Intrastrand locks increase duplex stability and base pairing selectivity

Article abstract of DOI:10.1039/c1cc14008f

Oligodeoxynucleotide probes with disulfide locks between neighboring nucleobases show increases in melting point for duplexes with RNA target strands of up to 7.6 °C. The weakly pairing TT dimers are replaced with locked 2′-deoxy-5-(thioalkynyl)uridine re

Full text of DOI:10.1039/c1cc14008f

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2011, 47, 10824–10826
www.rsc.org/chemcomm
COMMUNICATION
Intrastrand locks increase duplex stability and base pairing selectivityw
Cora Prestinari and Clemens Richert*
Received 5th July 2011, Accepted 6th August 2011
DOI: 10.1039/c1cc14008f
Oligodeoxynucleotide probes with disulfide locks between neigh-
boring nucleobases show increases in melting point for duplexes
with RNA target strands of up to 7.6 8C. The weakly pairing TT
dimers are replaced with locked 20-deoxy-5-(thioalkynyl)uridine
residues via automated synthesis.
oligonucleotides in biology. Cross links between strands
(Fig. 1c) are detrimental for cells, but synthetic constructs, such
as methylene bridged base pairs,6 metal–salen complexes,7
bis-maleimides,8 or triazolides,9 produce valuable bioorganic
compounds. For triplexes, interstrand cross links and links in
circular DNA can increase stability and sequence selectivity.10
A third possible approach involves intrastrand locks between
nucleotides of a hybridization probe (Fig. 1b). Experimental
realizations are rare.11 We were interested in the effect of
intrastrand locks on hybridization probes. We chose disulfide
links between neighboring nucleotides, assuming that they
would form spontaneously in the presence of air.12–14 We
replaced TT dinucleotides as the most weakly pairing
dinucleotide, as a step towards isostable duplexes for high
fidelity detection in massively parallel fashion.15,16 High fidelity is
very important for probes binding short RNA species, such as
microRNAs,17 that are difficult to amplify by PCR.18
The affinity of oligonucleotides for target strands is weaker
than the number of base pairs may suggest, due to a large
entropy–enthalpy compensation during duplex formation.1
Hydrogen bonding, stacking, and other energetically favorable
interactions are compensated by
a loss in translation,
conformational, and vibrational degrees of freedom, so that
the bulk of the binding enthalpy is not translated into a gain in
free energy. The large number of conformers accessible to
oligonucleotides also reduces sequence selectivity. Non-
Watson–Crick pairing is compatible with the backbone of
DNA, making duplex formation less predictable than desirable
for hybridization probes binding a single target in a genomic
context. The likelihood of mismatch formation is greatest at
the termini, where fraying and wobbling is common.2
Restricting the conformational flexibility of oligonucleotide
probes to the conformation found in the desired duplex should
increase target affinity and selectivity.
The intrastrand locks were introduced at the 5-position of
20-deoxyuridine residues via butynyl and hexynyl linkers
(Scheme 1). Other oligodeoxynucleotides with 20-deoxy-
uridines as the site of disulfide cross linking are known.10,19
Alkynyl substituents at the 5-position point into the major
groove of duplexes,20–23 where they provide additional
stacking interactions.24
The concept of rigidifying oligonucleotides to gain target affinity
has been demonstrated impressively for nucleosides with covalent
Phosphoramidites 1 and 2 were prepared following a route
for related compounds.10a The route of Benner and coworkers25
gave lower yields. Silyl-protected 5-iodo-20-deoxyuridine (3)
was cross-coupled to the alkynyl alcohol, and the resulting
deoxyuridines (4 or 5) were mesylated and reacted with
thiobenzoate to give protected thioalkynyldeoxyuridines 6
and 7. Desilylation with the known mixture of HFꢀPy and
TBAF10a gave 8/9, and dimethoxytritylation gave 10/11.
Phosphitylation produced phosphoramidites 1 and 2 in 30%
and 44% overall yield, respectively. Oligodeoxynucleotides
were assembled via automated solid-phase synthesis.
Pre-treatment with NEt3 was followed by full deprotection
and cleavage from the support with NH3, MeNH2, and
DTT.10a,26 Intrastrand locks formed upon diluting with buffer
(pH 5) and exposure to air. Remaining cyanoethyl adducts
were removed by HPLC, giving 12–22 in up to 22% yield.
Dithiol and disulfide differ in mass by just 2 Da, so MALDI
spectra were also measured in reflectron mode with internal
standard to confirm structures (Fig. S22–S43, ESIw).
links that lock
a desired conformation of the ribose
(Fig. 1a). Bicyclo-3 or tricyclo-DNA,4 and locked nucleic acids
(LNA)5 are among the most successful classes of modified
Fig. 1 Locks for nucleic acids (bold lines). (a) Locked nucleoside
(LNA), (b) intrastrand-lock, and (c) interstrand cross link.
Institute for Organic Chemistry, University of Stuttgart,
Pfaffenwaldring 55, 70565 Stuttgart, Germany
E-mail: lehrstuhl-2@oc.uni-stuttgart.de; Fax: 0049 (0)711/685-64311;
Tel: 0049 (0)711/685-64321
w Electronic supplementary information (ESI) available: Synthetic
and analytical details for nucleosides and oligodeoxynucleotides. See
DOI: 10.1039/c1cc14008f
Self-complementary control duplex (50-TTGCGCAA-30)2
(23)2 and locked duplexes (12)2–(14)2 were subjected to
UV-melting. The latter showed an increase in melting point
c
10824 Chem. Commun., 2011, 47, 10824–10826
This journal is The Royal Society of Chemistry 2011

View Article Online
Scheme 1 (I) Synthesis of intrastrand-locked oligodeoxynucleotides 12–21.a (II) Structure of intrastrand-locked d(US4US2):r(AA) dimer duplex,
as generated by molecular modeling for A-form geometry using Macromodel.
of 8.6–13.7 1C (4.3–6.9 1C per lock, Table 1). The short-hand
for a residue with an ethylene unit in the side chain is US2, and
for a residue with a butylene chain it is US4. The mixed
construct with a lock containing one US2 and one US4 gave
the largest DTm. Next, we studied non-self complementary
sequence 50-TTTTCCAC-30 (24) with DNA and RNA targets.
A total of six locks was introduced at different positions
(15–20, Scheme 1). The highest UV-melting point was found
for RNA as a target, with +7.6 1C for the US2/US4 lock at the
terminus of probe (16:25).
Table 1 UV melting points with perfect match target strands
Much smaller increases in Tm were found for locks bridging
non-neighboring residues (18:25 and 19:25), and a long lock in
the middle of the duplex (20:25). When DNA was the target
(26), slightly lower melting point increases were found, and the
shortest of the links gave the strongest effect (+7.0 1C, 17:26).
Locks clamping a base triplet gave melting point depressions
of 2.2–2.4 1C (Table 1).
Probe strand
Target strand
Tma/1C
DTmb/1C
23c
12c
13c
14c
41.9 ꢂ 0.6
53.0 ꢂ 0.5
55.6 ꢂ 0.9
50.5 ꢂ 0.9
—
+11.1/5.6d
+13.7/6.9d
+8.6/4.3d
r(GUGGAAAA) (25)
24
15
16
17
18
19
20
30.3 ꢂ 0.5
36.7 ꢂ 0.5
37.9 ꢂ 0.6
37.7 ꢂ 0.6
31.9 ꢂ 0.6
33.5 ꢂ 0.5
31.8 ꢂ 0.5
—
A third sequence motif containing the locks in the middle of
the strand was based on 50-CCTTTTAC-30 (27). Melting points
of 21 and 22 with both RNA and DNA target strands showed
that the short lock with two ethylene units flanking the disulfide
is not well tolerated in the interior of the helix. Locks with
butylene chains gave melting point increases, though, with the
higher DTm for the duplex with the RNA target (+4.2 1C; 21:28).
Next, we studied base pairing selectivity by measuring UV-
melting points of duplexes with mismatched bases opposite locks
(Fig. S44, ESIw, Table 2). In every single case, the locked strands
showed greater selectivity than their unmodified counterparts.
For locked strands 21 and 22, melting points with such targets
were even o10 1C. The effect is pronounced at the terminus,
where the control duplex (24:30) gives barely any melting point
depression for a T:U mismatch (ꢁ1.7 1C). Increases in DTm over
that of the control were also found for the penultimate position
and the third position from the terminus. When DNA is the
target, mismatch discrimination is greater to begin with, but
increases further in the presence of the lock. For the butylene/
butylene lock, a DTm of almost 10 1C for a T:T mismatch at the
very terminus was found, a value exceeding that induced by the
best commercial fidelity-enhancing cap.27
+6.4
+7.6
+7.4
+1.6
+3.2
+1.5
GTGGAAAA (26)
24
15
16
17
18
19
20
24.9 ꢂ 0.4
31.3 ꢂ 0.8
30.4 ꢂ 0.6
31.9 ꢂ 0.4
22.7 ꢂ 0.9
22.5 ꢂ 0.9
27.8 ꢂ 0.8
—
+6.4
+5.5
+7.0
ꢁ2.2
ꢁ2.4
+2.9
r(GUAAAAGG) (28)
GTAAAAGG (29)
27
21
22
24.8 ꢂ 0.5
29.0 ꢂ 0.7
23.5 ꢂ 1.7
—
+4.2
ꢁ1.3
27
21
22
23.8 ꢂ 1.0
26.2 ꢂ 0.5
19.0 ꢂ 0.6
—
+2.4
ꢁ4.8
a
Average of 4 curves ꢂ SD 1.5 mM strand concentration and 1 M
b
NaCl, 10 mM phosphate buffer, pH 7.0. DTm to unmodified control
duplex. Self-complementary sequence; no separate target strand.
c
d
DTm per lock.
c
This journal is The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 10824–10826 10825

View Article Online
Table 2 UV melting points of duplexes containing U:U or U:T
mismatches. Mismatched bases are given in boldface
and a two-step deprotection. Intrastrand-locked nucleic acids
may find use in biomedicine where high fidelity hybridization
is required. We are currently studying other nucleobases and
nucleotide distances.
Probe strand
Target strand
Tma/1C
DTmb/1C
r(GUGGAAAU) (30)
The authors thank C. Deck for oligonucleotide synthesis, M.
Fichte for CD spectra, and DFG for funding (RI 1063/9-1).
24
15
16
17
18
28.6 ꢂ 0.6
30.3 ꢂ 1.0
31.9 ꢂ 0.5
32.9 ꢂ 0.7
28.6 ꢂ 0.7
ꢁ1.7
ꢁ6.4
ꢁ6.0
ꢁ4.8
ꢁ3.3
Notes and references
r(GUGGAAUA) (31)
r(GUGGAUAA) (32)
GTGGAAAT (33)
24
19
27.0 ꢂ 0.3
27.9 ꢂ 0.3
ꢁ3.3
ꢁ5.6
1 (a) G. Vesnaver and K. J. Breslauer, Proc. Natl. Acad. Sci. U. S. A.,
1991, 88, 3569–3573; (b) M. S. Searle and D. H. Williams, Nucleic
Acids Res., 1993, 21, 2051–2056.
24
20
20.7 ꢂ 0.4
21.4 ꢂ 0.8
ꢁ9.6
2 (a) D. Andreatta, S. Sen, J. L. Perez Lustres, S. A. Kovalenko,
´
ꢁ10.4
N. P. Ernsting, C. J. Murphy, R. S. Coleman and M. A. Berg,
J. Am. Chem. Soc., 2006, 128, 6885–6892; (b) K.-W. Tsai,
W.-C. Chan, C.-N. Hsu and W.-C. Lin, BMC Mol. Biol., 2010,
11, 34–42.
3 (a) M. Tarkoy, M. Bolli, B. Schweizer and C. J. Leumann, Helv.
Chim. Acta, 1993, 76, 481–510; (b) A. Stauffinger and
C. J. Leumann, Eur. J. Org. Chem., 2009, 1153–1162.
4 (a) R. Steffens and C. J. Leumann, J. Am. Chem. Soc., 1997, 119,
11548–11549; (b) D. Ittig, A. B. Gerber and C. J. Leumann, Nucleic
Acids Res., 2011, 39, 373–380.
5 (a) S. K. Singh, P. Nielsen, A. A. Koshkin and J. Wengel, Chem.
Commun., 1998, 455–456; (b) A. A. Koshkin, P. Nielsen,
M. Meldgaard, V. K. Rajwanshi, S. K. Singh and J. Wengel,
J. Am. Chem. Soc., 1998, 120, 13252–13253.
6 X. Qiao and Y. Kishi, Angew. Chem., Int. Ed., 1999, 38, 928–931.
´
7 K. V. Gothelf, A. Thomsen, M. Nielsen, E. Clo and R. S. Brown,
24
15
16
17
18
21.5 ꢂ 0.8
21.7 ꢂ 1.4
22.9 ꢂ 1.0
22.8 ꢂ 1.8
19.1 ꢂ 0.8
ꢁ3.4
ꢁ9.6
ꢁ7.2
ꢁ9.2
ꢁ3.6
GUGGAATA (34)
GUGGATAA (35)
24
19
18.0 ꢂ 1.0
14.4 ꢂ 0.4
ꢁ6.9
ꢁ8.1
24
20
11.2 ꢂ 1.3
ꢁ13.7
11.1 ꢂ 1.9
ꢁ16.7
a
Average of 4 curves ꢂ SD 1.5 mM strand concentration and 1 M NaCl,
b
10 mM phosphate buffer, pH 7.0. Melting point difference to perfect
match control duplex (compare Table 1).
J. Am. Chem. Soc., 2004, 126, 1044–1046.
8 R. Tona and R. Haner, Mol. BioSyst., 2005, 1, 93–98.
¨
9 A. H. El-Sagheer, R. Kumar, S. Findlow, J. M. Werner,
A. N. Lane and T. Brown, ChemBioChem, 2008, 9, 50–52.
10 (a) N. C. Chaudhuri and E. T. Kool, J. Am. Chem. Soc., 1995, 117,
10434–10442; (b) E. T. Kool, Chem. Rev., 1997, 97, 1473–1487.
11 (a) A. Kumar, E-J. Chem., 2011, 8, 507–512; (b) Y. Liu, R. Wang,
L. Ding, R. Sha, P. S. Lukeman, J. W. Canary and N. C. Seeman,
ChemBioChem, 2008, 9, 1641–1648.
12 A. E. Ferentz and G. L. Verdine, J. Am. Chem. Soc., 1991, 113,
4000–4002.
13 J. Milton, B. A. Connolly, T. T. Nikiforov and R. Cosstick,
J. Chem. Soc., Chem. Commun., 1993, 779–780.
Molecular modeling suggests that the ethylene/butylene
combination can form an unstrained link (Scheme 1–II), and
so do CD spectra of duplexes of 16 with 25 and 26 (ESIw,
Fig. S45–S46). Thermodynamic data show that the stabilizing
effect is indeed of entropic origin (Table S1, ESIw). Covalently
locked neighboring nucleotides are most stabilizing at the very
terminus, where pairing is weak and wobbling is pronounced.
This is also where the fidelity-enhancing effect is most desirable,
to overcome the poor discriminating ability of unmodified
probes.
14 (a) G. D. Glick, J. Org. Chem., 1991, 56, 6746–6747;
(b) G. D. Glick, Biopolymers, 1998, 48, 83–96.
15 C. Ahlborn, K. Siegmund and C. Richert, J. Am. Chem. Soc., 2007,
129, 15218–15232.
Disulfides have several advantages over other locking
chemistries, such as reductive amination and amide formation
that were also tested in the early phase of our study. Firstly,
there is no need for additional synthetic steps. The lock forms
spontaneously after automated DNA synthesis and deprotec-
tion. Secondly, there is excellent orthogonality to other
functional groups, and no side reactions must be feared, other
than alkylation and intermolecular locking, which appears to
compete poorly with intramolecular locking under our conditions.
Thiols also have minimal steric demand, making it easy to
close very large rings, and disulfide formation has a large free
energy driving force, avoiding incomplete reactions. Still,
disulfide locking can be reversed through reduction. Exploratory
melting curves with duplex 16:25 show a Tm drop of 5.3 1C
when DTT is added. The Tm of the duplex of 18, with its poor
three nucleotide lock, and 25 increased by 1.5 1C upon
addition of DTT.
16 K. Siegmund, C. Ahlborn and C. Richert, Nucleosides, Nucleotides
Nucleic Acids, 2008, 27, 376–388.
17 D. Baek, J. Villen, C. Shin, F. D. Camargo, S. P. Gygi and
´
D. P. Bartel, Nature, 2008, 455, 64–71.
18 A. W. Wark, H. J. Lee and R. M. Corn, Angew. Chem., Int. Ed.,
2008, 47, 644–652.
19 (a) J. T. Goodwin and G. D. Glick, Tetrahedron Lett., 1993, 34,
5549–5552; (b) S. E. Osborne, R. J. Cain and G. D. Glick, J. Am.
Chem. Soc., 1997, 119, 1171–1182.
20 T. Kottysch, C. Ahlborn, F. Brotzel and C. Richert, Chem.–Eur.
J., 2004, 10, 4017–4028.
21 T. J. Bandy, A. Brewer, J. R. Burns, G. Marth, T. Nguyen and
E. Stulz, Chem. Soc. Rev., 2011, 40, 138–148.
22 M. Ahmadian, P. Zhang and D. E. Bergstorm, Nucleic Acids Res.,
1998, 26, 3127–3135.
23 I. Luyten and P. Herdewijn, Eur. J. Med. Chem., 1998, 33,
515–576.
24 (a) T. W. Barnes III and D. H. Turner, J. Am. Chem. Soc., 2001,
123, 4107–4118; (b) B. M. Znosko, T. W. Barnes III, T. R. Krugh
and D. H. Turner, J. Am. Chem. Soc., 2003, 125, 6090–6097.
25 (a) H. A. Held, A. Roychowdhury and S. A. Benner, Nucleosides,
Nucleotides Nucleic Acids, 2003, 22, 391–404; (b) H. A. Held and
S. A. Benner, Nucleic Acids Res., 2002, 30, 3857–3869.
26 F. Hofbauer and I. Frank, Chem.–Eur. J., 2010, 16, 5097–5101.
In conclusion, we have demonstrated that intrastrand locks
of proper length elevate the melting point of duplexes with
DNA and RNA target strands. The locks can be introduced
through DNA synthesis with phosphoramidite building blocks
27 Z. Dogan, R. Paulini, J. A. Rojas Stutz, S. Narayanan and
¨
C. Richert, J. Am. Chem. Soc., 2004, 126, 4762–4763.
c
10826 Chem. Commun., 2011, 47, 10824–10826
This journal is The Royal Society of Chemistry 2011