Cellular activity of siRNA oligonucleotides containing synthetic isomorphic nucleoside surrogates

Article abstract of DOI:10.1039/c4cc08809c

Singly and multiply modified synthetic siRNA oligonucleotides, containing isomorphic surrogate nucleobases, show high interference potency in cell culture, suggesting the highly isomorphic RNA alphabet, based on a thieno[3,4-d]-pyrimidine core, is tolerat

Full text of DOI:10.1039/c4cc08809c

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Cellular activity of siRNA oligonucleotides
containing synthetic isomorphic nucleoside
surrogates†
Cite this: Chem. Commun., 2015,
51, 1662
Received 5th November 2014,
Accepted 5th December 2014
Dongwon Shin,a Peter Lonn,b Steven F. Dowdyb and Yitzhak Tor*a
¨
DOI: 10.1039/c4cc08809c
www.rsc.org/chemcomm
Singly and multiply modified synthetic siRNA oligonucleotides,
containing isomorphic surrogate nucleobases, show high interfer-
ence potency in cell culture, suggesting the highly isomorphic RNA
alphabet, based on a thieno[3,4-d]-pyrimidine core, is tolerated
well by the cellular silencing machinery.
Fig. 2 Isomorphic RNA alphabet.
RNA interference (RNAi), a regulatory process induced by short
interfering RNA (siRNA), is a powerful process capable of altering
cellular phenotypes,1–6 deciphering genetic pathways7,8 and identi- activity, selectivity, stability and potency of therapeutically-relevant
fying new therapeutic targets.9–11 Short double stranded RNA siRNAs.
oligonucleotides are either generated endogenously by Dicer from
Numerous backbone and ribose modifications have been exam-
longer precursors, or delivered exogenously. The passenger strand ined in siRNA.13–17 Much less has been done, however, with
gets degraded, while the guide strand, loaded onto the RNA-induced nucleobase surrogates capable of closely mimicking the native
silencing complex (RISC), guides the degradation of matched heterocycles.18–20 Such modified isomorphic nucleosides have the
homologous mRNA sequences at a specific site (Fig. 1). Related potential to impact the cellular interference activity and to serve as
pathways mediated by micro- (miRNAs) or dicer-independent RNAs built-in probes, potentially assisting in correlating biophysical
(piRNAs) have also been described.12 The exquisite selectivity for all properties with biological activity. Here we systematically explore
human mRNA targets has made RNAi the focus of intense research. the cellular silencing activity and biophysical properties of eighteen
Its promise to facilitate personalized medicine and to keep pace siRNA oligonucleotides, modified with a recently introduced iso-
with evolving pandemics has made this technology vital for future morphic RNA alphabet (Fig. 2).21 We modify the guide strand at
therapeutic development. This potential has prompted the explora- 13 out of the 21 possible nucleotide positions, including the seed
tion of diverse chemical modifications with the hope of altering the region, investigating the effect of individual as well as multiple
modifications on the stability of the guide/passenger duplex and
on the in vitro interference potency. Our results illustrate that this
alternative alphabet is well tolerated biophysically and cellularly,
with oligonucleotides containing single substitutions displaying
the same (or enhanced) activity compared to the native unmodified
siRNAs. Multiple modifications were also well-accommodated,
even with multiple base substitutions within the critical seed
region. Importantly, a qualitative correlation is seen between the
thermodynamic stability of the guide/passenger duplex and cellular
Fig. 1 Basic features of siRNA-mediated gene silencing.
interference activity.
We have recently reported the preparation of an alternative RNA
a Department of Chemistry and Biochemistry, University of California, San Diego,
alphabet, where all nucleosides were derived from thieno[3,4-d]-
pyrimidine as the common heterocyclic nucleus (Fig. 2).21 Struc-
La Jolla, CA 92093-0358, USA. E-mail: ytor@ucsd.edu
b Cellular & Molecular Medicine, University of California, San Diego, La Jolla,
tural analysis of their crystal structures suggested high level of
CA 92093-0686, USA
‘‘isomorphicity’’. While some perturbations to the D-ribose
pucker were seen, all modified ribonucleosides displayed
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4cc08809c
1662 | Chem. Commun., 2015, 51, 1662--1665
This journal is ©The Royal Society of Chemistry 2015

View Article Online
Communication
ChemComm
an anti orientation at their glycosidic linkage, similar to the Additionally, substitutions were made to explore their impact in
preference seen with the native nucleosides.22 A more critical test, distinct domains within the 21-mer siRNA guide strand, including
assessing the potential utility of this alternative RNA alphabet, residues 2–8 at the 50-end, known as the seed region, and substitu-
would be to evaluate their impact on the biophysical features and tions neighboring or opposite the RISC induced cleavage site.
function of biologically relevant oligonucleotides. Although likely
The modified 21-mer RNA oligonucleotides were all prepared
to be context, system and sequence dependent, we felt rigorously using solid-phase synthesis (Table 1). The required protected
evaluating their in vitro RNA interference activity in cell culture phosphoramidites of thG, thA, thU and thC have been synthesized
would provide useful insight into their ability to properly replace as shown in Scheme 1 relying, in the case of the purine analogs th
G
the native nucleobases in a functionally-demanding context and and thA (whose synthesis is more demanding), on the simultaneous
perhaps inspire future utilizations.
protection of the 30 and 50 hydroxyls using di-tert-butylsiloxane to
To explore the compatibility of the isosteric RNA alphabet facilitate selective TBDMS protection of the 20-hydroxyl.24 Following
shown in Fig. 2 with the cellular interference machinery, an completion of the solid phase synthesis, the oligonucleotides were
established siRNA sequence, known to target a destabilized deprotected and purified by PAGE. They were further characterized
green fluorescent protein (dGFP) in human H1299 lung adeno- by MALDI-TOF mass spectrometry (Table S1, ESI†).
carcinoma reporter cell line,23 was selected. This assay has
Prior to evaluating the cellular activity of the modified siRNAs,
proven to be a sensitive and robust platform for evaluating their duplex stability was assessed. The effect of base substitutions
cellular siRNA activity.7,8 Following the selection of a reporter on the thermal stability of the resulting siRNA duplexes formed
assay, which identifies a specific siRNA sequence, specific upon hybridization to the corresponding passenger strand was first
modification positions needed to be selected. Although all examined and compared to the unmodified siRNA duplex 1ꢀ2
positions can, in principle, be modified, we have identified (Table 1). Singly modified siRNAs, incorporating the purine mimics
selected positions to challenge the alternative alphabet. Speci- thA and thG, appear to be more stable than their unmodified
fically, modifications have been introduced into the seed region natural counterpart, where duplex stabilization of about +0.5 1C
as well as opposite the cleavage site.
per modification is observed at pH 7.0 (20 mM sodium cacodylate,
Eighteen single stranded oligonucleotides, 5–22, containing 20 mM NaCl). The single exception was the modification in
thieno[3,4-d]pyrimidine-based nucleosides replacing their natural position G4, which was slightly destabilizing by ꢁ0.5 1C. Duplex
counterparts at a single or multiple substitutions in the guide siRNA oligonucleotides, containing the pyrimidine mimics thU and
strand, were prepared (Table 1). The new nucleobases were strate- thC, were destabilized by ꢁ0.4 to ꢁ2.9 1C per base modification,
gically placed to explore the impact of the surrogate nucleobases on when compared to the native duplex. Multiple incorporation of
positions naturally occupied by both pyrimidines and purines. pyrimidine mimics yielded siRNA oligonucleotides, which were
Table 1 siRNA Duplexes studied. All siRNA, excluding Luc2 (see below), have a native sugar phosphate backbone
Sequencea
b
c
No.
Duplex
WTd
Tm (1C)
DTm
1
2
P
G
74.8
ND f
0.0
3
Luc2e
P
ND
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G4
G7
74.3
75.4
75.3
75.3
75.8
75.8
71.8
74.4
73.9
71.9
73.0
71.8
72.4
73.9
72.9
75.3
75.3
77.9
ꢁ0.5
+0.6
+0.5
+0.5
+1.0
+1.0
ꢁ3.0
ꢁ0.4
ꢁ1.1
ꢁ2.9
ꢁ1.8
ꢁ3.0
ꢁ2.4
ꢁ1.1
ꢁ1.9
+0.5
+0.5
+3.1
G13
G16
G19
G4ꢀ5
5Gs
U2
U6
U9
U14
U2ꢀ6
C8
C10
C8ꢀ10
A11
A15
A11ꢀ15
a
b
Passenger strands are on top and guide strands below. Tm values were determined at duplex concentrations of 1 mM in 20 mM sodium
c
cacodylate, pH 7.0, 20 mM NaCl, and 0.5 mM EDTA. Errors of at least two independent measurements were less than 0.8 1C. Difference in Tm
d
e
compared to 1ꢀ2. Wild type. Luciferase2 siRNA served as a negative control and had a modified backbone (20-hydroxyl groups were fully
f
modified as OMe for purines and replaced with F for pyrimidines). Not determined.
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun., 2015, 51, 1662--1665 | 1663

View Article Online
ChemComm
Communication
Fig. 3 Gene expression of dGFP at 48 h for all siRNAs; Lipo = lipofect-
amine only; CTR = buffer; Luc2 = luciferase2; WT = wild type; grey = 1 nM,
red = 5 nM, blue = 10 nM, cyan = 50 nM.
points (for 24 and 72 hours) are provided in the ESI† (see Fig. S2).
Fig. 4 depicts the normalized interference activity with respect to
the native siRNA shown along the thermal stability of each duplex.
The following sections summarize the interference observed per
modification.
Interference results for siRNA duplexes containing guide
strand modifications with thG replacing the native G in posi-
tions 4, 7, 13, 16 and 19 are shown in Fig. 3. These modified
duplexes were nearly as effective as the WT sequence in all
concentrations, with the exception of the G4 substitution (1ꢀ5),
which showed slightly attenuated activity and G13 (1ꢀ7), which
consistently showed enhanced activity when compared to the
positive control (Fig. 4).
Scheme 1 Synthesis of thieno[3,4-d]pyrimidine phosphoramidites. Reagents
and conditions: (a) (i) di-t-BuSi(OTf)2, DMF, 0 1C; (ii) TBDMSCl, Im, 0 1C–RT,
68% for 2 steps; (iii) N,N-diisobutylformamidine dimethyl acetal, DMF, 87%;
(b) (i) HF-Py, DCM, 0 1C, 86%; (ii) DMTrCl, Py, 69%; (iii) 2-cyanoethyl N,N-
diisopropylchlorophosphoramidite, iPr2NEt, DCM, 0 1C–RT, 95%; (c) (i)
DMTrCl, Py, 76%; (ii) TOMCl, di-tert-butyltin dichloride, iPr2NEt, DCE,
80 1C, 19%; (d) 2-cyanoethyl N,N-diisopropylchlorophosphoramidite, iPr2-
NEt, DCM, 0 1C–RT, 94%; (e) (i) DMF–DMA, DMF, 89%; (ii) di-t-BuSi(OTf)2,
DMF, 0 1C–RT; (iii) TBDMSCl, Im, 66% for 2 steps; (f) (i) HF-Py, DCM, 0 1C,
92%; (ii) DMTrCl, Py, 0 1C, 95%; (iii) 2-cyanoethyl N,N-diisopropylchloro-
phosphoramidite, iPr2NEt, DCM, 0 1C–RT, 57%; (g) (i) N,N-diisobutyl-
formamidine dimethyl acetal, DMF, 85%; (ii) DMTrCl, Py, 74%; (iii) TOMCl,
di-tert-butyltin dichloride, iPr2NEt, DCE, 80 1C, 41%; (h) 2-cyanoethyl
N,N-diisopropylchlorophosphoramidite, iPr2NEt, DCM, 0 1C–RT, 81%.
This particular siRNA contains only two A residues at posi-
tions 11 and 15 in the guide strand. Their substitution for thA
yielded little effect on cellular mRNA degradation compared to
the unmodified WT for both singly and doubly modified
siRNAs (Fig. 3). This is of significance, since A11 is near the
cleavage site of the target mRNA, demonstrating unperturbed
function by a strategically modified siRNA.
Substituting U for thU at positions 2, 6, 9 and 14 in the
guide strands 12–15, respectively, was somewhat disruptive
destabilized by ꢁ1.5 1C and ꢁ1.0 1C per modification for U2ꢀ6 for RNAi activity (Fig. 3). In particular, incorporating thU at
(1ꢀ16) and C8ꢀ10 (1ꢀ19), respectively. In contrast, a duplex positions 2 and 9 (1ꢀ12 and 1ꢀ14, respectively), flanking the
incorporating two thA at positions A11 and A15 (1ꢀ22) was found seed region, were most detrimental (Fig. 4) with up to 25%
to be stabilized by +1.6 1C per modification, yielding a double attenuation in interference activity with 5 nM duplex concen-
stranded construct which appears more stable than the corres- tration at 48 hours. Interestingly, and in stark contrast to the
ponding singly modified duplexes with thA at positions A11 or behavior seen with residues 2 and 9, substituting U6 within
A15 (1ꢀ20 and 1ꢀ21, respectively).
the seed region with thU (1ꢀ13) retained gene silencing activity
siRNA-mediated interference studies were conducted in H1299 at levels close to the RNAi activity displayed by the WT
cells with a dGFP reporter gene as described.23 All modified siRNA oligonucleotide. Incorporating thU at position 14, toward the
duplexes containing thA, thC, thG and thU substitutions were exam- 30-end of the guide strand (siRNA 1ꢀ15) did not disrupt RNAi
ined and compared to the activity displayed by the unmodified RNA activity, with potency similar to that displayed by the siRNAs
(WT) and a luciferase2-specific siRNA (as a negative control).23
Although the deviation in thermal stability of all modified
containing the G mimic.
Since the siRNA used in this study has only three C residues
siRNAs appears to be relatively subtle when compared to the in the guide strand, with one of them at the 50-end, we examined
unmodified siRNA duplex, native polyacrylamide gels were run the replacement of the internal residues only. Remarkably, the
before all transfection experiments (see Fig. S1, ESI†). The cells siRNA oligonucleotide (17) where C at position 8 within the seed
were transfected using lipofectamine with 1–50 nM of dsRNA in region was substituted by thC, exhibited better gene knockdown
the cell media. Interference was evaluated by measuring the than that displayed by the WT. Incorporating thC at position 10
level of GFP expression at 24, 48 and 72 h using flow cytometry. had minimal impact on silencing activity, with potency similar
Fig. 3 shows the data corresponding to 48 h and additional time to that of the unmodified siRNA.
1664 | Chem. Commun., 2015, 51, 1662--1665
This journal is ©The Royal Society of Chemistry 2015

View Article Online
Communication
ChemComm
Fig. 4 Gene knockdown and duplex stability compared to WT; black square ’ = 5 nM, red dot
= 10 nM, blue inverted triangle = Tm.
In addition to the singly substituted siRNAs discussed Nevertheless, the overall excellent performance of highly mod-
above, we also investigated the impact of multiple modifica- ified letters, where a thiophene heterocycle replaces the native
tions in the guide strand on the ability of the resulting siRNAs imidazole, coupled to additional recent observations25–27 suggest
to inhibit protein expression. Two substitutions with each one great utility for such isomorphic RNA building blocks.
of the nucleoside analogs thC, thG, thU and thA were explored and
We thank the NIH (grant No. GM 069773 to YT) and the
one challenging case with five incorporations of thG, was exam- W.M. Keck Foundation (to SFD and YT) for support.
ined. Substitutions at positions 2 + 6 with thU, as well as at
positions 4 + 5 and 4 + 7 + 13 + 16 + 19 with thG (oligonucleotides
16 and 10–11, respectively) were somewhat detrimental to RNAi
Notes and references
1 G. J. Hannon, Nature, 2002, 418, 244.
activity (Fig. 4). siRNAs with double incorporation of thC and th
A
2 C. C. Mello and D. Conte, Nature, 2004, 431, 338.
3 G. Meister and T. Tuschl, Nature, 2004, 431, 343.
4 C. C. Mello, Angew. Chem., Int. Ed., 2007, 46, 6985.
5 A. Z. Fire, Angew. Chem., Int. Ed., 2007, 46, 6967.
6 J. Kurreck, Angew. Chem., Int. Ed., 2009, 48, 1378.
7 A. E. Carpenter and D. M. Sabatini, Nat. Rev. Genet., 2004, 5, 11.
8 J. Moffat and D. M. Sabatini, Nat. Rev. Mol. Cell Biol., 2006, 7, 177.
9 Y. Dorsett and T. Tuschl, Nat. Rev. Drug Discovery, 2004, 3, 318.
at positions 8 + 10 and 11 + 15 (1ꢀ19, 1ꢀ22, respectively), were
nearly as effective as the WT oligonucleotide. Interestingly, the
oligonucleotide with thC at positions 8 and 10 (19) exhibited
slightly better activity that the native siRNA in all concentrations.
In summary, all modified siRNA duplexes containing sub-
stitutions of the native nucleosides with thA, thC, thG and thU, 10 C. J. Echeverri and N. Perrimon, Nat. Rev. Genet., 2006, 7, 373.
synthetic alphabet letters, were found to display potent cellular
11 P. H. E. Johnson, ‘‘RNA Interference, Application to Drug Discovery and
Challenges to Pharmaceutical Development’’, John Wiley & Sons, NJ,
interference activity. In general, replacing native pyrimidines
2011, p. 307.
with their synthetic surrogates, thC and thU, appears to be 12 C. Ender and G. Meister, J. Cell Sci., 2010, 123, 1819.
13 M. A. Behlke, Oligonucleotides, 2008, 18, 305.
14 J. K. Watts, G. F. Deleavey and M. J. Damha, Drug Discovery Today,
slightly more disruptive when compared to the purine replace-
ments. This is not entirely unexpected, due to the sterically
2008, 13, 842.
more demanding fusion of the thiophene at the pyrimidine’s 15 M. Terrazas and E. T. Kool, Nucleic Acids Res., 2009, 37, 346.
16 E. L. Chernolovskaya and M. A. Zenkova, Curr. Opin. Mol. Ther.,
5,6-positions. Multiple incorporations of modified letters, while
overall more detrimental, resulted in highly modified siRNA
2010, 12, 158.
17 J. B. Bramsen, et al., Nucleic Acids Res., 2009, 37, 2867.
duplexes with respectable interference activity. Even with the 18 H. Peacock, R. V. Fucini, P. Jayalath, J. M. Ibarra-Soza, H. J.
Haringsma, W. M. Flanagan, A. Willingham and P. A. Beal, J. Am.
Chem. Soc., 2011, 133, 9200.
19 A. S. Wahba, F. Azizi, G. F. Deleavey, C. Brown, F. Robert, M. Carrier,
poorest performance, seen for duplexes containing two mod-
ifications within the seed region, interference activity of above
50% (compared to the wild type activity) was seen at all concentra-
tions. While not without exceptions, the interference activity followed
the relative thermal stability of the siRNA duplexes. Although duplex
A. Kalota, A. M. Gewirtz, J. Pelletier, R. H. E. Hudson and M. J. Damha,
ACS Chem. Biol., 2011, 6, 912.
20 H. Peacock, A. Kannan, P. A. Beal and C. J. Burrows, J. Org. Chem.,
2011, 76, 7295.
stability alone cannot necessarily serve as a reliable ‘‘high resolution’’ 21 D. Shin, R. W. Sinkeldam and Y. Tor, J. Am. Chem. Soc., 2011, 133, 14912.
22 Theoretical support for their high similarity to the native nucleo-
predictor for cellular potency, this suggests that disrupting the
conformational and solvation integrity of the siRNA duplex at certain
bases has been reported; P. K. Samanta, A. K. Manna and S. K. Pati,
J. Phys. Chem. B, 2012, 116, 7618.
key domains is likely to negatively impact its interaction with the 23 A. Eguchi, B. R. Meade, Y. C. Chang, C. T. Fredrickson, K. Willert,
N. Puri and S. F. Dowdy, Nat. Biotechnol., 2009, 27, 567.
interference machinery.
24 For the synthesis of thU, see: S. G. Srivatsan, H. Weizman and Y. Tor,
We recognize the broad spectrum of RNA interference activities
Org. Biomol. Chem., 2008, 6, 1334.
and their dependency on numerous factors, including transfection 25 W. Liu, D. Shin, Y. Tor and B. S. Cooperman, ACS Chem. Biol., 2013,
8, 2017.
efficiency, duplex stability, and recognition by the required
proteins, etc. Hence, some of the observations made above might
26 R. W. Sinkeldam, L. S. McCoy, D. Shin and Y. Tor, Angew. Chem.,
Int. Ed., 2013, 52, 14026.
be somewhat specific to this particular dGFP-specific siRNA. 27 L. S. McCoy, D. Shin and Y. Tor, J. Am. Chem. Soc., 2014, 136, 15176.
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun., 2015, 51, 1662--1665 | 1665