time taken for 50% degradation of the starting oligodeoxy-
nucleotide; 8 = phosphate DNA; 9 = thiophosphate DNA. *Legends
1ꢀ9 correspond to ODNs 1ꢀ9, respectively, as listed in Tables 2 and 3.
degradation of the vicinal phosphodiester linkages for a longer
time (t1/2 = 6 h; Figure 3). ODN 2 with two modifications at each
end was almost twice as stable (t1/2 > 12 h). Under the same
conditions, the all-phosphate control was completely degraded in
less than 20 min. As anticipated, when the phosphate gap was
substituted with phosphorothioate linkages (ODNs 4, 6), the
nuclease resistance increased considerably (t1/2 >20 h).
Early reports on methylphosphonate dimers suggest that one
diastereoisomer is slowly hydrolyzed by SVPDE.29 It has also
been demonstrated that only the (Sp)-stereoisomer of the
deoxyribonucleoside borane phosphonate dimer is digestible
by SVPDE.27 It was interesting to see whether SVPDE can
specifically degrade one stereoisomer (Rp or Sp) of a methylbor-
ane phosphine linkage. A pentameric oligonucleotide (50-
TTTbTT-30) bearing a single methylborane phosphine linkage
flanked by phosphates was subjected to SVPDE assay under the
above experimental conditions. Analytical RP-HPLC data indi-
cate that there is no enzyme specificity toward either isomer;
both isomers were degraded at a similar rate (Supporting
Information, S19, Figure 15).
Hybridization Experiments. The ability of an oligonucleo-
tide to bind to its target is an important parameter for its
biological activity. To assess the target binding ability of the
methylborane phosphine modified ODNs, duplex hybridization
studies were performed. ODNs 1ꢀ7 were mixed with the cDNA
9848
dx.doi.org/10.1021/ja201314q |J. Am. Chem. Soc. 2011, 133, 9844–9854

Journal of the American Chemical Society
ARTICLE
Table 3. Tm Data for ODN/DNA Duplexes
Table 4. Tm Data for ODN/RNA Duplexes
no. of modifications
no. of modifications
Tma (°C) ΔTm (°C) ΔTm d (°C)
duplex
MePBH3
thio
Tma (°C) ΔTm (°C) ΔTm d (°C)
0
0
duplex
MePBH3
thio
a
1/8
2/8
3/8
4/8
5/8
6/8
7/8
8b/8
9c/8
2
4
6
2
3
4
4
0
0
0
0
53.5
48.0
42.5
45.9
45.0
41.9
40.3
54.5
47.5
1
0.5
1.6
2.0
0.8
0.8
1.4
1.8
1/800
2
4
6
2
4
0
0
0
0
55.5
53.5
48.5
47.1
44.2
56.5
48.0
1
0.5
0.8
1.3
0.5
0.9
6.5
12.0
1.6
2.5
5.6
7.2
2/800
3/800
4/800
6/800
8b/800
9c/800
3.0
8.0
1
0
0
13
12
11
11
0
13
11
0
3.8
15
a All Tm’s represent an average of at least two experiments. b 8 =
phosphate DNA. c 9 = thiophosphate DNA. d ΔTm represents the
0
15
ΔTm per modification; 800 = phosphate RNA.
a All Tm’s represent an average of at least two experiments. b 8 =
phosphate DNA. c 9 = thiophosphate DNA. d ΔTm represents the
0
ΔTm per modification.
their DNA:DNA duplexes. Thus, a stereorandom methylborane
phosphine linkage is more helix stabilizing in a DNA:RNA
duplex.
strand in a 1:1 ratio in citrate buffer (150 mM NaCl, 15 mM
sodium citrate, pH 7.0) to obtain an overall concentration of
0.2 A260 units of the duplex. The samples were denatured at
80 °C and cooled to 15 °C. They were then heated at a rate of
1 °C/min, and A260 versus time was recorded. Melting tempera-
tures were taken as the temperature of half dissociation and were
obtained from the first derivative plots. Tm values for unmodified
phosphate and thiophosphate DNA of the same sequence with
cDNA were determined for comparison.
The duplexes comprising ODNs 1ꢀ7 (Table 3) with com-
plementary DNA showed increase in A260 upon increasing
temperature and gave the characteristic sigmoidal melting curve.
As the number of modifications increased from zero to two, the
Tm dropped from 54.5 to 53.5 °C. This modest depression,
however, was followed by a steep decline to 48.5° when four
modifications were introduced. Two additional modifications
depressed it further to 42.5 °C. Thus, the depression in Tm for
the ODN:DNA duplexes was 0.5 °C/mod for the first two
modifications and by 1.6ꢀ2 °C/mod for successive modi-
fications (Table 3). In the case of the oligomers 4ꢀ7 with
thiophosphate gaps, a similar trend was observed, although the
ΔTm was ꢀ(0.8ꢀ1.8) °C/mod. In comparison, a fully modified
stereorandom borane phosphonate DNA with all four bases
depressed the Tm of a DNA:DNA duplex by 1.3 °C/mod,30
whereas a fully modified stereorandom methyl phosphonate
DNA:DNA duplex was found to have a Tm that was almost
equivalent to the corresponding unmodified phosphodiester
duplex.31
Because the intended intracellular antisense targets are mRNAs,
it was of interest to explore the binding affinity of methylborane
phosphine ODNs toward complementary all-phosphate RNA.
Hybridization studies on ODNs 1ꢀ4 and 6 with RNA showed
that for a given ODN, the ODN:RNA duplex (Table 4) is more
stable relative to its ODN:DNA duplex (Table 3). As observed
for the ODN1:DNA duplex, the ΔTm for the ODN:RNA duplex
was ꢀ0.5 °C/mod for the first two modifications introduced.
Four modifications depressed the Tm by ꢀ0.8 °C/mod as
opposed to ꢀ1.6 °C/mod for the ODN2:DNA duplex, and six
modifications depressed the Tm by ꢀ1.3 °C/mod as opposed
to ꢀ2.0 °C/mod for the ODN3:DNA duplex. Following the
same trend, the depression in Tm was 0.5 and 0.9 °C/mod in
a DNA:RNA duplex for ODNs 4 and 6 with thiophosphate
gaps, as opposed to 0.8 and 1.4 °C/mod, respectively, for
A comparable trend has been observed in the case of borane
phosphonate DNA duplexes. Shimizu et al. have reported that a
DNA:RNA duplex of stereorandom borane phosphonate DNA
with its complementary phosphate RNA [d(CbAbGbTb)3:r-
(ACUG)3, (Tm = 45.0 °C)] is more stable as compared to a
DNA:DNA duplex of stereorandom borane phosphonate with
complementary phosphate DNA30 [d(CbAbGbTb)3:d(ACTG)3]
(Tm = 39.7 °C). The control all-phosphate DNA:RNA duplex
has a Tm of 52.4 °C under the same conditions. In contrast, the
DNA:RNA duplex of the stereorandom methylphosphonate
oligodeoxynucleotide (CmTm)7A:RNA has a highly depressed
Tm (ΔTm = ꢀ26.5 °C) relative to the all-phosphate DNA:RNA
control.32 In the case of borane phosphonates as well as
methylborane phosphine ODNs, the reasons for the higher
binding affinity for a complementary RNA strand as opposed
to a DNA strand could arise from the different conformations of
the resultant DNA:DNA and DNA:RNA duplexes.33 However,
more in-depth thermodynamic and structural studies need to be
undertaken to precisely determine the reasons behind this effect.
Barawkar et al. have reported that DNA:DNA and DNA:RNA
duplexes of oligonucleotides containing internucleoside guani-
dinium linkages are stabilized when the ionic strength is
lowered.2 This salt-effect is opposite of that observed in the case
of phosphate DNA:DNA and DNA:RNA duplexes where in-
creasing the salt concentration improves the stability of the
duplex by reducing electrostatic repulsion between the negatively
charged phosphate backbones. The reverse salt effect arises due
to the strong interactions of the positively charged guanidinium
moiety with the phosphodiester linkage of the complementary
strand, which becomes the predominant electrostatic interaction
under low salt concentrations.
If the methylborane phosphine linkage were positively
charged, we expected to see such an effect during the hybridiza-
tion of chimeric ODNs with cDNA and RNA. To test this
hypothesis, the hybridization experiments of ODNs 1ꢀ3 with
RNA were carried out at low salt concentrations (10 mM NaCl,
10 mM Na2HPO4, pH 7.1). Indeed, under these conditions, the
binding affinity of ODNs 1ꢀ3 was found to be higher. ODN 1
with two methylborane phosphine linkages exhibited a Tm of
39.5 °C. As expected for an increasing number of positive
charges, the Tm for the ODN2:RNA duplex was 41.5 °C, and
that for the ODN3:RNA duplex was 41.9 °C. The Tm for the
9849
dx.doi.org/10.1021/ja201314q |J. Am. Chem. Soc. 2011, 133, 9844–9854

Journal of the American Chemical Society
ARTICLE
Chart 1. Bar Diagram Showing the Percentage of HeLa Cells
Showing Uptake at Various Concentrations of ODNs 1ꢀ3, 8,
10, 11 after 16 h of Transfectiona
These experiments revealed that the uptake of the methylbor-
ane phosphine ODNs was highly dependent on concentration as
well as the percentage of modifications introduced. The cells
treated with the DNA control exhibited a very weak fluorescence
at all concentrations. For all ODNs studied, there was an increase
in uptake efficiency when the concentration was increased from
1.0 to 4.0 μM. In the case of ODNs 1ꢀ3, a plateau of maximum
uptake was reached around 6ꢀ8.0 μM. The cellular uptake
pathway appeared to be saturable at higher concentrations,
suggesting the possible involvement of cell-surface receptors.
In contrast, the increase in cellular accumulation of methylpho-
sphonate oligonucleotides which predominantly occurs by
endocytotic pathways was shown to be unsaturable even at
100 μM.35 Another interesting observation was that the percentage of
cells showing uptake was not directly proportional to the number
of modifications on the ODNs. Thus, ODN 2 with two mod-
ifications at both the 50 and 30 ends seemed to be more efficiently
taken up by cells as compared to ODNs 1, 3, and 10; ODN 3 with
three caps (six modifications) was comparable to ODN 1, but
slightly less efficient than ODN 2. ODN 10 with eight modifica-
tions and the ODN 11 variant did not show improved cell uptake
properties relative to ODN 3. It was also noted that cap-gap
ODN 10 was more efficient than ODN 11 at all concentrations
studied even though both had 8 methylborane phosphine
linkages. Although there are numerous reports in the literature
wherein a single molecule has been conjugated to DNA or RNA
to improve cell delivery,36ꢀ39 there are no studies detailing the
effect of increasing number of borane phosphonate modifications
on uptake. Further cell uptake experiments and mechanistic
studies dealing with single-stranded borane phosphonate oligo-
nucleotides and methylborane phosphine ODNs will be needed
to confirm whether this effect is brought about by the BH3 moiety,
or if it is specific to the methylborane phosphine linkage or the
positive charge, which appears to be associated with this site.
Fluorescence Microscopy. It has been demonstrated that
chemically modified oligonucleotides are generally taken up by
cells through various endocytotic pathways35,40 leading to en-
trapment in endocytotic vesicles, which results in poor biological
activity. Subcellular distribution patterns of methylborane phos-
phine ODNs after 24 h of passive transfection were visualized by
fluorescence microscopy. HeLa cells were incubated with 5.0 μM
fluorescein-tagged ODN 12 (TbTbTpTpTpTpTpTpTpTpTpTp-
TpTbTbT, 16-mer) for 24 h. The cells were washed with D-PBS,
trypsinized to remove membrane-bound fluorescence, pelleted,
and fixed by formalin before analysis. The fluorescence intensity
was found to be both vesicular and diffuse in the cytoplasm, but
no intensity was detected in the nucleus (Figure 4).
aThe data represent the means ( SEM of three independent experiments.
all-phosphate DNA:RNA control duplex was 41.9 °C. Thus, the
binding affinity of ODN3 with complementary RNA approached
that for the unmodified DNA:RNA duplex at 10 mM NaCl (close
to physiological conditions), suggesting that the methylborane
phosphine moiety might be positively charged. These results are
encouraging relative to the use of methylborane phosphine
containing ODNs in various biological experiments carried out
under physiological salt conditions.
Flow Cytometric Experiments. Encouraged by the results
obtained from preliminary biochemical experiments, we next
explored the cell-uptake properties of single-stranded methyl-
borane phosphine ODNs as a function of an increasing number
of modifications. For this purpose, 50-fluorescein labeled ODNs
1ꢀ3 (Table 2) were synthesized. In an attempt to compare the
efficiency of a cap-gap ODN versus an ODN having alternating
methylborane phosphine and phosphate internucleotide lin-
kages, fluorescein labeled ODN 10 (TbAbAbCbApCpGpApTpAp-
CpGbCbGbAbT) and 11 (TbApAbCpAbCpGbApTbApCbGpCbGp-
AbT) were synthesized. The fluorescein tags on ODNs 1ꢀ3, 8,
10, and 11 were introduced via borane phosphonate linkages to
reduce the rate of enzymatic hydrolysis resulting in the release of
fluorescein. The fluorescence detected by FACS was expected to
arise from fluorescein attached to intact ODNs because Fisher
et al. have shown that fluorescein, if detached from ODNs, is
quickly discarded from cells.34
We first studied the uptake of methylborane phosphine ODNs
1ꢀ3, 10, and 11 in HeLa cells. Exponentially growing HeLa cells
maintained as subconfluent monolayers in Dulbecco’s Modified
Eagle Medium (DMEM) were seeded on 12-well culture plates
24 h before transfection. During transfection, the medium was
replaced with DMEM premixed with the ODNs at various
concentrations (1.0, 4.0, 6.0, and 8.0 μM), and the cells were
further incubated for 16 h. The cells were then rinsed three times
with D-PBS (Dulbecco’s phosphate buffer saline), trypsinized,
and resuspended in D-PBS. Fluorescence intensity was analyzed
for cells presenting higher fluorescence than the background
(Chart 1). The background was defined as the autofluorescence
of cells incubated only with DMEM during transfection; single-
stranded 50-fluorescein tagged phosphate DNA (8) was used as
the control.
Because cellular uptake efficiency is known to be highly
dependent on cell type, we extended our studies to a human
melanoma cell line, WM-239A. WM-239A cells were transfected
with fluorescein tagged ODN 3 premixed in RPMI to give a final
concentration of 8.0 μM. The cells were fixed with formalin
before analysis, as described in the Experimental Section. After
20 h of incubation, fluorescence was found mainly within
the cytoplasm and to a very low extent in the nucleus. The
cytoplasmic fluorescence appeared to be distributed throughout
the cytoplasm in a diffuse, not punctated manner (Figure 5).
These experiments imply that methylborane phosphine ODNs
are not confined to vesicular structures inside cells and therefore
might show potent biological activity. Even as the mechanism of
uptake of these oligonucleotides remains largely unclear, further
studies willaddressthe temperaturedependenceof internalization,
9850
dx.doi.org/10.1021/ja201314q |J. Am. Chem. Soc. 2011, 133, 9844–9854

Journal of the American Chemical Society
ARTICLE
Figure 4. Uptake of fluorescein labeled ODN 12 into HeLa cells in the absence of transfecting agents. (A) Auto fluorescence of HeLa cells. (B)
Incubation of cells with ODN 12 (5.0 μM, 24 h) followed by formalin fixation. (C,D) Incubation of cells with ODN 12 (5.0 μM, 24 h) followed by
formalin fixation, nucleus stained with DAPI. (C) versus (D) corresponds to the cells photographed under phase-contrast versus the same cells under
fluorescence microscopy.
Figure 5. (A) Fluorescence microscopy image (20ꢁ magnification) of WM239A cells transfected with 8.0 μM ODN 2 for 20 h followed by formalin
fixation. (B) versus (C) corresponds to WM239A cells photographed at 40ꢁ magnification under phase-contrast (B) versus the same cells under
fluorescence microscopy using the same transfection conditions as for the cells shown in (A).
time-dependent efflux rates, and also how well these modifications
improve the potency of ODNs when applied to various gene
knock-down assays such as the Dual Luciferase Reporter assay41
and the Let-7b assay (Applied Biosystems, assay ID 002619).
Preliminary cytotoxicity assays have been carried out using
fluorescein labeled ODN 3 on HeLa cells and the WM-239A cell
line. The cells were transfected with ODN 3 at 6 μM concentra-
tion for 24 h and stained with propidium iodide as described in
the Supporting Information (S20). The number of dead cells was
determined by flow cytometry. About 5% dead cells were found
in the HeLa cell line (Supporting Information, S20, Figure 16) as
opposed to 3% for the control with no ODN. In the WM-239A
cell line, the apoptosis was 7% (Supporting Information, S20,
Figure 17) as opposed to 5% in the control experiment.
We have successfully developed a method for synthesizing
ODNs containing multiple methylborane phosphine internu-
cleotide linkages. This strategy required the choice of suitable
protecting groups and cleavage conditions that are compatible
9851
dx.doi.org/10.1021/ja201314q |J. Am. Chem. Soc. 2011, 133, 9844–9854

Journal of the American Chemical Society
ARTICLE
with both the PꢀC internucleotide bond and borane chemistry
for removing the ODNs from the support. Using this chemical
approach, chimeric methylborane phosphine ODNs (16-mers)
with phosphate and thiophosphate internucleotide linkages can
be synthesized in reasonable yields. A single modification at each
end generated an oligomer that was approximately 20 times more
stable than native DNA toward SVPDE digestion; the nuclease
resistance toward CSPDE was enhanced by at least 30 fold.
Despite the methyl and borane groups being nearly isosteric with
the phosphate, the polarity and charge distribution of the
methylborane phosphine linkage are unlike those of either the
methylphosphonate or the borane phosphonate, and are ex-
pected to give rise to interesting and unforeseen biological
activity.
Hybridization experiments suggest that, like borane phospho-
nates, these linkages are more helix destabilizing in a DNA:DNA
duplex than a DNA:RNA duplex. It should also be noted that
stereorandom methylborane phosphineꢀDNA chimeras have
much higher binding affinity to a complementary RNA strand
relative to stereorandom methylphosphonate oligonucleotides in
a DNA:RNA duplex, possibly due to the structural differences
between the duplexes.
linkage being positively charged. Further research will include
structural characterization of these molecules so as to gain a
better understanding of their three-dimensional conformation.
Preliminary studies on the cell-uptake of chimeric ODNs with
varying number of modifications indicate that they are very
efficiently taken up by cells without the aid of transfecting agents
even when the ODN is only 13% modified. Fluorescence
microscopy shows that the ODN is localized in the cytoplasm
and not bound to the cell membrane. These results indicate that
methylborane phosphine chimeras could possibly show potent
biological activity when applied to in vivo experiments.
’ CONCLUSION
In summary, we have demonstrated that methylborane phos-
phine linkages have increased lipophilicity, impart resistance to
cellular nucleases, have duplex stabilities similar to natural DNA:
RNA duplexes under physiological salt conditions, and are
efficiently taken up by various cell lines without the aid of
transfecting agents. These desirable qualities make the methyl-
borane phosphine modification an ideal candidate for various
biological applications such as antisense therapy, RNAi experi-
ments, and as antagomirs against miRNAs.
Observations on the slower electrophoretic mobility of chi-
meric ODNs (relative to its all-phosphate control) and the low
salt effect on the stability of their duplexes with unmodified RNA
lead to a discussion of the nature of charge distribution as well as
the structure of the tetra-coordinated phosphorus in the methyl-
borane phosphine linkage. It has been theorized, on the basis of
Gaussian 94 calculations, that the reduction of atomic charge on
phosphorus roughly correlates with chemical shifts observed for
borane phosphonates and phosphodiester linkages.42 Phosphorus
that bears a theoretical atomic charge of þ2.78 (phosphate, P(V))
resonates at ꢀ2.41 ppm, whereas the phosphorus of the borane
phosphonate linkage (atomic charge: þ2.36, P(V)) resonates at
95.0 ppm. In comparison, (MeO)3P (P(III)) with an atomic
charge of þ1.77 resonates at þ140.0 ppm. Following this trend, a
chemical shift of þ148.0 ppm would suggest that the atomic
charge on the MeꢀPꢀBH3 phosphorus (P(IV) species) might
be slightly lower than þ1.77. Alternatively, it could also be
hypothesized that the phosphorus (PꢀIV) in the MeꢀPꢀBH3
linkage exists as a phosphonium species. Tetra-coordinated
phosphonium-based ionic liquids are well-known in the literature
and have been reported to be stable even at elevated temperatures.43
X-ray crystallographic data on decyl(tricyclohexyl) phosphonium
cation have shown that the phosphorus center adopts a near-
symmetrical tetrahedral environment, although with consistently
shorter PꢀC bond lengths (ꢀ0.038 Å) than its phosphine (PꢀIII)
counterpart.44 Borane phosphonate and methyl phosphonate di-
esters are also known to adopt a tetrahedral geometry around the
phosphorus.42,45 Interestingly, in contrast to the bond shortening
that was observed for the phosphonium species, the PꢀB bond
(∼1.905 Å) in a borane phosphonate diester42 and the PꢀC bond
(∼1.77 Å) in a methylphosphonate diester45 are longer than their
respective PdO bonds (1.52, 1.48 Å). Thus, irrespective of the
charge on the phosphorus center or the distortion in PꢀC andPꢀB
bond lengths, we speculate that the MeꢀPꢀBH3 internucleotide
linkage might presumably adopt a tetrahedral geometry. Our current
research has yielded no definitive proof to support either the
existence of a positively charged phosphonium species or that a
neutral species as has been postulated.46 However, both the
electrophoretic mobility data and the Tm measurements at low salt
conditions suggest the possibility of the methylborane phosphine
’ EXPERIMENTAL SECTION
Chemical Synthesis. The general procedure for the synthesis of
protected 20-deoxynucleoside-30-O-methyl phosphinoamidites (13ꢀ16)
is as follows: Protected 20-deoxynucleosides (8ꢀ11, Supporting Infor-
mation, S2ꢀS3) were dried in vacuo for 6 h. The protected
20-deoxynucleosides (0.02 mol), 4,5-dicyano imidazole (0.018 mol,
0.9 equiv) (DCI), and anhydrous dichloromethane (100 mL) were
taken up in a 250 mL round-bottom flask and stirred under nitrogen.
Anhydrous CH2Cl2 (250 mL), bis(N,N-diisopropylamino)methyl-
phosphine (0.02 mols, 1.0 equiv), and a magnetic stir bar were placed
in a 500 mL round-bottom flask fitted with a 150 mL addition funnel.
The freshly prepared deoxynucleosideꢀDCI solution was taken up in
the addition funnel and added slowly over 30 min to the 500 mL round-
bottom flask, under nitrogen. The solution was allowed to stir for an
additional 30 min. A small amount of triethylamine (approximately
0.2 mL) was added to neutralize the solution, and the solvent was removed
in vacuo. The crude product was isolated by chromatography using a
30ꢀ80% gradient of ethylacetate in hexane containing 1% triethylamine.
Compound 13. 50-O-Silyl-N6-trimethoxytrityl-20-deoxyadenosine-
30-O-methylphosphinoamidite.
Yield 47.7%. 31P NMR (CDCl3): δ 121.36, 119.95. 13C NMR
(CDCl3): δ 158.2, 154.12, 152.19, 148.56, 144.11, 143.06, 144.03,
138.19, 130.02, 128.23, 128.21, 128.18, 127.14, 127.11, 126.36,
126.30, 126.24, 121.23, 121.19, 116.41, 113.08, 86.86, 86.62, 84.35,
82.21, 76.18, 76.07, 75.95, 75.85, 70.17, 63.47, 55.20, 44.26, 44.13, 39.32,
24.70, 24.02, 18.30, 18.21, 18.15, 18.05, 1.53, 1.51, 1.49.
Compound 14. 50-O-Silyl-N2-trimethoxytrityl-20-deoxyguanosine-
30-O-methylphosphinoamidite.
Yield 40.5%. 31P NMR (CDCl3): δ 119.65, 118.16. 13C NMR
(CDCl3): δ 158.4, 158.28, 151.32, 151.19, 150.46, 150.28, 144.06,
144.00, 139.72, 134.86, 129.85, 128.24, 127.18, 126.35, 126.27,
117.39, 113.56, 113.36, 113.24, 113.11, 86.30, 83.92, 83.66, 75.51,
69.43, 63.21, 55.24, 55.15, 44.22, 44.09, 40.34, 24.71, 24.13, 24.03,
18.25, 18.18, 18.09, 18.02, 1.94, 1.85, 1.54, 1.52, 1.51, 1.15.
Compound 15. 50-O-Silyl-N4-trimethoxytrityl-20-deoxycytidine-30-
O-methylphosphinoamidite.
Yield 45.3%. 31P NMR (CDCl3): δ 120.04, 119.47. 13C NMR
(CDCl3): δ 158.55, 158.40, 158.28, 151.32, 151.19, 150.46, 50.28,
144.04, 143.95, 139.72, 134.86, 128.24, 127.18, 126.35, 126.27, 117.39,
9852
dx.doi.org/10.1021/ja201314q |J. Am. Chem. Soc. 2011, 133, 9844–9854

Journal of the American Chemical Society
ARTICLE
113.56, 113.36, 113.24, 113.11, 86.30, 83.92, 83.66, 75.51, 69.81, 69.77,
63.43, 63.21, 55.24, 55.15, 44.22, 44.09, 40.34, 24.71, 24.13, 24.03, 18.25,
18.18, 18.09, 18.02, 1.94, 1.85, 1.54, 1.52, 1.51, 1.14.
20-deoxyoligonucleotide transfection. The cells were used at passages
12ꢀ18. WM-239A cell line was a gift from the Natalie Ahn laboratory,
Department of Chemistry and Biochemistry, University of Colorado,
Boulder. Cells were routinely maintained as monolayers in RPMI contain-
ing 10% fetal bovine serum, penicillin (100 U/mL), and streptomycin
(100 U/mL), and used at passages 7ꢀ12.
Compound 16. 50-O-Silyl-20-deoxythymidine-30-O-methylphosphinoamidite.
Yield 60%. 31P NMR (CDCl3): δ 121.18, 120.96. 13C NMR
(CDCl3): δ 163.53, 150.15, 135.85, 126.11, 110.79, 110.73, 87.43,
86.47, 85.00, 87.46, 76.21, 75.95, 63.43, 63.33, 44.28, 44.15, 40.14,
39.61, 24.74, 24.16, 18.26, 18.11, 12.70, 1.54, 1.52, 1.49.
Cell-Uptake Experiments. For uptake experiments, ∼1 ꢁ 105
cells/well (12-well plates) were incubated in appropriate medium
containing 10% fetal bovine serum, penicillin (100 U/mL), and strep-
tomycin (100 U/mL) for 24 h. The concentration (OD) of 50-
fluorescein labeled, chimeric methylborane phosphine ODN dissolved
in siRNA buffer 1X (Thermoscientific) was measured by UV spectros-
copy. The medium was removed, and the cells were transfected with the
ODN in DMEM or RPMI to give the required final concentration. The
cells were then incubated at 37 °C for 16 h. After the incubation, the
medium was removed from the wells and cells washed three times with
0.5 mL of D-PBS. Cells were then treated for 3 min at 37 °C with a
prewarmed solution of trypsin-EDTA (1X) until all cells became round
and detached from the plates; the cells from each plate were taken up in
1 mL of D-PBS and centrifuged at 1000 rpm for 5 min to form cell
pellets. The pellets were washed and resuspended in D-PBS and kept at
0 °C in the dark for flow cytometric analysis.
Soild-Phase Synthesis. The chemical synthesis of chimeric
methylborane phosphine 20-deoxyoligonucleotides was accomplished
using an ABI model 392 automated DNA synthesizer. The synthesis
cycle (Table 1) was a slightly modified version of the previously
reported procedure.47 The 50-O-silyl-N-trimethoxytrityl-30-O-meth-
oxy-20deoxynucleoside phosphoramidites of A (17), G (18), C (19)
and the 50-O-silyl-20-deoxythymidine-30-O-methoxyphosphoramidites
(20) were prepared as previously reported.47 Suitably protected 30O-
methylphosphinoamidites 13ꢀ16 as well as the 30O-methoxyphosphor-
amidites 17ꢀ20 were dissolved in anhydrous acetonitrile at 100 mM and
placed on the appropriate ports of the synthesizer. The coupling wait time
was 500 s for the 30-O-methylphosphinoamidites 13ꢀ16; it was decreased
to 180 s for the 30-O-methoxyphosphoramidites 17ꢀ20. The activator
was 5-ethylthio-1H-tetrazole (0.25 M in anhydrous acetonitrile, Glen
Research, VA). After each amidite condensation step, the support was
washed with acetonitrile for 40 s, and Unicap Phosphoramidite (Glen
Research, VA) was used as per the manufacturer’s recommendations for
capping the failure sequences. Anhydrous tert-butyl hydroperoxide17 or
Beaucage reagent19 was used to generate the phosphate or the thiophos-
phate triester linkage, respectively. Boronation of the methylphosphonite
internucleotide linkages was carried out using a freshly prepared solution
Flow Cytometry. Flow cytometric data on at least 10 000 cells per
sample were acquired on a Moflow flow cytometer (Beckman-Coulter)
equipped with a single 488 nm argon laser, 530/40 nm emission filter
(Fluorescein). Raw flow cytometry data were manipulated and visua-
lized using Summit 4.3 software (Beckman-Coulter). Fluorescence
intensity of the 50-fluorescein tag was analyzed for cells presenting
higher fluorescence than the background. The background was defined
asthe auto fluorescenceof cells incubatedonlywith the vehicle (DMEMþ
siRNA buffer).
Fluorescence Microscopy. An inverted microscope (Olympus
IX 81) equipped with a Hamamatsu C4742-95 CCD and CoolSNAP ES
digital camera (Photometrics) was used for fluorescence microscopy.
For microscopic analysis, ∼1 ꢁ 105 cells were seeded onto sterile
coverslips (18 mm diam, round, VWR) placed in 12-well plates and
incubated in appropriate medium containing 10% fetal bovine serum,
penicillin (100 U/mL), and streptomycin (100 U/mL) for 24 h. The
cells were transfected with the chimeric ODNs and incubated at 37 °C
for 20 h. After the incubation, the medium was removed from the wells
and cells washed four times with 0.5 mL of D-PBS. The cells were
covered with 1 mL of 10% neutral, buffered formalin solution (Sigma-
Aldrich) for 15 min. The formalin solution was removed, and the cells
were washed with 2 mL of D-PBS for 10 min at room temperature. The
coverslips were carefully removed from the wells and mounted upside-
down on cover slides (25 ꢁ 75 mm, 1.0 mm thick) using Fluoromount-
G (Southern Biotech) as the mounting medium, and were observed
under the microscope.
of 25 mM BH3 THF in THF.
3
Postsynthesis, the polystyrene (PS) was washed with anhydrous
acetonitrile for 60 s and then flushed with a stream of argon until dry.
The bases were detritylated using an 80% aqueous acetic acid solution
for 24 h. The methyl groups were then removed from the phosphate
backbone using 2-carbamoyl-2-cyanoethylene-1,1-dithiolate trihydrate
(Acacia Lifesciences, IL) in DMF. The polystyrene beads were thor-
oughly washed with water followed by anhydrous acetonitrile and
dichloromethane and flushed with argon until dry. The PS support
was removed from the column and placed in a 1.5 mL screw-cap, conical
glass reaction vial. A 50% solution of ethylenediamine in ethanol/
acetonitrile/water (47:48:5 v/v) was added, and the vial was sealed.
The vial was vortexed to stir the contents, and the cleavage was allowed
to proceed for 120 min at room temperature. The vial’s contents were
transferred to a 1.5 mL Eppendorf tube and evaporated to dryness in a
SpeedVac. The crude ODN was redissolved in 5% acetonitrileꢀwater
and purified by RP-HPLC.
Nuclease Stability Determination. Snake Venom Phosphodies-
terase I (Crotalus adamanteus) and calf spleen phosphodiesterase II were
purchased from Sigma (St. Louis, MO). While performing enzymatic
hydrolysis experiments, the oligodeoxynucleotide (1 OD) was incu-
bated at 37 °C in either 150 μL of 100 mM Tris-HCl buffer (pH 9.0),
10 mM MgCl2 (5 μg), Snake Venom Phosphodiesterase I, and the total
volume made up to 200 μL; or 160 μL of 125 mM ammonium acetate
buffer (pH 6.8), 2.5 mM EDTA, and 0.1 U of CSPDE, made up to a total
volume of 200 μL. Aliquots of the reaction mixtures were removed at the
indicated time points, quenched by the addition of 1.0 M EDTA
(SVPDE) or 2.0 M Urea (CSPDE), and stored in dry ice until analyzed
by analytical RP-HPLC.
Preparation of Cells. HeLa cells were obtained from American
Type Culture Collection (Rockville, MD) and serially maintained
at monolayer cultures in a humidified atmosphere of 5% CO2 at
37 °C in Dulbecco’s Modified Eagle’s Medium (DMEM) containing
10% fetal bovine serum, penicillin (100 U/mL), and streptomycin
(100 U/mL) in 12-well culture plates, and incubated for 24 h prior to
’ ASSOCIATED CONTENT
S
Supporting Information. Procedures for the chemical
b
synthesis of 50-O-silyl-20-deoxynucleotides-30-O-methylphos-
phoroamidites, 31P NMR data, analytical RP-HPLC profiles,
and MALDI-TOF data for ODNs, data on PAGE, enzyme
degradation experiments performed using SVPDE to study the
relative stability of (Rp)- and (Sp)- enantiomers, and apoptosis
assay protocol. This material is available free of charge via the
Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
marvin.caruthers@colorado.edu
9853
dx.doi.org/10.1021/ja201314q |J. Am. Chem. Soc. 2011, 133, 9844–9854

Journal of the American Chemical Society
ARTICLE
’ ACKNOWLEDGMENT
(26) Miller, P. S.; Dreon, N.; Pulford, S. M.; McParland, K. B. J. Biol.
Chem. 1980, 255, 9659.
(27) Li, P.; Sergueeva, Z. A.; Dobrikov, M.; Shaw, B. R. Chem. Rev.
2007, 107, 4746.
(28) Chen, Y.-Q.; Qu, F.-C.; Zhang, Y.-B. Tetrahedron Lett. 1995,
36, 745.
(29) Agarwal, K. L.; Riftina, F. Nucleic Acids Res. 1979, 6, 3009.
(30) Shimizu, M.; Saigo, K.; Wada, T. J. Org. Chem. 2006, 71, 4262.
(31) Vyazovkina, E. V.; Savchenko, E. V.; Lokhov, S. G.; Engels,
J. W.; Wickstrom, E.; Lebedev, A. V. Nucleic Acids Res. 1994, 22, 2404.
(32) Reynolds, M. A.; Hogrefe, R. I.; Jaeger, J. A.; Schwartz, D. A.;
Riley, T. A.; Marvin, W. B.; Daily, W. J.; Vaghefi, M. M.; Beck, T. A.;
Knowles, S. K.; Klem, R. E.; Arnold, L. J. Nucleic Acids Res. 1996,
24, 4584.
We thank Rich Shoemaker for assistance with the NMR
facility, Theresa Nahreini for FACS data collection, and the
University of Colorado Central Analytical Laboratories for use of
the mass spectral facility. This research was supported by the
University of Colorado at Boulder.
’ REFERENCES
(1) Maier, M. A.; Guzaev, A. P.; Manoharan, M. Org. Lett. 2000,
2, 1819.
(2) Barawkar, D. A.; Bruice, T. C. Proc. Natl. Acad. Sci. U.S.A. 1998,
95, 11047.
(33) Lesnik, E. A.; Freier, S. M. Biochemistry 1995, 34, 10807.
(34) Fisher, T. L.; Terhorst, T.; Cao, X.; Wagner, R. W. Nucleic Acids
Res. 1993, 21, 3857.
(35) Levis, J. T.; Butler, W. O.; Tseng, B. Y.; Ts’O, P. O. P. Antisense
Res. Dev. 1995, 5, 251.
(36) Habus, I.; Zhao, Q.; Agrawal, S. Bioconjugate Chem. 1995,
6, 327.
(37) Ugarte-Uribe, B. A; Perez-Rentero, S.; Lucas, R.; Avino, A.;
Reina, J. J.; Alkorta, I.; Eritja, R. n.; Morales, J. C. Bioconjugate Chem.
2010, 21, 1280.
(38) Jeong, J. H.; Mok, H.; Oh, Y.-K.; Park, T. G. Bioconjugate Chem.
2008, 20, 5.
(39) Manoharan, M. Antisense Nucleic Acid Drug Dev. 2002, 12, 103.
(40) Gray, G. D.; Basu, S.; Wickstrom, E. Biochem. Pharmacol. 1997,
53, 1465.
(41) Hawkins, E.; Beck, M.; Butler, B.; Wood, K. Promega Notes
2003, 85, 31.
(42) Summers, J. S.; Roe, D.; Boyle, P. D.; Colvin, M.; Shaw, B. R.
Inorg. Chem. 1998, 37, 4158.
(43) Tsunashima, K.; Sugiya, M. Electrochem. Commun. 2007,
9, 2353.
(44) Ermolaev, V.; Miluykov, V.; Rizvanov, I.; Krivolapov, D.;
Zvereva, E.; Katsyuba, S.; Sinyashin, O.; Schmutzler, R. Dalton Trans.
2010, 39, 5564.
(3) Lin, J.; Shaw, B. R. Nucleosides, Nucleotides Nucleic Acids 2001,
20, 1325.
(4) Miller, P. S. In Progress in Nucleic Acid Research and Molecular
Biology; Waldo, E. C., Kivie, M., Eds.; Academic Press: New York, 1996;
Vol. 52, p 261.
(5) Other instances of oligonucleotide “hybrids” known in the
literature include oligonucleotide N30ꢀP50 methanephosphonami-
dates: Nawrot, B.; Boczkowska, M.; Wꢀojcik, M.; Sochacki, M.; Kazmierski,
S.; Stec, W. J. Nucleic Acids Res. 1998, 26, 2650. N30ꢀP50 thiopho-
sphoramidates: Pongracz, K.; Gryaznov, S. Tetrahedron Lett. 1999, 40,
7661. Alkyl/aryl phosphonothioates:Mag, M.; Muth, J.; Jahn, K.; Peyman,
A.; Kretzschmar, G.; Engels, J. W.; Uhlmann, E. Bioorg. Med. Chem. 1997,
5, 2213.
(6) Miller, P. S.; McParland, K. B.; Jayaraman, K.; Tso, P. O. P.
Biochemistry 1981, 20, 1874.
(7) Hall, A. H. S.; Wan, J.; Spesock, A.; Sergueeva, Z.; Shaw, B. R.;
Alexander, K. A. Nucleic Acids Res. 2006, 34, 2773.
(8) McCuen, H. B.; Noꢀe, M. S.; Sierzchala, A. B.; Higson, A. P.;
Caruthers, M. H. J. Am. Chem. Soc. 2006, 128, 8138.
(9) When H-phosphonate chemistry is used for boranephosphonate
synthesis, thymine has to be protected at N3: Higson, A. P.; Sierzchala,
A.; Brummel, H.; Zhao, Z.; Caruthers, M. H. Tetrahedron Lett. 1998,
39, 3899.
(10) Sergueeva, Z. A.; Sergueev, D. S.; Shaw, B. R. Nucleosides,
(45) Han, F.; Watt, W.; Duchamp, D. J.; Callahan, L.; Kꢀezdy, F. J.;
Agarwal, K. Nucleic Acids Res. 1990, 18, 2759.
(46) Shaw, B. R.; Dobrikov, M.; Wang, X. I. N.; Wan, J.; He, K.; Lin,
J.-L.; Li, P.; Rait, V.; Sergueeva, Z. A.; Sergueev, D. Ann. N. Y. Acad. Sci.
2003, 1002, 12.
Nucleotides Nucleic Acids 2000, 19, 275.
(11) Sergueeva, Z. A.; Sergueev, D. S.; Shaw, B. R. Nucleosides,
Nucleotides Nucleic Acids 2001, 20, 941.
(12) Higson, A. P.; Sierzchala, A.; Brummel, H.; Zhao, Z.; Caruthers,
M. H. Tetrahedron Lett. 1998, 39, 3899.
(47) McCuen, H. B.; Noe, M. S.; Olesiak, M.; Sierzchala, A. B.;
Caruthers, M. H.; Higson, A. P. Phosphorus, Sulfur Silicon Relat. Elem.
2008, 183, 349.
(13) Shimizu, M.; Wada, T.; Oka, N.; Saigo, K. J. Org. Chem. 2004,
69, 5261.
(14) Roelen, J. C. P. F.; van den Elst, H.; Dreef, C. E.; van der Marel,
G. A.; van Boom, J. H. Tetrahedron Lett. 1992, 33, 2357.
(15) CPG support could not be used due to its incompatibility
toward the HF mixture used for removing the 50-O-silyl ether.
(16) Glen research, Sterling, VA, used as per manufacturer’s
instructions.
(17) Hayakawa, Y.; Wakabayashi, S.; Kato, H.; Noyori, R. J. Am.
Chem. Soc. 1990, 112, 1691.
(18) Undesired side reactions were observed when 0.10 M iodine in
tetrahydrofuran/pyridine/water was used as oxidizer.
(19) Iyer, R. P.; Egan, W.; Regan, J. B.; Beaucage, S. L. J. Am. Chem.
Soc. 1990, 112, 1253.
(20) Hogrefe, R. I.; Vaghefi, M. M.; Reynolds, M. A.; Young, K. M.;
Arnold, L. J. Nucleic Acids Res. 1993, 21, 2031.
(21) Trityl yields per condensation could not be calculated because
50-DMT chemistry was not employed.
(22) Belabassi, Y.; Antczak, M. I.; Tellez, J.; Montchamp, J.-L.
Tetrahedron 2008, 64, 9181.
(23) http://www.glenresearch.com/ProductFiles/10-4410.html.
(24) Hashimoto, H.; Nelson, M. G.; Switzer, C. J. Am. Chem. Soc.
1993, 115, 7128.
(25) Wei, Z.; Tung, C.-H.; Zhu, T.; Stein, S. Bioconjugate Chem.
1994, 5, 468.
9854
dx.doi.org/10.1021/ja201314q |J. Am. Chem. Soc. 2011, 133, 9844–9854