Templated DNA ligation with thiol chemistry

Article abstract of DOI:10.1039/c4ob01552e

We describe two DNA-templated ligation strategies: native chemical ligation (NCL), and thiol-disulfide exchange. Both systems result in successful ligation in the presence of a DNA template. The stability of the product from the NCL reaction relies on exogenous thiol, while the thiol-disulfide reaction proceeds in a catalyst-free manner. This journal is

Full text of DOI:10.1039/c4ob01552e

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
COMMUNICATION
Templated DNA ligation with thiol chemistry†
Dadong Li,‡ Xiaojian Wang,‡ Fubo Shi,‡ Ruojie Sha, Nadrian C. Seeman* and
James W. Canary*
Cite this: Org. Biomol. Chem., 2014,
12, 8823
Received 23rd July 2014,
Accepted 24th September 2014
DOI: 10.1039/c4ob01552e
www.rsc.org/obc
We describe two DNA-templated ligation strategies: native chemi-
cal ligation (NCL), and thiol-disulfide exchange. Both systems
result in successful ligation in the presence of a DNA template. The
stability of the product from the NCL reaction relies on exogenous
thiol, while the thiol-disulfide reaction proceeds in a catalyst-free
manner.
Template-directed DNA ligation reactions are of great interest.
For example, such reactions are essential in self-replicating
systems, where a template initiates a process and released pro-
ducts serve as templates for subsequent reactions.1 In nature,
DNA replication is one of the fundamental processes of any
living system. Not surprisingly, systems utilizing DNA as tem-
plates have drawn great attention for decades. Various reac-
tions have been reported to ligate two DNA strands, including
phosphodiester formation activated by carbodiimide2 or N-cya-
noimidazole,3 imine formation,4 photoinduced [2 + 2] cyclo-
addition,5 azide–alkyne [3 + 2] cycloaddition,6 and the Wittig
reaction.7 The key challenge has always been to achieve high
yields with simple reaction conditions. Our groups have
reported DNA-templated ligations utilizing amide formation
with excellent coupling yields. However, catalysis was required
to activate the carboxylates.8 Here we report two DNA-tem-
plated systems utilizing thiol chemistry, including native
chemical ligation (NCL) and disulfide bond formation (DSF)
(Scheme 1).
The concept of NCL was first introduced in 1990s by the
Kent group,9 and it has been shown to be the most robust and
practical of the available ligation chemistries for covalently
conjugating two unprotected peptide segments.10 Not limited
to protein ligations, NCL has been applied to the conjugation
of a variety of molecules, including peptide-oligonucleotide,11
small molecule-oligonucleotide,12 protein-liposome,13 and
Scheme 1 Templated DNA ligation using thiol-based chemistry. (A)
General ligation scheme; (B) native chemical ligation (NCL); (C) disul-
phide exchange (DSF). The template strand was omitted in both (B) and
(C). RSH: thiophenol or 4-mercaptophenylacetic acid (MPAA). TCEP: tris
(2-carboxyethyl) phosphine.
Department of Chemistry, New York University, 100 Washington Sq. East, New York,
NY 10003, USA. E-mail: james.canary@nyu.edu, ned.seeman@nyu.edu
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4ob01552e
‡These authors contributed equally to this work.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8823–8827 | 8823

View Article Online
Communication
Organic & Biomolecular Chemistry
templated PNA ligations.14 Nevertheless, it has not been
reported that NCL can be applied to DNA–DNA ligation.
Table 1 Strand sequences
Code
S0
Oligonucleotide sequences
5′-TTTTGTAGGACAGCTCTCTG(T)n -
GAGTTACTGGACAAGATTTT-3′
5′-CAGAGAGCTGTCCTAC-3′
5′-CAGAGAGC-3′
Our initial plan to achieve DNA–DNA ligation followed the
principle of NCL: an exogenous thiol activates the thioester and
catalyzes the ligation, transthioesterification with the side chain
thiol of an N-terminal cysteine creates a thioester-linked inter-
mediate, and the thioester-linked intermediate undergoes intra-
molecular nucleophilic rearrangement to form an amide bond
at the ligation site.15 In addition, the presence of a template
strand should provide substantial acceleration of the reaction.
As shown in Fig. 1, when the template strand (S0, Table 1)
was absent, the coupling product (S1_1–S2) was not observed.
When the template was present, the NCL reaction proceeded
smoothly. The conjugation proceeded rapidly upon initiation
by reducing the disulfide bond, yielding substantial conju-
gated product almost immediately, regardless of whether
exogenous thiol catalysts were present or not. The conjugation
yield was between 60% to 90% based on quantitation of elec-
trophoretograms.16 This finding clearly demonstrates that tem-
plation accelerated the coupling so effectively that the catalytic
effect from exogenous thiol could not be observed. Indeed, the
exogenous thiol showed a surprising protective effect on the
conjugate, especially at higher temperature. When the reaction
mixture was incubated for a longer time at 37 °C, much more
conjugate decomposed in the thiol-free condition compared to
the mixture containing an exogenous thiol. About 50% of the
conjugate decomposed after 8 h of incubation, and less than
10% conjugate was left after 50 h of incubation. Since two
strands were conjugated through linkers, the T spacing on the
template was tested. With slight differences from 0 T to 5 T,
S1a
S1′ a
S2b
5′-TCTTGTCCAGTAACTC-3′
5′-AGTAACTC-3′
S2′ b
S1_1–S2
S1′_1–S2′
S1′_2–S2
S1′_2–S2′
5′-CAGAGAGCTGTCCTAC∼TCTTGTCCAGTAACTC-3′
5′-CAGAGAGC∼AGTAACTC-3′
5′-CAGAGAGC∼TCTTGTCCAGTAACTC-3′
5′-CAGAGAGC∼AGTAACTC-3′
a S1_1 and S1′_1 were modified with thioester at 5′ end. S1_2 and S1′_2
were modified with MPAA at 5′ end. b S2 and S2′ were modified with
tBu-S-Cys at 3′ end. Detailed structures of the modifications were
shown in Scheme 1. “∼” denotes the variant linkage as shown in
Scheme 1.
maximum at 3 T, the yields dropped significantly with longer
spacers, which imposed extra distance and thus decreased the
templating effect (Fig. 2).
MALDI-TOF MS showed that the conjugated product
obtained from the reaction with 1% thiophenol (PhSH)
present contained an extra thiophenol motif (Fig. S3A,†
S1_1–S2–PhSH), and this extra thiophenol was cleaved by treatment
with dithiothreitol (DTT) (Fig. S3B,† S1_1–S2), which indicated
that a disulfide bond was formed between the cysteine side
chain thiol and thiophenol. Systems with shorter strands (S1′_1,
S2′) and two different exogenous thiols, thiophenol (Fig. 3)
or MPAA (ESI, Fig. S4†) groups were also tested. MALDI-TOF
spectra with better resolution were obtained (about 20 Da at
5 kDa), and similar results were observed. This indicated that
an N–S acyl shift might occur with an uncapped cysteine thiol
group and lead to the decomposition of the conjugated
product (ESI, Fig. S5†).13,17 We also noticed that the decompo-
sition was slower at lower temperature (Fig. 4).
To explore further the role of exogenous thiols, experiments
were carried out under TCEP-free conditions. As expected,
Fig. 1 Templated DNA native chemical ligation reaction. (A) Schematic
of the designed strands: top, template strand, S0; bottom left, strand
containing thioester, S1-thioester (S1_1); bottom right, strand containing
cysteine, S2-Cys-S-tBu (S2). (B) Time course of templated DNA NCL:
denaturing gels developed with Stains-All (Sigma) show the NCL conju-
gate product. Gels contained 15% polyacrylamide, and were run in 1X
TBE buffer at 55 °C. The reaction solution was sampled after incubation
at 37 °C for the following times (left to right) 5 min, 1 h, 3 h, 8 h, 29 h,
50 h, and 74 h. Within each time slot, from left to right, the lanes include
Fig. 2 Templated NCL reaction with differently spaced templates. The
system consists of a (40 + n) mer template S0, 16mer S1-thioester
(S1_1), 16mer S2-Cys-S-tBu (S2), 1% v/v thiophenol, and incubated at
37 °C for 72 h. Different spaced templates are used, and n indicates the
number of T spacing used for each template. The marker lane on the left
is a 10 nucleotide ladder. The gel contains 15% polyacrylamide, run in 1X
TBE buffer at 55 °C, stained with ethidium bromide (EB).
a
10 nucleotide ladder, reaction with both a 43-mer template
(S0–43mer) and 1% v/v thiophenol, reaction with S0–43-mer but no
thiophenol, and reaction with 1% v/v thiophenol but no template. The
reaction was initialized by adding 20 mM TCEP.
8824 | Org. Biomol. Chem., 2014, 12, 8823–8827
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Communication
Fig. 5 Denaturing gel stained with EB showing the conjugate product
with thiol-disulfide exchange deprotection. From left to right, lanes
include a 10 nucleotide ladder, 3 h reaction with S0, S1_1, S2 and 1% v/v
thiophenol, 1 h reaction with S0, S1_1, S2 and 1% v/v thiophenol, 3 h
reaction with S0, S1_1, S2 and 200 mM MPAA, 1 h reaction with S0, S1_1,
S2 and 200 mM MPAA, 3 h reaction with S0, S1_1 and S2 and (no
exogenous thiol), and 1 h reaction with S0, S1_1 and S2 and (no exogen-
ous thiol).
Fig. 3 MALDI-TOF MS of conjugation product from shorter strands. (A)
Before DTT treatment: m/z calculated (S1’_1–S2’–PhSH,
5490.2, found: 5490. (B) After DTT treatment: m/z calculated (S1’_1–S2’,
M − H+): 5380.1, found: 5384. The reaction mixture was incubated with
1% v/v thiophenol at room temperature for 21 h.
M −
H+):
when there was no exogenous thiol, the capped cysteine strand
did not react with the thioester strand. This is consistent with
the accepted mechanism of NCL, wherein the transthioesterifi-
cation step is critical for NCL to proceed.15 Interestingly, with
exogenous thiol present, coupling of the two strands pro-
ceeded smoothly (Fig. 5). A thiol-disulfide exchange18 was
possibly the initial step that freed the cysteine side chain thiol
for the NCL reaction.
Inspired by successful deprotection via the thiol-disulfide
exchange reaction, a DNA-templated conjugation system was
developed based on disulfide chemistry. Instead of a thioester
motif, the S1 strand was modified with MPAA (S1_2), which
Fig. 6 Characterization of DNA-templated disulfide conjugation. (A)
Denaturing gels stained with EB showed the DSF conjugate product.
From left to right, lanes contain S0; S1_2; S2; 24 h reaction with S0,
S1_2, and S2; a 10 nucleotide ladder; 24 h reaction with S0, S1_2, and
S2’; S2’. (B) MALDI-TOF spectrum for the crude reaction mixture of S0,
S1_2, and S2. Inset: enlarged view of the mass range between 5 to
5.3 kDa. Peaks are assigned as follows: S0 m/z calculated (M − H+):
12 347, found: 12 343; S1_2–S2 m/z calculated (M − H+): 10 283, found:
10 280; S1_2 m/z calculated (M − H+): 5195.6, found: 5199; S2 m/z
calculated (M − H+): 5177.7, found: 5179. (C) MALDI-TOF spectrum for
the crude reaction mixture of S0, S1_2, and S2’. Peaks are assigned as
follows: S0 m/z calculated (M − H+): 12 347, found: 12 349; S1_2–S2’
m/z calculated (M − H+): 7869.5, found: 7871; S1_2 m/z calculated
(M − H+): 5195.6, found: 5199; S2’ m/z calculated (M − H+): 2764.1,
found: 2765.
Fig. 4 Time course of DNA-templated NCL with S0, S1_1 and S2 at
23 °C. They were all 15% acrylamide, run in 1X TBE buffer at 55 °C and
stained with EB. The reaction solution was sampled at incubation time
t = 0 (adding thiol), 5 min, 10 min, 20 min, 30 min, 1 h, 1.5 h, 3 h, 24 h,
48 h, 72 h. In each gel from left to right, the lanes show a 10 nucleotide
ladder, t = 0, 5 min, 10 min, 20 min, 30 min, 1 h, 1.5 h, 3 h, 24 h, 48 h,
72 h. The thiols added in were (A) 2% v/v thiophenol, (B) 20 mM MPAA,
(C) no exogenous thiol added.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8823–8827 | 8825

View Article Online
Communication
Organic & Biomolecular Chemistry
Acknowledgements
This research was supported by NSF grant CMMI-1120890 to
JWC and NCS and by the following grants to NCS: GM-29554
from NIGMS, CCF-1117210 from the NSF, MURI W911NF-11-1-
0024 from ARO, grants N000141110729 and N000140911118
from ONR. We acknowledge the support of DE3C0007991 from
DOE for DNA purchase and synthesis expenses.
Notes and references
Fig. 7 Denaturing gels stained with Stains-all showed the DSF conju-
gate product. From left to right, lanes include 24 h reaction with 16mer
S1_2 and 16mer S2; 24 h reaction with S1_2 and 8mer S2’; 24 h reaction
with S0, S1_2, and S2; 24 h reaction with S0, S1_2, and S2’; a 10 nucleo-
tide ladder; template strand S0; S1_2; 16mer S2; 24 h reaction with
S0, S1_2, and S2’; S2’.
1 (a) L. E. Orgel, Nature, 1992, 358, 203–209;
(b) A. Robertson, A. J. Sinclair and D. Philp, Chem. Soc.
Rev., 2000, 29, 141–152; (c) S. Liao and N. C. Seeman,
Science, 2004, 306, 2072–2074; (d) T. Wang, R. Sha,
R. Dreyfus, M. E. Leunissen, C. Maass, D. J. Pine,
P. M. Chaikin and N. C. Seeman, Nature, 2011, 478, 225–
228.
2 (a) G. von Kiedrowski, Angew. Chem., Int. Ed. Engl., 1986,
25, 932–935; (b) R. Naylor and P. T. Gilham, Biochemistry,
1966, 5, 2722–2728.
3 T. Li and K. C. Nicolaou, Nature, 1994, 369, 218–221.
4 J. T. Goodwin and D. G. Lynn, J. Am. Chem. Soc., 1992, 114,
9197–9198.
5 K. Fujimoto, S. Matsuda, N. Takahashi and I. Saito, J. Am.
Chem. Soc., 2000, 122, 5646–5647.
6 (a) A. H. El-Sagheer and T. Brown, Proc. Natl. Acad.
Sci. U. S. A., 2010, 107, 15329–15334; (b) M. Shelbourne,
X. Chen, T. Brown and A. H. El-Sagheer, Chem. Commun.,
2011, 47, 6257–6259.
7 M. L. McKee, P. J. Milnes, J. Bath, E. Stulz, R. K. O’Reilly
and A. J. Turberfield, J. Am. Chem. Soc., 2012, 134, 1446–
1449.
8 (a) Y. Liu, A. Kuzuya, R. Sha, J. Guillaume, R. Wang,
J. W. Canary and N. C. Seeman, J. Am. Chem. Soc., 2008,
130, 10882–10883; (b) M. Ye, J. Guillaume, Y. Liu, R. Sha,
R. Wang, N. C. Seeman and J. W. Canary, Chem. Sci., 2013,
4, 1319–1329.
reacted with a tert-butyl mercaptan-capped cysteine motif (tBu-
S-Cys) in an S2 strand (S2), yielding cross-linked products
(Fig. 6). Initial attempts with 16-mer S1 and 16-mer S2 were
promising, as MALDI-TOF MS of the reaction crude products
showed the target mass peak corresponding to the S1_2–S2
conjugate. However, the gel was hard to interpret because the
S1 strand dimerized with itself, yielding a 32-mer S1_2–S2 con-
jugate which had similar mobility to the 32mer S1_2–S2. To
confirm the conjugation reaction, an 8-mer S2′ was prepared
and tested. Both MALDI-TOF MS and the gel showed clearly
that the S1_2–S2′ conjugate was formed. The conjugation yield
was about 90%.
Similar to the NCL reaction, the template strand was essen-
tial for the coupling reaction, without which no observable
conjugated product formed (Fig. 7). While this manuscript was
in preparation, another report of a sulfur exchange-based DNA
coupling reaction in forming 5′-3′ disulfide linkages appeared.
Although similar chemistry was studied, a rather complicated
protocol was required to prepare the solid support for the syn-
thesis of oligonucleotide, while our system utilized only com-
mercially available materials.19
9 P. E. Dawson, T. W. Muir, I. Clarklewis and S. B. H. Kent,
Science, 1994, 266, 776–779.
10 S. B. H. Kent, Chem. Soc. Rev., 2009, 38, 338–351.
Conclusions
In conclusion, two DNA-templated conjugation systems were 11 D. A. Stetsenko and M. J. Gait, J. Org. Chem., 2000, 65,
developed based on the chemistry of thiols. One of them fol- 4900–4908.
lowed native chemical ligation, and interestingly a protective, 12 J. Michaelis, A. Maruyama and O. Seitz, Chem. Commun.,
instead of catalytic, role of exogenous thiols was observed. The 2013, 49, 618–620.
other system utilized a thiol-disulfide exchange reaction and the 13 S. W. A. Reulen, W. W. T. Brusselaars, S. Langereis,
desired cross-linked product formed without the need for any
additional reagents but the template DNA. Ultimately, these two
W. J. M. Mulder, M. Breurken and M. Merkx, Bioconjugate
Chem., 2007, 18, 590–596.
systems provide new methods to ligate together two oligonucleo- 14 (a) A. Mattes and O. Seitz, Chem. Commun., 2001, 2050–
tides via enzyme-free processes. There is considerable potential
for developing new DNA nano-materials by applying these chem-
istries, such as constructing catenanes;20 or producing asym-
2051; (b) S. Ficht, A. Mattes and O. Seitz, J. Am. Chem. Soc.,
2004, 126, 9970–9981; (c) A. Roloff and O. Seitz, Chem. Sci.,
2013, 4, 432–436.
metric three-way branched oligonucleotides,21 which can be 15 E. C. B. Johnson and S. B. H. Kent, J. Am. Chem. Soc., 2006,
processed further to assemble more complex nanostructures.
128, 6640–6646.
8826 | Org. Biomol. Chem., 2014, 12, 8823–8827
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Communication
16 X. Li, Z. J. Gartner, B. N. Tse and D. R. Liu, J. Am. Chem. 19 V. Patzke, J. S. McCaskill and G. von Kiedrowski, Angew.
Soc., 2004, 126, 5090–5092. Chem., Int. Ed., 2014, 53, 4222–4226.
17 J. Kang and D. Macmillan, Org. Biomol. Chem., 2010, 8, 20 R. Kumar, A. El-Sagheer, J. Tumpane, P. Lincoln,
1993–2002.
L. M. Wilhelmsson and T. Brown, J. Am. Chem. Soc., 2007,
129, 6859–6864.
18 (a) P. A. Fernandes and M. J. Ramos, Chem. – Eur. J., 2004,
10, 257–266; (b) A. Fava, A. Iliceto and E. Camera, J. Am. 21 M. F. Jacobsen, J. B. Ravnsbaek and K. V. Gothelf, Org.
Chem. Soc., 1957, 79, 833–838.
Biomol. Chem., 2010, 8, 50–52.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8823–8827 | 8827