Chirality Dependent Potency Enhancement and Structural Impact of Glycol Nucleic Acid Modification on siRNA

Article abstract of DOI:10.1021/jacs.7b02694

Here we report the investigation of glycol nucleic acid (GNA), an acyclic nucleic acid analogue, as a modification of siRNA duplexes. We evaluated the impact of (S)- or (R)-GNA nucleotide incorporation on RNA duplex structure by determining three individu

Full text of DOI:10.1021/jacs.7b02694

Article
Chirality Dependent Potency Enhancement and Structural
Impact of Glycol Nucleic Acid Modification on siRNA
Mark K Schlegel, Donald J. Foster, Alexander V Kel'in, Ivan Zlatev, Anna Bisbe,
Muthusamy Jayaraman, Jeremy G. Lackey, Kallanthottathil G Rajeev, Klaus Charisse,
Joel Harp, Pradeep S. Pallan, Martin A. Maier, Martin Egli, and Muthiah Manoharan
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 01 Jun 2017
Downloaded from http://pubs.acs.org on June 2, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
Chirality Dependent Potency Enhancement and Structural Impact of Glycol
Nucleic Acid Modification on siRNA
7
8
9
Mark K. Schlegel1*, Donald J. Foster1, Alexander V. Kel’in1, Ivan Zlatev1, Anna Bisbe1,
Muthusamy Jayaraman1, Jeremy G. Lackey1#, Kallanthottathil G. Rajeev1, Klaus Charissé1, Joel
Harp2, Pradeep S. Pallan2, Martin A. Maier1, Martin Egli2 and Muthiah Manoharan1*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1Alnylam Pharmaceuticals, 300 Third Street, Cambridge, Massachusetts 02142
2Vanderbilt University School of Medicine, Department of Biochemistry, Nashville, Tennessee
37232
#present address: Quantum-Si Inc, 530 Old Whitfield Street, Guilford, Connecticut, 06437
Abstract
Here we report the investigation of glycol nucleic acid (GNA), an acyclic nucleic acid
analog, as a modification of siRNA duplexes. We evaluated the impact of (S)ꢀ or (R)ꢀGNA
nucleotide incorporation on RNA duplex structure by determining three individual crystal
structures. These structures indicate that the (S)ꢀnucleotide backbone adopts a conformation that
has little impact on the overall duplex structure, while the (R)ꢀnucleotide disrupts the phosphate
backbone and hydrogen bonding of an adjacent base pair. In addition, the GNAꢀT nucleobase
adopts a rotated conformation in which the 5ꢀmethyl group points into the minor groove, rather
than the major groove in a normal WatsonꢀCrick base pair. This observation of reverse Watsonꢀ
Crick base pairing is further supported by thermal melting analysis of GNAꢀC and GNAꢀG
containing duplexes where it was demonstrated that a higher thermal stability was associated
with isoguanine and isocytosine base pairing, respectively, over the canonical nucleobases.
Furthermore, it was also shown that GNA nucleotide or dinucleotide incorporation increases
resistance against snake venom phosphodiesterase. Consistent with the structural data,
modification of an siRNA with (S)ꢀGNA resulted in greater in vitro potencies over identical
sequences containing (R)ꢀGNA. A walk of (S)ꢀGNA along the guide and passenger strands of a
conjugate duplex targeting mouse transthyretin (TTR) indicated that GNA is well tolerated in the
seed region of both strands in vitro, resulting in an approximate 2ꢀfold improvement in potency.
Finally, these conjugate duplexes modified with GNA were capable of maintaining in vivo
potency when subcutaneously injected into mice.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 28
1
2
3
4
Introduction
5
6
7
8
RNA interference (RNAi) is the endogenous biological process through which doubleꢀ
stranded small interfering RNA (siRNA) mediates sequenceꢀspecific gene silencing of target
mRNA.1ꢀ4 Recognition of the intended mRNA target is accomplished after the siRNA duplex is
recognized and bound by the endonuclease protein Argonaute 2 (Ago2), the catalytic component
within the RNAꢀinduced silencing complex (RISC). Once bound, the passenger strand (also
referred to as the sense strand) is specifically cleaved between positions 10 and 11 and
subsequently dissociates from RISC.3ꢀ5 The guide strand (also referred to as the antisense strand)
remains in the active RISC complex whereupon it can recognize and cleave the target mRNA.
While exogenous synthetic siRNAs also have the capability to induce geneꢀsilencing via the
endogenous RNAi pathway, they must include various chemical modifications in order to
stabilize these molecules against nuclease degradation, reduce their immunostimulatory
potential, and facilitate their uptake into cells.6 For example, strategically placed 2’ꢀribose
modifications and conjugation to a targeting triꢀNꢀacetylgalactosamine (GalNAc) ligand,
specifically recognized by the asialoglycoprotein receptor, have resulted in siRNAs which are
currently being evaluated in preꢀclinical and clinical studies for therapeutic silencing of genes in
hepatocytes.7ꢀ10
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
With regards to chemical modifications of siRNAs, it has been demonstrated that acyclic,
thermally destabilizing modifications have the capability to improve stability and/or enhance
RNAiꢀmediated gene silencing via several different mechanisms.11ꢀ14
For instance,
differentiation between the two strands of an siRNA that is loaded within Argonaute strongly
depends on the thermal stability of the terminal regions of the duplex; Ago2 favors loading of the
5’ꢀend of the strand from the duplex end which has a lower thermal stability.15ꢀ18 It has been
shown that thermally destabilizing modifications incorporated at certain positions of an siRNA
duplex can lead to an increase in potency by improving strand bias, thereby favoring loading of
the desired guide strand into RISC. In addition, the introduction of thermally destabilizing
modifications may help facilitate sense strand dissociation during Ago2 loading, thereby
increasing the rate of formation of active RISC.3
Along these lines, we wanted to investigate the potential of glycol nucleic acid (GNA) as
a modification within siRNA (Figure 1). Previous work has demonstrated that the (S)ꢀisomer of
GNA, comprised of a flexible acyclic, threeꢀcarbon backbone with only one stereocenter,
ACS Paragon Plus Environment

Page 3 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
crosspairs with RNA in an A/T rich context, whereas the (R)ꢀisomer does not.19,20 Analysis of
the crystal structures of several GNA homoduplexes have shown that (S)ꢀGNA forms a rightꢀ
handed duplex with helical parameters that resemble the Aꢀform, whereas (R)ꢀGNA would
presumably form a leftꢀhanded duplex.21ꢀ23
Although it has been demonstrated that the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
incorporation of GNA nucleotides is thermally destabilizing in a DNA duplex20, it is unclear as
to what extent the incorporation of GNA affects the thermal and/or metabolic stability of an
RNA duplex. Herein we report the impact of RNA nucleotide or base pair substitution using
both stereoisomers of GNA on in vitro and in vivo siRNA potency, duplex thermal stability,
resistance to degradation by snake venom phosphodiesterase (SVPD), and RNA duplex
structure.
Figure 1. Structural overview of an siRNAꢀGalNAc conjugate duplex.7 Also shown are the
structures of the nucleotides used in this study, including 2’ꢀOꢀmethyl (2’ꢀOMe), 2’ꢀdeoxyꢀ2’ꢀ
fluoro (2’ꢀF), and GNA. Backbone modifications include both natural phosphate (PO) and
phosphorothioate (PS) linkages.
Results
Thermal Stability of GNA/RNA Heteroduplexes. The impact of GNA incorporation on
RNA duplex thermal stability was measured in a series of 12ꢀmer RNA duplexes. Single
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 28
1
2
3
4
5
6
7
8
9
substitution of an RNA nucleotide with GNA resulted in an average destabilization of 10.3 °C
and 11.6 °C for the (S)ꢀ and (R)ꢀGNA isomers, respectively (Table 1). The difference in stability
between the two isomers was specific to the identity of the nucleobase; substitution of an A, T,
or C nucleotide resulted in more stable duplexes for the (S)ꢀisomer compared to the (R)ꢀisomer,
whereas G nucleotide substitution with the (R)ꢀisomer resulted in a slightly more stable duplex
compared to the (S)ꢀisomer (Table 1). In addition, the impact of GNA substitution on stability
was much greater within the context of G and C nucleotides (from 12ꢀ18 °C) compared to A or T
nucleotides (from 5ꢀ9 °C). These results are consistent with previous reports by Meggers and
colleagues who observed a similar discrepancy between A:T and G:C base pairs within
RNA:GNA heteroduplexes.20 Replacement of a single RNA base pair with GNA indicated little
difference between the two isomers in the A:T base pairs, but a clear preference for the (S)ꢀ
isomer in a G:C base pair (Table 1) with a ꢁꢁTm of 7.5 °C.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Impact of GNA substitution on thermal stability of duplexes.
Sequence Context
5’ꢀUACAGUCUAUGUꢀ3’
ꢀTm (°C)
3’ꢀAUGUCAGAUACAꢀ5’
(S)ꢁisomer
(R)ꢁisomer
Single substitution
ꢀ10.3 ± 5.3
ꢀ7.7
ꢀ11.6 ± 4.7
ꢀ9.1
A
T
ꢀ4.7
ꢀ7.6
G
C
ꢀ12.1
ꢀ16.8
ꢀ11.6
ꢀ18.2
Base pair substitution
ꢀ17.7 ± 3.4
ꢀ15.3
ꢀ21.0 ± 9.3
ꢀ14.4
A:T
G:C
ꢀ20.1
ꢀ27.6
The underlined RNA nucleotides were substituted with the indicated GNA
nucleotide or base pair. All values are the average of two independent measurements
at a duplex concentration of 2 µM in 1x PBS buffer (10 mM Na/K phosphate buffer,
pH 7.4, with 137 mM NaCl and 2.7 mM KCl). The values shown for single or base
pair substitution represent the average destabilization across all four nucleotides or
two base pairs, respectively.
Resistance to 3’ꢁexonuclease Degradation. Previous reports have demonstrated that the
incorporation of acyclic nucleotides at the 3’ꢀend of DNA oligonucleotides can enhance
resistance towards 3’ꢀexonuclease degradation.13 To assess the stability of GNAꢀmodified single
ACS Paragon Plus Environment

Page 5 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
stranded oligonucleotides towards 3’ꢀexonucleases, a series of strands containing single or
double GNA substitution at the 3’ꢀend of dT20 were analyzed in the presence of snake venom
phosphodiesterase (SVPD). These results were compared with that of a control containing a
single phosphorothioate linkage (PS) at the 3’ꢀend (halfꢀlife of 2.7 hours, Figure 2). The
oligonucleotides containing a single (S)ꢀ or (R)ꢀGNA nucleotide with a natural phosphodiester
linkage (PO) at the 3’ꢀend were more stable than the control (PS) with halfꢀlives of 5.6 and 8.0
hours, respectively (Figure 2A). Interestingly, the replacement of the PO linkage with PS
between the 3’ꢀterminal nucleotides had a consistent negative impact on nuclease resistance
resulting in halfꢀlives of only 2.1 and 2.7 hours for the (S)ꢀ and (R)ꢀisomer, respectively. When
the GNA nucleotide was moved to the penultimate position from the 3’ꢀend, nuclease resistance
was diminished in an all phosphate (PO) context, but greatly enhanced in the presence of PS to
20.1 and 19.6 hours for the (S)ꢀ and (R)ꢀisomer, respectively (Figure 2B). Finally, the
substitution of both 3’ꢀterminal nucleotides with GNA, connected through a PO linkage, resulted
in roughly an 8ꢀ or 5ꢀfold increase in nuclease resistance for the (S)ꢀ or (R)ꢀisomer, respectively
(Figure 2C). In contrast to the case of single substitution, the presence of a terminal PS linkage
between the GNAꢀGNA dinucleotide had only a minor effect on resistance to degradation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Identity Half-Life (hours)
Identity Half-Life (hours)
Identity Half-Life (hours)
TT
T•T
TT
< 1
2.7
5.6
8.0
2.1
2.7
TT
T•T
TT
< 1
2.7
TT
T•T
TT
< 1
2.7
0.7
20.5
13.6
27.5
14.5
Tt
tT
0.5
tt
T•T
T•t
T•T
t•T
20.1
19.6
T•T
t•t
Figure 2. Resistance of GNAꢀcontaining oligonucleotides toward degradation by snake venom
phosphodiesterase (SVPD) as a function of time. All dT20 oligonucleotides were modified by the
indicated dinucleotide on the 3’ꢀend which contained a single GNA nucleotide at the 3’ꢀend (A),
a single GNA at the penultimate nucleotide (B), or a GNA dinucleotide (C). Uppercase bold
underlined and lowercase bold underlined letters indicate (S)ꢀGNA and (R)ꢀGNA nucleotides,
respectively. Phosphorothioate linkages are indicated by the “●” symbol.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 28
1
2
3
4
5
6
7
8
9
XꢁRay Crystal Structures of GNA/RNA Heteroduplexes. To better understand the
implication of GNA modification on nucleic acid structure in the context of the observed thermal
and nucleolytic stability, we used Xꢀray crystallography to examine how well GNA can be
accommodated within an RNA duplex. We determined crystal structures of the selfꢀ
complementary RNA duplexes 5’ꢀr(CGCGAAU)ꢀTꢀ(AGCG)ꢀ3’ (SgnTꢀ12ꢀmer; T=(S)ꢀGNAꢀT;
G:A mismatch italicized), 5’ꢀr(CGCGAABrU)ꢀtꢀ(AGCG)ꢀ3’ (RgnTꢀ12ꢀmer; BrU=5ꢀBrꢀUridine;
t=(R)ꢀGNAꢀT; G:A mismatch italicized), and 5’ꢀr(BrCGAA)ꢀTꢀ(UCG)ꢀ3’ (SgnTꢀ8ꢀmer, BrC=5ꢀ
BrꢀCytidine; T=(S)ꢀGNAꢀT) at resolutions of between 1.08 and 1.20 Å (Table S10). All three
structures were phased by singleꢀwavelength anomalous dispersion (SAD), using either
strontium (SgnTꢀ12ꢀmer) or bromine (RgnTꢀ12ꢀmer, residue U7; SgnTꢀ8ꢀmer, residue C1) as the
anomalously scattering ion/atom. Residues are numbered 1 to 12 in the dodecamers and 1 to 8 in
the octamer. The asymmetric units of the SgnTꢀ12ꢀmer and ꢀ8ꢀmer crystals contain two
independent duplexes, whereas a single duplex comprises the asymmetric unit in the RgnTꢀ12ꢀ
mer crystal.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
These structures demonstrate that the GNA nucleotide is accommodated within the RNA
duplex in a variety of ways as far as backbone geometry, intraꢀstrand phosphateꢀphosphate
distances, and GNA conformation are concerned (Figure 3). In the SgnTꢀ12ꢀmer duplexes, two
of four (S)ꢀGNAꢀT residues per asymmetric unit exhibit twofold disorder, such that together with
the two (R)ꢀGNAꢀT residues in the RgnTꢀ12ꢀmer structure and the four (S)ꢀGNAꢀT residues in
the SgnTꢀ8ꢀmer structure, there are a total of 10 independent observations for the conformation
adopted by a GNAꢀT and its immediate environment in the three structures. In one of the
duplexes in the SgnTꢀ8ꢀmer structure, the GNAꢀT assumes an anti conformation (torsion around
the C2’ꢀC3’ bond) while the other GNAꢀT adopts the syn conformation (Figure 3B). In the other
duplex, the conformations of both GNA residues fall into the anti range. Interestingly, the more
limited distance that can be spanned by a GNA residue, in comparison to RNA which contains
an additional atom in its backbone, leads to a slight intrastrand stretch between phosphate groups
5’ꢀadjacent to the site of the glycol moiety. More or less regular Aꢀtype duplex spacing is
observed between 3’ꢀadjacent phosphates (Figure 3A). As expected, the anti conformation of
GNA is associated with a longer distance between 5’ꢀ and 3’ꢀintrastrand phosphate groups than a
…
syn conformation (Figure 3B), with P P distances in GNA residues ranging from 4.8 Å to 5.8 Å
in the three structures. The slight stretching of the RNA backbone on the 5’ꢀside of the GNA
ACS Paragon Plus Environment

Page 7 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
residue observed in the 8ꢀmer duplex is not as clearly evident in the 12ꢀmer duplexes. These
longer duplexes feature a G:A mismatch adjacent to the GNA residue that may also affect
backbone geometry. All ribose sugars that bracket GNAꢀT adopt a C3’ꢀendo (North)
conformation, with the exception of a single ribose in the SgnTꢀ12ꢀmer that exhibits a C2’ꢀendo
(South) pucker.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 28
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Crystal structures of RNA duplexes modified with both GNAꢀT stereoisomers.
…
(A) Variations in intrastrand P P distances in SgnTꢀ8ꢀmer RNA as a consequence of the
incorporated GNAꢀT residues (carbon atoms highlighted in green). (B) GNA nucleotides adopt
both gauche and anti conformations within the structures. (C) Example of an (S)ꢀGNAꢀT:RNAꢀ
A base pair showing a rotated nucleobase conformation for the GNA nucleotide (arrow). (D)
Global structures of the RNA duplexes incorporating both (S)ꢀ and (R)ꢀisomers of GNA which
highlight the phosphate backbones. The two isomers are accommodated differently within the
global RNA structure and result in a slight kink in the (R)ꢀisomerꢀcontaining duplex (arrow). (E)
ACS Paragon Plus Environment

Page 9 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
(R)ꢀGNAꢀT residues distort RNA duplex and pairing geometry to a greater extent than (S)ꢀGNAꢀ
T residues. Superimposition of A:U and G:A base pairs flanking (S)ꢀGNAꢀT(green):RNAꢀA and
(R)ꢀGNAꢀT(yellow):RNAꢀA in the two 12ꢀmer duplexes reveals a disruption of the neighboring
A:U pair in the (R)ꢀGNAꢀTꢀmodified 12ꢀmer (arrow). (F) An (S)ꢀGNAꢀT residue from the
crystal structure of the SgnTꢀ12ꢀmer can seamlessly and with optimal geometry replace an RNA
nucleotide at position 7 of the guide strand RNA bound to human Ago 2.24 The RNA strand
assumes a kink at that site that is associated with Ileꢀ365 and results in unstacking of bases of
nucleotides 6 and 7.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The most surprising conformational change, a shared feature of all GNA nucleotides in
the three structures, is the rotation of the GNA thymine nucleobase such that the C5ꢀmethyl
group of GNAꢀT points into the minor groove as opposed to the major groove in a standard
WatsonꢀCrick base pair (Figure 3C). For an A:T(U) pair, this change remains without
consequence as far as the number of Hꢀbonds (two) are concerned. However, as the
thermodynamic data demonstrate, G:C pairs involving GNA residues, unlike A:U pairs, are
poorly tolerated in RNA duplexes (Table 1). This difference may possibly be rationalized by the
crystallographic data, which would suggest that the rotation of the GNA cytosine base would
result in the loss of hydrogen bonding compared to a standard WatsonꢀCrick G:C pair. In order
to investigate this further, we paired GNAꢀC and GNAꢀG with isoguanosine and isocytidine
RNA nucleotides in our 12ꢀmer RNA duplex (Figure 4 and Table 2). Contrary to previous
reports25,26, the RNA duplex containing a central isoC:isoG base pair resulted in a slight
destabilization of 2.5 °C. While the incorporation of GNAꢀC against riboꢀG resulted in a
destabilization of 16.8 °C, switching the pairing nucleotide to riboꢀisoG resulted in a duplex
destabilized by only 11.5 °C, representing a 5.3 °C stabilization of the duplex over the GNAꢀC:G
pair. The incorporation of riboꢀisoC opposite GNAꢀG in the duplex had a similar stabilizing
effect over the C:GNAꢀG pair of 1.6 °C. This additional Tm data strongly supports the
hypothesis that, as suggested by the crystal structures, reverse WatsonꢀCrick base pairing is
common to GNA pyrimidine nucleotides in RNA duplexes, and helps explain the discrepancy
between A:T and G:C base pairs of (S)ꢀGNA with RNA.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 28
1
2
3
4
5
6
7
8
9
O
O
N
NH2
N
O
HN
O
N
N
N
N
N
N
NH
N
N
NH2 H2N
RNA-A:(S)-GNA-T
RNA-G:(S)-GNA-C
(observed)
(proposed)
Change to
isoC
Change to
isoG
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
H2N
NH2
NH
O
O
N
N
N
N
N
N
N
NH
NH2
N
O
N
N
H2N
GNA-G:RNA-isoC
RNA-isoG:GNA-C
NH2
NH
O
N
N
O
N
O
O
N
NH2
N
O
O
O
OH
O
P
OH
O
P
O
O
O
RNA-isoC
RNA-isoG
Figure 4. Structures of isocytidine and isoguanosine nucleotides and their potential to form fully
complementary base pairs to “rotated” GNAꢀC or GNAꢀG.
Table 2. Investigation of (S)ꢀGNA crosspairing with isoC and isoG RNA
nucleotides.
5’ꢀUACAGUXUAUGUꢀ3’
3’ꢀAUGUCAYAUACAꢀ5’
X:Y
Tm (°C) ꢀTm (°C)
X:Y
Tm (°C) ꢀTm (°C) ꢀꢀTm (°C)
C:G
C:isoG
C:G
51.2
48.5
34.4
39.2
0.0
isoC:isoG
isoC:G
48.7
35.8
39.7
40.7
ꢀ2.5
ꢀ15.4a
ꢀ11.5
ꢀ10.5
ꢀ
ꢀ
ꢀ2.7
ꢀ16.8
ꢀ12.1
C:isoG
isoC:G
+5.3
+1.6
C:G
The red indicated RNA nucleotides were substituted as specified in the table. Uppercase bold
underlined letters represent (S)ꢀGNA nucleotides. All values are the average of two
independent measurements at a duplex concentration of 2 µM in 1x PBS buffer (10 mM
Na/K phosphate buffer, pH 7.4, with 137 mM NaCl and 2.7 mM KCl). a) The magnitude of
destabilization in this base pair has been discussed before.25
ACS Paragon Plus Environment

Page 11 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Comparison of the SgnTꢀ and RgnTꢀ12ꢀmer structures provided insight into the
differential destabilization afforded by individual GNA residues when incorporated in RNA
duplexes. A closer look at each of the global duplex structures suggests that the (R)ꢀGNA
nucleotide introduces a kink within the phosphate backbone, whereas the (S)ꢀGNA nucleotide
appears to be well accommodated with only minimal distortion (Figure 3D). The increased
backbone distortion caused by the (R)ꢀGNA nucleotide is unsurprising considering that the (R)ꢀ
isomer presumably forms a leftꢀhanded duplex in contrast to the (S)ꢀisomer which has been
shown to form a rightꢀhanded structure in homoꢀduplexes of GNA.20,22,23 An overlay of baseꢀ
pair steps adjacent to (S)ꢀ and (R)ꢀGNAꢀT residues in the two 12ꢀmer duplexes reveals that the
latter results in a significant distortion of the 5’ꢀadjacent A:U pair that would be expected to
destabilize the duplex (Figure 3E). In general, the structural data support the conclusion that,
consistent with the thermal stability data, (R)ꢀGNAꢀT is less well accommodated in an RNA Aꢀ
form duplex than (S)ꢀGNAꢀT.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In Vitro Gene Silencing in Unmodified RNA. The impact of GNA modification and the
influence of C2’ stereochemistry on silencing was first assessed within the context of an
unmodified siRNA duplex targeting phosphatase and tensin homolog (PTEN) mRNA.
Modification of the center of the siRNA duplex with GNA resulted in both nucleobaseꢀ and
isomerꢀspecific effects (Figure 5). Whereas GNAꢀA and GNAꢀC were tolerated regardless of
their absolute configuration, GNAꢀG and GNAꢀT showed a strong stereoisomer dependence on
inherent potency (Figure 5A). In the cases where the C2’ stereochemistry had an impact on
potency, the (S)ꢀisomer was roughly 2ꢀ or 15ꢀfold more potent than the (R)ꢀisomer for G or T
substitution, respectively. We additionally investigated the influence of single GNA base pair
incorporation on siRNA potency using the same PTEN sequence (Figure 5B). The siRNA
containing a single (S)ꢀGNAꢀA:T base pair showed activity that was roughly 60ꢀfold greater than
the parent. Interestingly, this duplex was almost 10,000ꢀfold more potent that the corresponding
duplex containing an (R)ꢀGNA base pair. The C2’ configuration did not have a significant
impact on the potency of the duplexes containing a GNAꢀG:C base pair.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 28
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Effect of GNA incorporation on in vitro potency of unmodified siRNA targeting
PTEN. Either a single nucleotide (panel A) or a single base pair (panel B) in the center of the
duplex was substituted as indicated. See Supporting Information for sequence information.
In Vitro and In Vivo Gene Silencing in HeavilyꢁModified Conjugate Duplexes. Based
on our initial results which suggested that the (S)ꢀisomer of GNA is not only more easily
accommodated within the RNA duplex, but also results in, on average, greater potency than the
(R)ꢀisomer, we only chose to further investigate the (S)ꢀisomer within the context of a heavily
modified siRNA duplex targeting the rodent transthyretin gene (TTR).7,27 The parent siRNA,
which is fully modified with 2’ꢀOMe, 2’ꢀF, and strategically placed PS modifications, was
designed to provide an ideal balance between stability towards nucleolytic degradation and
productive association with Ago2.7,27,28 Since the use of these 2’ꢀmodifications results in a high
increase of thermal stability, one would presume that additional modification with GNA would
still provide a duplex with a higher thermal stability than natural RNA. To determine the
positional effect of GNA incorporation on thermal stability and silencing activity, we walked a
single (S)ꢀGNA nucleotide through both the guide and passenger strand of a 21/23ꢀmer duplex
targeting mouse TTR mRNA. The incorporation of (S)ꢀGNA led to a variation of thermal
stability ranging from +1 °C to ꢀ10 °C that was specific to both the position and nature of the
nucleobase (see Tables S5 and S6). As expected, the nature of the nucleobase had a large impact
on the observed ꢁTm values with incorporation of (S)ꢀGNAꢀG or (S)ꢀGNA–C being significantly
more destabilizing than either (S)ꢀGNAꢀA or (S)ꢀGNA–T.
ACS Paragon Plus Environment

Page 13 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
To evaluate the potential of this set of modified duplexes to silence TTR gene expression
in vitro, a twoꢀdose screen was performed in primary mouse hepatocytes (Figure 6). The
incorporation of (S)ꢀGNA nucleotides was generally well tolerated in both the guide and
passenger strands, with the exception of positions that have been previously shown to be
intolerant to modification.12,18,29 Those positions, 1ꢀ2 on the guide strand and 10ꢀ12 on the
passenger strand, are likely sensitive due to their importance in Ago2 association/loading and
sense strand cleavage/removal, respectively.24,30,31 In addition, loss of potency was observed
when GNA was incorporated towards the 3’ꢀend of the guide strand or the 5’ꢀend of the
passenger strand, respectively. The loss in silencing activity may have been the result of an
improper thermal balancing towards loading of the passenger strand. In contrast, positions 13ꢀ15
of the guide and 6ꢀ9 of the passenger strand maintained in vitro efficacy comparable to the
parent. Substitution of positions 3ꢀ9 of the guide strand (within the seed region) and positions
13ꢀ19 of the passenger strand were also well tolerated. Of particular interest are nucleotides 6ꢀ7
of the guide strand where it has been shown that an isoleucine side chain causes a kink in the
guide RNA structure complexed with Ago2.24,30 Modeling an (S)ꢀGNA nucleotide in position 7
of the guide strand demonstrated that GNA can efficiently replace the RNA nucleotide with a
geometry that appears optimal for interaction with Ago2 (Figure 3F).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 28
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Positional impact of single (S)ꢀGNA nucleotide substitution on in vitro silencing
activity at a concentration of 10 nM siRNA. The nucleotide at the indicated position of the guide
or passenger strand was substituted with the corresponding GNA nucleotide. See Table S4
(Supporting Information) for sequence information.
A subsequent study investigated the effect of (S)ꢀGNA base pair incorporation within the
same TTR siRNA duplex on in vitro silencing activity (Figure 7). Although base pair
substitution resulted in a loss of activity in many positions, we observed similar trends with
respect to the regions where the GNA base pair was tolerated. Those positions were within the
seed region of the duplex, mainly at positions 3 and 5ꢀ8 relative to the 5’ꢀend of the guide strand.
ACS Paragon Plus Environment

Page 15 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Positional impact of single (S)ꢀGNA base pair substitution on in vitro silencing
activity at a concentration of 10 nM siRNA. The base pair at the indicated position of the guide
or passenger strand was substituted with the corresponding GNA base pair. See Table S4
(Supporting Information) for sequence information.
Utilizing several of the siRNA duplexes which were promising in our in vitro assays
(Table 3), we evaluated the ability of (S)ꢀGNAꢀmodified siRNAꢀGalNAc conjugate duplexes to
silence TTR gene expression in vivo. After subꢀcutaneous administration of each siRNA at a
dose of 2.5 mg/kg in mice (approximate ED80 of the parent siRNA7), both liver mRNA and
serum TTR levels were monitored over a period of 21 days (Figure 8). Administration of the
parent duplex (D1) led to a 77% suppression of TTR mRNA in the liver and an 85% reduction in
circulating TTR protein at day 7. The administration of GNAꢀmodified duplexes resulted in
similar levels of silencing at day 7 with 84%, 87%, and 84% mRNA suppression for guide (D2),
passenger (D3), and baseꢀpair (D4) modified duplexes, respectively (Figure 8A). Knockdown
was also reflected in the levels of circulating TTR protein where a reduction of 85%, 84%, and
79% was observed at day 7 for duplexes D2, D3, and D4, respectively (Figure 8B). While the
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 28
1
2
3
4
5
6
7
8
9
efficacy of TTR suppression was comparable across the siRNAs, the duration of silencing
activity appeared to be somewhat dependent on the placement of the GNA nucleotides. Whereas
the duplex modified in the passenger strand with GNA (D3) exhibited levels of suppression of
TTR mRNA through day 21 comparable to the parent (D1), guide (D2) and base pair (D4)
modified duplexes exhibited less durable mRNA knockdown than duplex D3 (Figure 8A). A
similar trend was observed with regards to suppression of the TTR protein. Although the levels
of protein suppression between the duplexes differed little at day 21, the passenger modified
duplex D3 (similar to the parent D1) showed higher levels of protein suppression on day 14
(approximately 7 days post Nadir) compared to duplex D4 (Figure 8B).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3. siRNA duplexes targeting rodent TTR.
In Vitro IC50
siRNA
duplex
Passenger (5’ꢁ3’)
Tm (°C)
(pM)
Guide (3’ꢁ5’)
A
●
a
●
●
CaGuGuUCUuGcUcUaUaA(L)
D1
D2
D3
D4
20.6 ± 1.9
13.0 ± 1.5
11.9 ± 1.0
8.1 ± 0.8
75.2
70.9
70.3
66.3
u
●
u●
uUgUcAcAagaAcGaGaUa
●
U●u
A●
a
CaGuGuUCUuGcUcUaUaA(L)
u●
u
●
uUgUcAcAagaAcGAGaUa
CaGuGuUCUuGcTcUaUaA(L)
uUgUcAcAagaAcGaGaUa
CaGuGuUCUuGcTcUaUaA(L)
uUgUcAcAagaAcGAGaUa
●
U●
u
A
u
● ●
a
u●
●
●
U●u
A
● ●
a
u●
u●
●
U●u
Italicized uppercase, lowerꢀcase, and uppercase bold underlined letters represent 2’ꢀF, 2’ꢀOMe, and
(S)ꢀGNA sugar modifications, respectively to Adenosine, Cytosine, Guanosine, and Uridine. (L)
represents the triꢀNꢀacetylgalactosamine ligand.7 Phosphorothioate linkages are indicated by the “
symbol.
●”
ACS Paragon Plus Environment

Page 17 of 28
Journal of the American Chemical Society
1
2
3
4
5
B
Day 7
Day 14
Day 21
6
7
8
9
100
75
50
25
0
p=0.0444
p=0.0164
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Position Modified with (S)-GNA
Figure 8. Knockdown of TTR in vivo with (S)ꢀGNA modified siRNA duplexes dosed at 2.5
mg/kg. Error bars represent the SD from each cohort (n=3). Only those comparisons which are
statistically significant are shown in the graph; all others are nonꢀsignficant with the exception of
all comparisons to PBS which were all significant. (A) TTR mRNA levels measured in the liver.
(B) TTR protein levels measured in the serum. G = guide strand, P = Passenger strand.
Discussion
In this study we evaluated the positional effect of GNA on the biophysical and biological
properties of siRNA duplexes. As expected, the incorporation of GNA into RNA or RNAꢀlike
duplexes led to a significant thermal destabilization. The impact of nucleobase identity on the
degree of thermal destabilization may be partially explained via the crystal structures of RNA
duplexes containing a single GNAꢀT nucleotide. If the reverse WatsonꢀCrick pairing mode, a
result of a rotated nucleobase, exhibited by the GNA nucleotide in these structures can be
extended to the other three nucleobases, it would likely explain the significantly higher degree of
destabilization observed for G:C base pairs, where one or more of the three hydrogen bonds of
the base pair would be suppressed, vs. A:T base pairs for which the two natural hydrogen bonds
of the base pair would be preserved. This observation is supported by thermal melting
experiments which show that pairing GNAꢀC or GNAꢀG with isoguanosine or isocytidine,
respectively, can restore the degree of destabilization back to the level of A and T GNA
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 28
1
2
3
4
5
6
7
8
9
nucleotides. The crystal structures additionally demonstrate that the acyclic backbone allows
GNA to sample many different backbone conformations within the duplex structure. Of
particular interest to siRNA, the model of a GNA nucleotide incorporated in the seed region
suggests that GNA’s acyclic nature may provide a unique potency benefit by facilitating
recognition and cleavage/release of the mRNA target by Ago2.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The potential of any new nucleotide modification incorporated within an siRNAꢀ
conjugate duplex will strongly depend on its ability to resist degradation by nucleases. Since
chemical modification of an siRNA administered via subꢀcutaneous injection is necessary in
order to stabilize it against nucleolytic attack along the path to its ultimate destination,
modifications which introduce a metabolic instability may negatively impact the efficacy and
duration of effect of the corresponding siRNA. Our experiments have shown that GNA has the
ability to stabilize linkages against 3’ꢀexonuclease attack in an in vitro setting. The ability of a
GNA dinucleotide at the 3’ꢀend to create an oligonucleotide highly resistant to exonuclease
degradation in the absence of phosphorothioate linkages is especially intriguing.
An increasing body of evidence has demonstrated that the effect of chemical
modifications on siRNA duplexes is highly dependent on the specific position within the
duplex.29 Certain sensitive positions typically only tolerate 2’ꢀOH (natural RNA) or 2’ꢀF
modifications to maintain RNAi activity.32 Following this trend, GNA is not well tolerated at
positions 1 and 2 of the guide strand or in the central region of the passenger strand. In
agreement with previous studies investigating the thermal destabilization of siRNA duplexes,
GNA was most well tolerated within the seed region of the duplex. Interestingly, the proposed
mechanism of target recognition and cleavage by RISC involves an initial binding of the target
mRNA via the seed of the guide strand. Once the mRNA target is recognized, the structure of
Ago2 undergoes a shift in which the alpha helix (helixꢀ7) neighboring the guide strand in the
seed region is moved away, simultaneously causing a positional shift in the PAZ domain and
thereby allowing the guide strand to relax into a near Aꢀform conformation.31 Of particular
relevance to siRNA, the model of a GNA nucleotide incorporated in the seed region suggests that
GNA’s acyclic nature may provide a potency benefit by facilitating the conformational changes
in Ago2 upon recognition and cleavage/release of the mRNA target. Modification with (S)ꢀGNA
was additionally well tolerated in the 3’ꢀsupplemental region (guide positions 12ꢀ16) of the
duplex; a region proposed to have an additional role in the removal of the sense strand in the
ACS Paragon Plus Environment

Page 19 of 28
Journal of the American Chemical Society
1
2
absence of cleavage.33 For the siRNAs used in this study, the potency improvement of guideꢀ,
passengerꢀ, or baseꢀpair modified duplexes observed in vitro did not translate into a significant
activity benefit in vivo. Of the three GNAꢀmodified siRNAs, the passenger modified duplex D3
performed most similar to the parent in vivo, whereas the guide (D2) and baseꢀpair (D4)
modified duplexes trended towards lower potency and duration compared to the parent. The
difference in thermal stability observed between these four duplexes may help explain the
discrepancy between their in vitro and in vivo activity. Since the base pair modified duplex D4
trended towards the lowest in vivo activity, but the highest inherent potency in vitro, it is possible
that the large loss in Tm (8.9 °C) due to the incorporation of the GNA base pair (Table 3) may
result in the introduction of metabolic liability into the duplex, while the smaller decrease in Tm
observed with D2 (4.3 °C) and D3 (4.9 °C) relative to the parent D1 may not have a large impact
on the overall pharmacokinetic profile of those siRNAs.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Conclusions
We have shown that (S)ꢀGNA can be utilized as a modification within siRNA, thereby
expanding the toolbox of thermally destabilizing modifications. While both the (S)ꢀ and (R)ꢀ
isomers were effective towards the suppression of mRNA in vitro, we have shown that
incorporation of the (S)ꢀisomer results in a potency improvement over the (R)ꢀisomer. Crystal
structures and Tm data of RNA duplexes containing GNA nucleotides indicate that (R)ꢀisomer
incorporation, preferring a leftꢀhanded duplex, resulted in a larger perturbation of the overall
duplex structure and thereby a larger thermal destabilization. The in vitro and in vivo data, along
with our modeling studies, demonstrate that GNA is well accommodated in certain positions of
the duplex and therefore make GNA an attractive modification for siRNA. It should be noted,
however, that in order to be able to consider GNA, or any other novel modification, for
therapeutic oligonucleotides, it will need to demonstrate an appropriate preclinical safety profile
in rodents along with a substantial (at least 3ꢀfold) pharmacological benefit in higher species.
Future work will focus on additional structureꢀactivity and pharmacokinetics studies in order to
better understand the in vivo translation and to fully harness the potential of GNA within the
context of therapeutic siRNAs.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 28
1
2
3
4
5
6
7
Experimental Section
8
9
Synthesis of (S)- and (R)-GNA phosphoramidites. Racemic GNA nucleosides were first
synthesized by the groups of Ueda34 and Imoto35 and as enantiomerically pure compounds by the
group of Holý.36 The first GNA phosphoramidites and GNAꢀcontaining oligonucleotides where
synthesized by Cook et. al.37,38 and Wengel.39 Meggers and colleagues subsequently published
improved and simplified methods for the synthesis of GNA phosphoramidites and GNA
oligonucleotides.19,40,41 To further reduce the total number of synthetic steps for GNA
phosphoramidites, we developed a procedure which allows for the ringꢀopening of enantiopure
DMTꢀglycidol using protected purine nucleobases (Scheme 1). Whereas previous reports
mention failure in the direct alkylation of protected purines 3c and 3d, we found that the reaction
in the presence of N,Nꢀdiisopropylethylamine resulted in the desired products while preserving
the exocylic protecting groups on 3cꢀ3d. The alkylation of purines 3c and 3d, in contrast to
alkylation using unprotected purines in the presence of NaH,40,41 also resulted in the formation of
a significant amount of the N7 regioisomer which required separation from the N9 product via
silica gel chromatography. This method could additionally be applied to the pyrimidine
nucleobases 3aꢀ3b resulting in yields similar to previous reports utilizing NaH for deprotonation
of the nucleobase. Phosphoramidites of both (S)ꢀ and (R)ꢀisomers of GNA were synthesized
from these nucleosides using previously reported procedures.40,41
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Procedure for the synthesis of GNA phosphoramidites 6aꢁ6d and 7aꢁ7d.
ACS Paragon Plus Environment

Page 21 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Oligonucleotide synthesis. All oligonucleotides were prepared on an ABI 394 or MerMade 192
synthesizer on a 1 µmole scale using universal or custom supports. All phosphoramidites were
used at a concentration of 100 mM in 100% acetonitrile or 9:1 acetonitrile:DMF (2’ꢀOMeꢀC, 2’ꢀ
OMeꢀU, GNAꢀG) with a standard protocol for 2ꢀcyanoethyl phosphoramidites and ETT
activator, except that the coupling time was extended to 400 seconds. Phosphite oxidation to
phosphate or sulfurization to phosphorothioate was achieved using a solution of 50 mM I2 in 9:1
acetonitrile:water or 100 mM DDTT in 9:1 pyridine:acetonitrile, respectively.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
MerMade 192 workꢀup: After the tritylꢀoff synthesis using the MerMade 192, columns
were incubated with 150 µL of 40% aqueous methylamine for 30 minutes at room temperature
and the solution was drained via vacuum into a 96ꢀwell plate. After repeating the incubation and
draining with a fresh portion of aqueous methylamine, the plate containing crude
oligonucleotides solution was sealed and shaken at room temperature for an additional 60
minutes to completely remove all protecting groups. Precipitation of the crude oligonucleotides
was accomplished via the addition of 1.2 mL of 9:1 acetonitrile:EtOH to each well, followed by
incubation at ꢀ20 °C overnight. The plate was then centrifuged at 3000 RPM for 45 minutes at 4
°C, the supernatant removed from each well, and the pellets resuspended in 950 µL of 20 mM
aqueous NaOAc. Each crude solution was finally desalted over a GE HiꢀTrap desalting column
(Sephadex G25 Superfine) using water to elute the final oligonucleotide products. The identities
and purities of all oligonucleotides were confirmed using ESIꢀMS and IEX HPLC, respectively.
ABI 394 workꢀup: Standard protocols were used for cleavage and deprotection. Crude
oligonucleotides were purified using strong anion exchange and phosphate buffers (pH=8.5)
containing sodium bromide. The identities and purities of all oligonucleotides were confirmed
using ESIꢀLC/MS and IEX HPLC, respectively.
Thermal melting experiments. Melting studies were performed in 1 cm path length quartz cells
on a Beckman DU800 spectrophotometer equipped with a thermoprogrammer. Each cuvette
contained 200 µL of sample solution covered by 125 µL of light mineral oil. Samples were
initially annealed in the instrument by heating at a rate of 5 °C/min from 15ꢀ95 °C followed by
cooling at the same rate from 95ꢀ15 °C. After a waiting period of 5 minutes, melting curves
were monitored at 260 nm with a heating rate of 1 °C/min from 15ꢀ95 °C. Melting temperatures
(Tm) were calculated from the first derivatives of the heating curves and the reported values are
the result of at least two independent measurements.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 28
1
2
3
4
5
6
7
8
9
Nuclease stability assays. DNA oligonucleotide solutions were prepared at a concentration of 0.1
mg/mL in 50 mM Tris*HCl buffer (pH=8.0), containing 10 mM MgCl2. Snake venom
phosphodiesterase I (150 mU/mL) was added to the DNA mixture prior to the first injection and
each sample was analyzed for degradation over a time period of 24 hours at room temperature.
An analytical sample was removed from the reaction every hour and was eluted (1 mL/min) over
a Dionex DNAPac PA200 column (4 x 250 mm, 8 µm particle size) at 30 °C with buffer A (20
mM sodium phosphate buffer, 10% acetonitrile, pH = 11.0) and a linear gradient from 37ꢀ52% of
buffer B (20 mM sodium phosphate buffer, 1M NaBr, 10% acetonitrile, pH = 11.0) over 10
minutes. Integration of the peak corresponding to the full length product at 260 nm was
normalized to the area at time zero and used to calculate the percentage of full length product
remaining at each time point. The data points were plotted in Microsoft Excel and firstꢀorder
decay kinetics were used to determine the halfꢀlife, defined as the point where 50% of the full
length product was degraded.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Crystallization, X-ray data collection and processing, phasing and refinement. Crystals of
SgnTꢀ12ꢀmer were grown by the hangingꢀdrop vapor diffusion technique using the Nucleic Acid
Miniscreen (Hampton Research, Aliso Viejo, CA).42 Droplets (2 µL) containing oligonucleotide
(0.6 mM), sodium cacodylate (20 mM, pH=6.0), lithium chloride (20 mM), strontium chloride
(40 mM), magnesium chloride (10 mM), spermine tetrahydrochloride (6 mM), and 2ꢀmethylꢀ2,4ꢀ
pentanediol (MPD; 5% (v/v)) were equilibrated against a reservoir of MPD (1 mL, 35%). RgnTꢀ
12ꢀmer crystals were obtained from droplets (2 µL) containing oligonucleotide (0.6 mM),
sodium cacodylate (20 mM, pH=6.0), potassium chloride (40 mM), spermine tetrahydrochloride
(6 mM), and MPD (5% v/v) that were equilibrated against a reservoir of MPD (1 mL, 35%).
SgnTꢀ8ꢀmer crystals were obtained from drops (0.6 µL) containing oligonucleotide (0.5 mM),
sodium cacodylate (20 mM, pH=5.5), sodium chloride (40 mM), magnesium chloride (10 mM),
hexamine cobalt(III) chloride (10 mM), and MPD (5% v/v) that were equilibrated against a
reservoir of MPD (70 µL, 40%). All the crystals were mounted without further cryoꢀprotection
and flashꢀcooled in liquid nitrogen.
Xꢁray data collection, phasing and refinement. Diffraction data were collected on the 21ꢀIDꢀF
or 21ꢀIDꢀD beam lines of the Life Sciences Collaborative Access Team (LSꢀCAT) at the
Advanced Photon Source (APS), located at Argonne National Laboratory (Argonne, IL). For
ACS Paragon Plus Environment

Page 23 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
SgnTꢀ12ꢀmer crystals, native data sets were collected on 21ꢀIDꢀF, but we were not able to phase
the structure using molecular replacement. Therefore, anomalous data were collected at 0.769 Å
(strontium edge). For RgnTꢀ12ꢀmer crystals the wavelength was tuned to 0.91697 Å (bromine
edge) to collect anomalous data. Crystals were kept at 100 K during data collection using a
MARCCD 300 detector. Diffraction data were integrated, scaled and merged with HKL2000.43
A summary of selected crystal data and data collection parameters is provided in Table S10. The
structures were phased by the singleꢀwavelength anomalous dispersion (SAD) technique using
the program HKL2MAP.44,45 The resulting experimental electron density maps allowed
visualization of all nucleotides in the unit cell (48, SgnTꢀ12ꢀmer; 24, RgnTꢀ12ꢀmer). Building of
the models was performed using the program COOT.46 The initial orientations of the duplex
models were optimized by several rounds of rigid body refinement while gradually increasing
the resolution of the diffraction data. The refinements were carried out with the program CNS47,
performing simulated annealing, gradient minimization, and refinement of individual isotropic
temperature factors. The later refinements were done using the program SHELX44, keeping
aside 5% of the reflections to compute the Rꢀfree. After a few cycles of refinement in SHELX,
GNA residues were built into the electron density, and the dictionary was adapted and the
refinement continued. Ions and water molecules were added on the basis of Fourier 2FoꢀFc sum
and FoꢀFc difference electron density maps, and accepted on the basis of standard distance and
Bꢀfactor criteria. Final refinements were carried out using anisotropic temperature factors for all
nucleic acid and solvent atoms and the final refinement parameters and deviations from ideal
geometries are listed in Table S10.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Data from the SgnTꢀ8ꢀmer crystal containing were collected at 100K using the Dectris
Eiger 9M detector on the 21ꢀIDꢀD beamline at LSꢀCAT tuned to 0.91836 Å (bromine high
energy remote). Diffraction data were integrated, scaled and merged with HKL2000.43
A
summary of selected crystal data and data collection parameters is provided in Table S10. The
structure was phased by SAD with the program HKL2MAP.44,45 The resulting experimental
electron density maps allowed visualization of all 32 nucleotides in the asymmetric unit.
Building of the models was performed using the program COOT.46 Refinements were carried
out using the PHENIX package.48,49 GNA residues were built into the electron density and
refinement continued with a dictionary created using PRODRG.50 Ions and water molecules
were added on the basis of Fourier 2FoꢀFc sum and FoꢀFc difference electron density maps, and
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 28
1
2
3
4
5
6
7
8
accepted on the basis of standard distance and Bꢀfactor criteria. Final refinements were carried
out using anisotropic temperature factors for all nucleic acid and solvent atoms and the final
refinement parameters and deviations from ideal geometries are listed in Table S10.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In vitro screening. Primary hepatocytes were obtained from Gibco and cultured in Williams E
Medium (Gibco) with 10% fetal bovine serum. Transfection of duplexes was accomplished in a
384ꢀwell format, containing 40 µL complete medium, 4.9 µL OptiMem I (Gibco), 0.1 µL
RNAiMax (Invitrogen), and 5 µL siRNA. Cells were plated at a density of 5,000 cells per well
on collagenꢀcoated plates, and siRNA was added to a final concentration of 10 nM or 0.1 nM.
Twentyꢀfour hours after transfection, samples were collected for processing. RNA extraction
was performed using Dynabeads (Inivtrogen), followed by reverse transcription according to the
manufacturer protocols (ABI, High Capacity). Levels of TTR mRNA were measured after
incubation with 10 nM or 0.1 nM siRNA in primary mouse hepatocytes for 24 hours. Mouse
TTR was quantified by RTꢀqPCR (Applied Biosystems # Mm00443267_m1), with mouse
GAPDH (4352339E) as loading control. Relative levels of TTR were determined by
normalizing to GAPDH RNA expression from the same sample, and these values were
subsequently expressed as percent of nonꢀtargeting control. All data points are the average of
four measurements. The siRNA concentration required to inhibit relative TTR expression by
50% (IC50) was calculated for a subset of the siRNAs by transfecting as described, eight
concentrations ranging from 10 nM – 37.5 fM in a sixꢀfold dilution series. IC50 values were
derived from a 4 parameter fit model with XLFit.
In vivo screening. All studies were conducted using protocols consistent with local, state and
federal regulations, as applicable, and approved by the Institutional Animal Care and Use
Committee (IACUC) at Alnylam Pharmaceuticals. Animals received a single subscapular
subcutaneous injection of 2.5 mg/kg siRNA, prepared as an injection volume of 10 µL/g in PBS.
At the indicated time preꢀ or postꢀdosing, animals were anesthetized with isofluorane and blood
obtained via retroorbital bleed. TTR protein was quantified by ELISA from serum isolated from
whole blood. ELISA was performed according to manufacturer protocol (ALPCO, 41ꢀPALMSꢀ
E01) after a 3025ꢀfold dilution of the serum samples. Data were normalized to preꢀbleed TTR
levels. All samples were assayed in duplicate and each data point is the average of all the mice
ACS Paragon Plus Environment

Page 25 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
within each cohort (n=3). Data were analyzed using a twoꢀway ANOVA with a Tukey postꢀhoc
test for multiple comparison in GraphPad Prism.
7
8
9
For evaluation of TTR mRNA in liver, mice were euthanized with carbon dioxide after
collection of blood, the liver harvested, and snapꢀfrozen in liquid nitrogen. The liver tissues
were processed in a SPEX GenoGrinder, and a small amount powderized tissue transferred to
tubes for lysis. RNA extraction was performed with PureLink Pro 96 total RNA Purification Kit
(Life Technologies), with RTꢀqPCR performed as described above. Data were analyzed using a
twoꢀway ANOVA with a Tukey postꢀhoc test for multiple comparisons in GraphPad Prism.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Associated Content
Detailed synthetic procedures, analytical data for oligonucleotides, crystallographic tables,
thermal melting data, and in vitro screening data. The coordinate files of all structures
determined using xꢀray crystallography have been deposited in the Protein Data Bank under PDB
codes 5V1L, 5V1K, and 5V2H.
Author Information
Corresponding authors: mschlegel@alnylam,com, mmanoharan@alnylam.com
Notes. The authors declare the following competing financial interest(s): All Alnylam authors
except J.G.L. are current employees of Alnylam.
Acknowledgements
The authors thank Nate Taneja for help in oligonucleotide purification and Dr. Stuart Milstein
for helpful discussions.
References
(1) Elbashir, S. M.; Harborth, J.; Lendeckel, W.; Yalcin, A.; Weber, K.; Tuschl, T. Nature
2001, 411, 494.
(2) Wang, H.ꢀW.; Noland, C.; Siridechadilok, B.; Taylor, D. W.; Ma, E.; Felderer, K.;
Doudna, J. A.; Nogales, E. Nat. Struct. Mol. Biol. 2009, 16, 1148.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 28
1
2
3
4
5
(3) Meister, G.; Landthaler, M.; Patkaniowska, A.; Dorsett, Y.; Teng, G.; Tuschl, T. Mol.
Cell 2004, 15, 185.
(4) Liu, J.; Carmell, M. A.; Rivas, F. V.; Marsden, C. G.; Thomson, J. M.; Song, J.ꢀJ.;
Hammond, S. M.; JoshuaꢀTor, L.; Hannon, G. J. Science 2004, 305, 1437.
(5) Martinez, J.; Patkaniowska, A.; Urlaub, H.; Lührmann, R.; Tuschl, T. Cell 2002, 110,
563.
(6) Deleavey, G. F.; Watts, J. K.; Damha, M. J. Current Protocols in Nucleic Acid
Chemistry; Wiley: New York, 2009; Chapter 16, Unit 16.3.
(7) Nair, J. K.; Willoughby, J. L. S.; Chan, A.; Charisse, K.; Alam, M. R.; Wang, Q.;
Hoekstra, M.; Kandasamy, P.; Kel’in, A. V.; Milstein, S.; Taneja, N.; O’Shea, J.; Shaikh, S.;
Zhang, L.; van der Sluis, R. J.; Jung, M. E.; Akinc, A.; Hutabarat, R.; Kuchimanchi, S.;
Fitzgerald, K.; Zimmermann, T.; van Berkel, T. J. C.; Maier, M. A.; Rajeev, K. G.;
Manoharan, M. J. Am. Chem. Soc. 2014, 136, 16958.
(8) Chan, A.; Liebow, A.; Yasuda, M.; Gan, L.; Racie, T.; Maier, M.; Kuchimanchi, S.;
Foster, D.; Milstein, S.; Charisse, K.; Sehgal, A.; Manoharan, M.; Meyers, R.; Fitzgerald, K.;
Simon, A.; Desnick, R. J.; Querbes, W. Mol. Ther.--Nucleic Acids 2015, 4, e263.
(9) Liebow, A.; Li, X.; Racie, T.; Hettinger, J.; Bettencourt, B. R.; Najafian, N.; Haslett, P.;
Fitzgerald, K.; Holmes, R. P.; Erbe, D.; Querbes, W.; Knight, J. J. Am. Soc. Nephrol. 2017,
28, 494.
(10) Fitzgerald, K.; White, S.; Borodovsky, A.; Bettencourt, B. R.; Strahs, A.; Clausen, V.;
Wijngaard, P.; Horton, J. D.; Taubel, J.; Brooks, A.; Fernando, C.; Kauffman, R. S.; Kallend,
D.; Vaishnaw, A.; Simon, A. New Engl. J. Med. 2017, 376, 41.
(11) Laursen, M. B.; Pakula, M. M.; Gao, S.; Fluiter, K.; Mook, O. R.; Baas, F.; Langklaer,
N.; Wengel, S. L.; Wengel, J.; Kjems, J.; Bramsen, J. B. Mol. Biosyst. 2010, 6, 862.
(12) Kenski, D. M.; Cooper, A. J.; Li, J. J.; Willingham, A. T.; Haringsma, H. J.; Young, T.
A.; Kuklin, N. A.; Jones, J. J.; Cancilla, M. T.; McMasters, D. R.; Mathur, M.; Sachs, A. B.;
Flanagan, W. M. Nucleic Acids Res. 2010, 38, 660.
(13) Kamiya, Y.; Takai, J.; Ito, H.; Murayama, K.; Kashida, H.; Asanuma, H. ChemBioChem
2014, 15, 2549.
(14) Alagia, A.; Terrazas, M.; Eritja, R. Molecules 2015, 20, 7602.
(15) Khvorova, A.; Reynolds, A.; Jayasena, S. D. Cell 2003, 115, 209.
(16) Sano, M.; Sierant, M.; Miyagishi, M.; Nakanishi, M.; Takagi, Y.; Sutou, S. Nucleic
Acids Res. 2008, 36, 5812.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(17) Bramsen, J. B.; Laursen, M. B.; Nielsen, A. F.; Hansen, T. B.; Bus, C.; Langkjær, N.;
Babu, B. R.; Højland, T.; Abramov, M.; Van Aerschot, A.; Odadzic, D.; Smicius, R.; Haas,
J.; Andree, C.; Barman, J.; Wenska, M.; Srivastava, P.; Zhou, C.; Honcharenko, D.; Hess, S.;
Müller, E.; Bobkov, G. V.; Mikhailov, S. N.; Fava, E.; Meyer, T. F.; Chattopadhyaya, J.;
Zerial, M.; Engels, J. W.; Herdewijn, P.; Wengel, J.; Kjems, J. Nucleic Acids Res. 2009, 37,
2867.
(18) Addepalli, H.; Meena; Peng, C. G.; Wang, G.; Fan, Y.; Charisse, K.; Jayaprakash, K. N.;
Rajeev, K. G.; Pandey, R. K.; Lavine, G.; Zhang, L.; JahnꢀHofmann, K.; Hadwiger, P.;
Manoharan, M.; Maier, M. A. Nucleic Acids Res. 2010, 38, 7320.
(19) Zhang, L.; Peritz, A.; Meggers, E. J. Am. Chem. Soc. 2005, 127, 4174.
(20) Schlegel, M. K.; Peritz, A. E.; Kittigowittana, K.; Zhang, L.; Meggers, E.
ChemBioChem 2007, 8, 927.
(21) Schlegel, M. K.; Essen, L.ꢀO.; Meggers, E. J. Am. Chem. Soc. 2008, 130, 8158.
ACS Paragon Plus Environment

Page 27 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
(22) Schlegel, M. K.; Essen, L.ꢀO.; Meggers, E. Chem. Comm. 2010, 46, 1094.
(23) Johnson, A. T.; Schlegel, M. K.; Meggers, E.; Essen, L.ꢀO.; Wiest, O. J. Org. Chem.
2011, 76, 7964.
7
8
9
(24) Elkayam, E.; Kuhn, C.ꢀD.; Tocilj, A.; Haase, Astrid D.; Greene, Emily M.; Hannon,
Gregory J.; JoshuaꢀTor, L. Cell 2012, 150, 100.
(25) Roberts, C.; Bandaru, R.; Switzer, C. J. Am. Chem. Soc. 1997, 119, 4640.
(26) Chen, X.; Kierzek, R.; Turner, D. H. J. Am. Chem. Soc. 2001, 123, 1267.
(27) Allerson, C. R.; Sioufi, N.; Jarres, R.; Prakash, T. P.; Naik, N.; Berdeja, A.; Wanders,
L.; Griffey, R. H.; Swayze, E. E.; Bhat, B. J. Med. Chem. 2005, 48, 901.
(28) Deleavey, Glen F.; Damha, Masad J. Chem. Biol. 2012, 19, 937.
(29) Kenski, D. M.; Butora, G.; Willingham, A. T.; Cooper, A. J.; Fu, W.; Qi, N.; Soriano,
F.; Davies, I. W.; Flanagan, W. M. Mol. Ther.--Nucleic Acids 2012, 1, e5.
(30) Schirle, N. T.; MacRae, I. J. Science 2012, 336, 1037.
(31) Schirle, N. T.; SheuꢀGruttadauria, J.; MacRae, I. J. Science 2014, 346, 608.
(32) Manoharan, M.; Akinc, A.; Pandey, R. K.; Qin, J.; Hadwiger, P.; John, M.; Mills, K.;
Charisse, K.; Maier, M. A.; Nechev, L.; Greene, E. M.; Pallan, P. S.; Rozners, E.; Rajeev, K.
G.; Egli, M. Angew. Chem., Int. Ed. 2011, 50, 2284.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(33) Nakanishi, K. Wiley Interdiscip. Rev.: RNA 2016, 7, 637.
(34) Ueda, N.; Kawabata, T.; Takemoto, K. J. Heterocycl. Chem. 1971, 8, 827.
(35) Seita, T.; Yamauchi, K.; Kinoshita, M.; Imoto, M. Makromol. Chem. 1972, 154, 255.
(36) Holý, A.; Ivanova, G. S. Nucleic Acids Res. 1974, 1, 19.
(37) Cook, P. D.; Acevedo, O. L.; Davis, P. W.; Ecker, D. J.; Hebert, N. Monomeric Diols
and Phosphate Linked Oligomers Formed Therefrom. PCT Int. Appl. WO 9518820, 1995.
(38) Cook, P. D.; Acevedo, O. L.; Davis, P. W.; Ecker, D. J.; Hebert, N. Phospate Linked
Oligomers. U.S. Patent 5,886,177, 1999.
(39) Nielsen, P.; Dreiøe, L. H.; Wengel, J. Bioorg. Med. Chem. 1995, 3, 19.
(40) Zhang, L.; Peritz, A. E.; Carroll, P. J.; Meggers, E. Synthesis 2006, 2006, 645.
(41) Schlegel, M. K.; Meggers, E. J. Org. Chem. 2009, 74, 4615.
(42) Jancarik, J.; Kim, S.ꢀH. J. Appl. Crystallogr. 1991, 24, 409.
(43) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307.
(44) Schneider, T. R.; Sheldrick, G. M. Acta Crystallogr., Section D 2002, 58, 1772.
(45) Pape, T.; Schneider, T. R. J. Appl. Crystallogr. 2004, 37, 843.
(46) Emsley, P.; Cowtan, K. Acta Crystallogr., Section D 2004, 60, 2126.
(47) Brunger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W. L.; Gros, P.; Grosseꢀ
Kunstleve, R. W.; Jiang, J.ꢀS.; Kuszewski, J.; Nilges, M.; Pannu, N. S.; Read, R. J.; Rice, L.
M.; Simonson, T.; Warren, G. L. Acta Crystallogr., Section D 1998, 54, 905.
(48) Adams, P. D.; Afonine, P. V.; Bunkoczi, G.; Chen, V. B.; Davis, I. W.; Echols, N.;
Headd, J. J.; Hung, L.ꢀW.; Kapral, G. J.; GrosseꢀKunstleve, R. W.; McCoy, A. J.; Moriarty,
N. W.; Oeffner, R.; Read, R. J.; Richardson, D. C.; Richardson, J. S.; Terwilliger, T. C.;
Zwart, P. H. Acta Crystallogr., Section D 2010, 66, 213.
(49) Afonine, P. V.; GrosseꢀKunstleve, R. W.; Echols, N.; Headd, J. J.; Moriarty, N. W.;
Mustyakimov, M.; Terwilliger, T. C.; Urzhumtsev, A.; Zwart, P. H.; Adams, P. D. Acta
Crystallogr., Section D 2012, 68, 352.
(50) Schuttelkopf, A. W.; van Aalten, D. M. F. Acta Crystallogr., Section D 2004, 60, 1355.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 28
1
2
3
4
Suggestion for Table of Contents Entry
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment