Full Paper
1
0th, 16th position and the 5’-end prephosphorylated anti-
se A (RNase A) was also examined (Figure S7). After 0.5 h incu-
bation, unmodified siRNA duplex was significantly hydrolyzed,
whereas AS23/Sc and AS24/Sc, corresponding to three and
four photolabile modifications at the antisense strand, suffered
much less degradation within 2 h, suggesting that siRNA du-
plexes carrying multiple caging groups were significantly more
resistant to RNase A compared with the native siRNA duplex.
Finally, the stability of siRNAs in PBS buffer containing fetal
bovine serum was investigated. Modified siRNA duplexes, simi-
lar to the native siRNA (ASc/Sc), suffered only low levels of hy-
drolysis after 1 h incubation, and the corresponding hydro-
lyzed bands increased to a similar level when the incubation
time was extended to 7 h (Figure S8). This indicated that
caging modification did not decrease the stability of siRNA du-
plexes in the presence of fetal bovine serum.
sense strand (FAScP) were also synthesized (Figure 3). RNAi ac-
tivities of their duplexes with natural sense strand (Sc) were
further investigated according to different siRNA interaction re-
gions with Ago2 (Figure S6). For the 5’-terminus caged siRNA
duplex (FAS00/Sc), a clear inactivation of siRNA activity was ob-
served, which was identical to that of the negative control
siRNA. After UV irradiation (2 min), the silencing activity of
FAS00/Sc was restored to the level of that of the positive con-
trol siRNA (ASc/Sc), leading to an approximate 2.4-fold de-
crease. However, duplex FAS06/Sc (bearing a caging group in
the seed region), duplex FAS10/Sc (bearing a caging group at
the cleavage site), and duplex FAS16/Sc (bearing a caging
group in dissociation region) exhibited similar levels of firefly
luciferase activity either with or without light irradiation. Fur-
ther evaluation of phosphorylated siRNAs (FAS06P, FAS10P, and
FAS16P) paired with the corresponding natural sense strand
also showed that no improvement of their RNAi activities was
observable (Figure S6). These results indicated that the in-
crease in siRNA stability with 2’-fluoro-modification had no
effect on the photomodulation of phosphate-caged siRNA
gene silencing activity.
Hydrolysis of the caging group in phosphate-caged siRNAs
and molecular dynamics simulation
The work described above shows that a single caged phos-
phate at the seed region and cleavage site is not sufficient to
block RNAi activity. Moreover, their failure to block siRNA activ-
ity is not due to the chemical or enzymatic instability of modi-
fied siRNAs themselves, as evidenced by experiments de-
scribed above (Figure 6 and Figure S6–8). The groups of Fried-
man and Monroe previously speculated that the caging group
on statistically labeled siRNA may be unstable in cellular envi-
Improving the nuclease resistance of siRNA is another im-
portant issue for therapeutic applications of synthetic
[
21]
siRNAs. The native siRNA sequence we used has previous
been shown to be highly stable towards enzyme-catalyzed
[
15]
degradation. To establish whether phosphate-caged modifi-
cation of siRNAs can decrease the nuclease resistance com-
pared with native siRNA, the susceptibility of siRNAs to snake
venom phosphodiesterase (SVPD), was first examined
[9,10]
ronments.
In our case, we introduced single caged phos-
phate group site-specifically in a RNA oligonucleotide and ob-
tained the pure caged antisense or sense RNA strand instead
of previous statistically distributed phosphate caged siRNA,
which facilitated a further investigation of detailed hydrolysis
mechanisms.
(Figure 6). Unmodified ASc/Sc and modified caged siRNA du-
plexes (AS00/Sc, AS21/Sc, AS22/Sc, AS23/Sc, AS24/Sc) with dif-
ferent numbers of caging groups, were incubated at 378C in
the presence of SVPD. After 1 h incubation, the hydrolyzed
band for the unmodified siRNA duplex was observed, however,
the hydrolyzed band for the modified siRNA duplexes was not
clearly visible. Thus, modified siRNA duplexes are generally
more stable than native siRNA duplexes in the presence of
SVPD. In addition, the tolerance of the siRNAs to Ribonuclea-
When the caged antisense RNA was mixed with the control
sense strand in standard PBS buffer at 378C, the siRNA duplex
(AS00/Sc and AS06/Sc) was formed. An identical experiment
was also carried out for only these caged antisense stands
(AS00 and AS06). Aliquots of the latter RNA solutions were re-
moved at recorded time points (0, 6, 14, and 24 h) and then
subjected to HPLC analysis. As shown in Figure 7, no caging
group was hydrolyzed for 5’-terminally phosphate-caged siRNA
duplex (AS00/Sc) even with long incubation time (24 h). How-
ever, hydrolyzed product, which was confirmed to be the cor-
responding siRNA duplex without the caging group (ASc/Sc),
was detected for siRNA duplex AS06/Sc even without light irra-
diation. The degree of caging group hydrolysis was different,
depending on the modified position in the RNA strand. More-
over, caging groups on the internal phosphate were more
easily hydrolyzed than caging groups close to the termini of
the siRNAs (Figure S9). However, no hydrolyzed product was
detected for any of the single-stranded RNAs, even after 24 h
incubation under the same conditions, even though they have
the same phosphotriester moiety. These results suggest that
hydrolysis of the caging group was not dependent on the cel-
lular environment, and that the internally phosphate-caged
siRNAs, but not the terminally phosphate-caged siRNA or
caged single-stranded RNAs, were more susceptible to hydroly-
Figure 6. Native PAGE of modified siRNA duplexes and control siRNA duplex
hydrolyzed by SVPD. The experimental conditions are as described in the Ex-
perimental Section.
&
&
Chem. Eur. J. 2014, 20, 1 – 10
6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!