Covalent split protein fragment-DNA hybrids generated through N-terminus-specific modification of proteins by oligonucleotides

Article abstract of DOI:10.1039/b720013g

Semisynthetic protein-DNA hybrid molecules have recently attracted much attention as valuable tools for bioanalytical chemistry and nanobiotechnology. Here we describe a synthetic method for conjugating oligonucleotides to the N-terminus of recombinant proteins. Our strategy involves the conversion of amine-terminated oligonucleotides to thioester-functionalized oligonucleotides by using a bifunctional reagent bearing an N-hydroxysuccinimide ester and benzyl thioester group, followed by native chemical ligation with proteins containing an N-terminal cysteine. We applied this technique to construct split luciferase fragment-DNA hybrid systems in which the catalytic activity of split luciferase is restored by the re-assembly of each fragment through a specific DNA-protein or DNA-DNA interaction. Split protein fragment-DNA hybrids will offer new opportunities to explore the potential of protein-DNA conjugates for various applications. The Royal Society of Chemistry 2008.

Full text of DOI:10.1039/b720013g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Covalent split protein fragment–DNA hybrids generated through
N-terminus-specific modification of proteins by oligonucleotides†
Shuji Takeda,a Shinya Tsukiji,a Hiroshi Uedaa and Teruyuki Nagamune*a,b,c
Received 2nd January 2008, Accepted 20th March 2008
First published as an Advance Article on the web 23rd April 2008
DOI: 10.1039/b720013g
Semisynthetic protein–DNA hybrid molecules have recently attracted much attention as valuable tools
for bioanalytical chemistry and nanobiotechnology. Here we describe a synthetic method for
conjugating oligonucleotides to the N-terminus of recombinant proteins. Our strategy involves the
conversion of amine-terminated oligonucleotides to thioester-functionalized oligonucleotides by using
a bifunctional reagent bearing an N-hydroxysuccinimide ester and benzyl thioester group, followed by
native chemical ligation with proteins containing an N-terminal cysteine. We applied this technique to
construct split luciferase fragment–DNA hybrid systems in which the catalytic activity of split luciferase
is restored by the re-assembly of each fragment through a specific DNA–protein or DNA–DNA
interaction. Split protein fragment–DNA hybrids will offer new opportunities to explore the potential
of protein–DNA conjugates for various applications.
crosslinkers. Although this strategy provides a general means for
Introduction
covalently attaching oligonucleotides to lysine or cysteine residues
within proteins, many proteins display multiple copies of the
targeted residues on their surfaces, resulting in the generation
of product mixtures. As a result, site-specific DNA modification
could be achieved only when the protein of interest contains a
single cysteine residue. We11 and other groups7,12,13 have recently
demonstrated that expressed protein ligation14 is applicable as
a general tool for the preparation of homogeneous protein–
DNA conjugates. The method is based on chemical ligation
of a recombinant protein containing a C-terminal a-thioester,
generated by an intein-fusion technology, with a cysteine-modified
oligonucleotide to form an amide bond. A key advantage of this
technique is that oligonucleotides can be joined specifically and
efficiently to the C-terminus of proteins under mild conditions.
However, there may be many situations in which the C-terminus of
the protein is functionally important. Therefore, the development
of an alternative strategy for conjugating oligonucleotides to the
N-terminus of proteins is also highly desirable.
We describe herein a new chemical method for ligating oligonu-
cleotides to recombinant proteins specifically at their N-terminus.
Our strategy involves the facile introduction of a thioester moi-
ety into amine-terminated oligonucleotides using a bifunctional
reagent bearing an N-hydroxysuccinimide ester and a benzyl
thioester group, and the subsequent native chemical ligation15 of
the thioester-derivatized oligonucleotides with proteins containing
an N-terminal cysteine residue (Scheme 1). Furthermore, we
applied this technique to construct split protein fragment–DNA
hybrid systems in which the re-assembly of each split protein
fragment is mediated by specific protein–DNA or DNA–DNA
interactions. Because the functional reconstitution of catalytically
(more) active split proteins occurs only upon complementary
association,16 the split protein fragment–DNA conjugates should
be capable of transducing molecular complexation events into an
enzymatic signal. Such properties will offer new opportunities to
Creating (semi-)artificial proteins with new functions that are
useful for a variety of applications is one of the major challenges in
protein engineering and nanobioscience. With significant advances
in DNA-based technologies, such as the fabrication of DNA
microarrays and 2D and 3D DNA nano-architectures,1 the
construction of DNA machines,2 and the in vitro selection of DNA
enzymes and aptamers,3 there has been considerable interest in
integrating the unique properties of DNA molecules into proteins
and enzymes. In this respect, a growing number of protein–DNA
hybrid molecules have been described.4 For example, semisynthetic
nucleases,5 biosensors for nucleic acids,6 and signal-amplifiable
“tadpoles” for sensitive detection of analytes7 were constructed by
covalently introducing oligonucleotides to proteins. It was also
demonstrated that the insertion of single-strand DNA springs
allows the preparation of enzymes of which the activity can be
controlled by mechanical tension induced by DNA hybridization.8
Additionally, protein–DNA conjugates have proved to be valu-
able as molecular scaffolds for the construction of protein
microarrays9 and nanometre-sized protein-based supramolecular
objects.10
Despite the broad utility of protein–DNA hybrid molecules,
the synthetic method for conjugating oligonucleotides to proteins
is still very limited. Most of the hybrids reported so far were
prepared by classical bioconjugation reactions using bifunctional
aDepartment of Chemistry and Biotechnology, School of Engineering, The
University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656, Japan
bDepartment of Bioengineering, School of Engineering, The University
of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656, Japan. E-mail:
nagamune@bioeng.t.u-tokyo.ac.jp; Fax: +81-3-5841-8657; Tel: +81-3-
5841-7328
cCenter for NanoBio Integration, The University of Tokyo, 7-3-1 Hongo,
Bunkyo-ku, Tokyo, 113-8656, Japan
† Electronic supplementary information (ESI) available: RP-HPLC and
SDS-PAGE analyses. See DOI: 10.1039/b720013g
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2187–2194 | 2187
©

View Article Online
Table 1 Oligonucleotides used in this study
ODN
Sequence (5ꢀ–3ꢀ)
Amino-modified site
a
TCG ACA TCA AGC
AGC ACT TCC ACG
TTT ACG CCC ACG CTT T
AAA GCG TGG GCG TAA A
GTA ATC ATG GTC ATA GCT GTT
AAC AGC TAT GAC CAT GAT TAC
AAC AGC TAT GAC CAT GAT TAC
GGA TCC TCT AGA TCG ACC TG
GGA TCC TCT AGA TCG ACC TG
5ꢀ
3ꢀ
5ꢀ
5ꢀ
5ꢀ
5ꢀ
3ꢀ
5ꢀ
3ꢀ
b
c
cꢀ-fl
5d
5dꢀ
3dꢀ
5e
3e
available amino-terminated oligonucleotides. The reagent 3 was
synthesized as shown in Scheme 1(A). In brief, the starting
material 1 was prepared according to the previous report19 and
reacted with sarcosine tert-butyl ester in dichloromethane to
yield the compound 2. Deprotection of 2 by TFA treatment
and subsequent condensation with N-hydroxysuccinimide using
HBTU and diisopropylethylamine (DIEA) in DMF afforded the
crosslinker 3.
We tested the reactivity of 3 with amino-oligonucleotides. The
oligonucleotides, ODN-a and ODN-b (Table 1), were incubated
with 3 in a 50 mM borate buffer at pH 8.5 containing 50%
DMF. The reaction was monitored by reverse-phase HPLC (RP-
HPLC) and MALDI-TOF-MS analysis. It was confirmed that the
reaction was sufficiently rapid and complete within 1 h, producing
the desired benzylthioester-terminated ODNs in quantitative
yield with no significant side reactions (see Fig. S1, ESI†). It
should be noted that the thioester functionality has stability
sufficient for the purification by RP-HPLC and subsequent
lyophilization.
Next, the obtained benzyl thioester-functionalized ODNs (TE-
ODNs) were subjected to NCL with an enhanced green fluorescent
protein that possesses a cysteine at its N-terminus (Cys-EGFP).
The Cys-EGFP was prepared as described previously.20 The
ligation reaction was monitored by SDS-PAGE under reducing
conditions. A new band corresponding to the ligation product
was clearly observed after 48 h incubation at 4 ◦C, indicating
the covalent attachment of the oligonucleotide to the protein
(Fig. 1). Ligation yields were approximately 20%. All these
results demonstrated that the N-terminus-specific modification of
Scheme 1 (A) Synthesis of thioester modification reagent 3. Reagents and
conditions: (a) sarcosine tert-butyl ester hydrochloride (1 equiv.), DIEA
(2 equiv.), CH2Cl2, rt, 4 h (2, 61%); (b) 50% TFA in CH2Cl2; (c) NHS
(1.2 equiv.), HBTU (1.2 equiv.), DIEA (1.2 equiv.) DMF, rt, 16 h (3, 26%).
(B) Strategy for the N-terminus-specific DNA modification of protein of in-
terest (POI): (a) affinity capture and purification of chitin binding domain
(CBD)-intein-POI fusion by chitin beads; (b) intein-catalyzed self-cleavage
reaction; (c) preparation of thioester-functionalized oligonucleotides using
3 (40–60 equiv. for ODN) in 10 mM borate buffer (pH 8.5) containing 50%
DMF; (d) native chemical ligation.
extend the use of protein–DNA hybrid molecules as a platform
for bioanalytical chemistry and nanobiotechnology.
Results and discussion
Development of the N-terminus-specific DNA modification method
Our strategy for developing an N-terminus-specific DNA modifi-
cation method makes use of native chemical ligation (NCL).15 To
this end, it is necessary to have access to both recombinant proteins
containing an N-terminal cysteine and oligonucleotides that have a
thioester group. The former can be obtained by several techniques
such as TEV protease-mediated digestion17 and intein-fusion
expression.18 On the other hand, to date, no method is available
for preparing thioester-appended oligonucleotides. Because solid-
phase oligonucleotide synthesis requires treatment with base for
full deprotection, the base-sensitive thioester group needs to be
incorporated into DNA in a post-synthetic manner. Therefore,
we designed a new bifunctional crosslinker 3 containing an N-
hydroxysuccinimide ester and a benzyl thioester group for the
facile introduction of the thioester functionality to commercially
Fig. 1 N-terminus-specific DNA modification. Reagents and conditions:
3.7 lM Cys-EGFP, 200 lM TE-ODN-a or -b, in ligation buffer for
48 h at 4 ◦C. Lane 1, molecular weight marker; lane 2, Cys-EGFP;
lane 3, Cys-EGFP mixed with TE-ODN-a; lane 4, Cys-EGFP mixed
with TE-ODN-b. Cys-EGFP and the ligation product ODN-a–EGFP or
ODN-b–EGFP are denoted by • and *, respectively.
2188 | Org. Biomol. Chem., 2008, 6, 2187–2194
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
proteins with DNA was accomplished by using the crosslinking
C-terminus of LucN-Zif268.24 The methionine addition resulted
in an approximately 50-fold improvement in the yield of LucN-
Zif268. It is important to note that LucN-Zif268 could be expressed
in a soluble fraction. The purity and identity of LucN-Zif268 were
confirmed by SDS-PAGE (Fig. 2(B)). Because the Zif268 domain
was obtained in an apo-form, the LucN-Zif268 was reconstituted
with zinc ion before assay.
reagent 3.21
Design of a split luciferase re-assembly system mediated by a
protein–DNA interaction
Split protein fragment complementation is now recognized as an
important tool for analyzing protein–protein interactions.16 A key
aspect of this technique is that the activity of the split protein is
regained only when a specific protein–protein interaction occurs
and induces the re-assembly of each fragment. Thus, we applied
the principle of protein fragment complementation to construct
a “semi-synthetic” split reporter system in which a protein–DNA
interaction induces an enzymatic signal.
The overall scheme of the split protein fragment–DNA hybrid
system is illustrated in Scheme 2. As a proof-of-principle study, we
chose the zinc finger protein, Zif268, a DNA-binding protein,22
and its target DNA sequence as an interaction pair. In addition,
firefly luciferase was used as a reporter protein and split at the
position between amino acids 437 and 438, a flexible hinge region
that connects the N- and C-terminal domains of the enzyme.23
The Zif268 was genetically fused to the C-terminus of the N-
terminal half of luciferase to give LucN-Zif268. The LucN-Zif268
was expressed as a fusion with Mycobacterium xenopi GyrA intein-
chitin binding domain (CBD) (IMPACT system, New England
Biolabs). Using this method, after bacterial expression, the LucN-
Zif268 can be readily affinity-purified on chitin beads and cleaved
off from the beads by incubating with dithiothreitol (DTT). In
initial experiments, we found that the yield of LucN-Zif268 was
low (approximately 300 lg per 1 L culture) due to the occurrence
of an unwanted in vivo self-cleavage reaction. Thus, according
to the protocol of IMPACT system, to suppress the in vivo self-
processing of GyrA intein, a methionine residue was added at the
Fig. 2 (A) SDS-PAGE analysis of preparation of dsODN-c–LucC.
Left, Coomassie staining image; Right, fluorescence image. Lane 1,
molecular weight marker; lane 2, Cys-LucC (12 kDa); lane 3, reaction
mixture of TE-dsODN-c and Cys-LucC. Cys-LucC and ligation product
dsODN-c–LucC are denoted by • and *, respectively. (B) SDS-PAGE
analysis of LucN-Zif268. Lane 1, molecular weight marker; lane 2,
LucN-Zif268 (60 kDa).
The C-terminal half of luciferase was also expressed by ge-
netically fusing Synechocystis sp. DnaB intein-CBD (IMPACT
system). The fusion protein was designed to expose a cysteine
residue at its N-terminus (Cys-LucC) after an intein-mediated self-
cleavage reaction.18 The Cys-LucC was bacterially expressed and
purified by chitin beads from a soluble fraction. A 5ꢀ-amino-
modified oligonucleotide containing Zif268 binding sequences,
5ꢀGCG TGG GCG3ꢀ (ODN-c), was annealed to a complementary
5ꢀ-fluorescein-labeled DNA (ODN-cꢀ-fl). The obtained double-
strand DNA (dsODN-c) was reacted with 3 as described above and
purified by gel filtration, affording thioester-functionalized TE-
dsODN-c. Subsequently, TE-dsODN-c was applied to the NCL
reaction with Cys-LucC. SDS-PAGE analysis using Coomassie
staining and fluorescence imaging clearly indicated the forma-
tion of the ligation product, dsODN-c–LucC (Fig. 2(A)). The
dsODN-c–LucC was purified to homogeneity by agarose gel
extraction.
The DNA binding activity of the zinc finger motif within LucN-
Zif268 was investigated by a gel shift assay (Fig. 3). It was clearly
observed that the electrophoretic mobility of fluorescent dsODN-
c was retarded by mixing with holo-LucN-Zif268. The fluorescence
intensity of the shifted band was increased in proportion to the
concentration of holo-LucN-Zif268. On the other hand, an apo-
form of LucN-Zif268 showed no mobility shift. These results
indicated that the zinc finger motif of LucN-Zif268 is functional
and is capable of forming a complex with dsDNA in a holo-form.
We subsequently examined the enzymatic activity of split
luciferase hybrids using D-luciferin as a luminogenic substrate
(Fig. 4). Consistent with previous reports,25 LucN-Zif268 showed
Scheme 2 (A) Schematic illustration of the plasmid constructs. (B)
Illustration of split enzyme fragment re-assembly driven by interaction
of Zif268 and dsDNA.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2187–2194 | 2189
©

View Article Online
Fig. 3 Polyacrylamide gel shift assay of LucN-Zif268 and dsODN-c.
Fluorescein-labeled dsODN-c and LucN-Zif268 complexed with dsODN-c
are denoted by • and *, respectively. Left, various concentrations
of holo-LucN-Zif298 were mixed with 10 lM dsODN-c; Right,
apo-LucN-Zif268 was mixed with 10 lM dsODN-c.
Scheme 3 (A) Schematic illustration of the plasmid constructs for com-
plementary DNA hybridization-based protein fragment complementation.
(B) Schematic illustration of split enzyme fragment complementation
induced by complementary DNA hybridization.
also feasible to control the reconstitution of a split protein
by sequence-specific DNA–DNA hybridization. This strategy is
shown in Scheme 3. The N-terminal half of split firefly luciferase
was expressed as a fusion protein with GyrA intein-CBD (IM-
PACT system). The expressed protein was purified on chitin beads
and incubated with 2-mercaptoethanesulfonic acid (MESNA) to
obtain it in a C-terminus thioester form (LucN-TE). The LucN-TE
was then modified with a cysteine-tagged single-strand ODN-5d
by expressed protein ligation as previously described, affording
LucN–ODN-5d. Several hybrids of the C-terminal half of luciferase
that have various single-strand DNA (Table 1) at their N-terminus
were also prepared as described above (ODN–LucC). As shown
in Fig. 5, both the N- and C-terminal halves of luciferase could
be successfully modified with DNA. After removing excess ODNs
by ion-exchange chromatography, size-exclusion filtration, and gel
chromatography, the ODN-modified protein fragments were used
for the subsequent reconstitution experiments as a mixture with
unmodified fragments. It should be noted that a pair of N- and
C-terminal halves of unmodified split luciferase fragment shows
no recovery of the enzymatic activity (data not shown).
Fig. 4 Split enzyme fragments complementation assay. Conditions: 4 lM
apo- or holo-LucN-Zif268, 2 lM of each dsODN-c–LucC, Cys-LucC, or
free dsODN-c, in PBS(−). Each luminescence intensity was normalized
against that of the basal LucN activity. RLU, relative light unit.
a weak remaining activity, whereas dsODN-c–LucC displayed no
activity. There was no significant difference in the activity of LucN-
Zif268 between the apo- and the holo-form. Incubation of two
hybrids, holo-LucN-Zif268 and dsODN-c–LucC, led to a two-fold
enhancement in the luminescent signal.26 The addition of either
dsODN-c or Cys-LucC to holo-LucN-Zif268 did not induce the
luciferase activation. Also, apo-LucN-Zif268 showed no significant
change in the activity in the presence of dsODN-c–LucC conjugate,
dsODN-c, or Cys-LucC. Taking all results together, we concluded
that the functional reconstitution of split luciferase occurs upon
the interaction between Zif268 and dsODN-c portions in the
present semisynthetic split protein system.
Fig. 5 SDS-PAGE analysis of preparation of ssDNA-modified luciferase
fragments. (A) Coomassie staining gel image after ligation of ODN-5d with
LucN-TE. M, molecular weight marker; -5d, ligation product. LucN-TE
and ligation product LucN–ODN-5d are denoted by · and *, respectively.
(B) Coomassie staining gel image of ssDNA-modified LucC fragments. M,
molecular weight marker. LucC and the ligation product ODN-x–LucC are
denoted by • and *, respectively.
DNA hybridization-regulated split luciferase reconstitution
By combining the terminus-specific protein–DNA conjugation
methods developed in the present study and previously,11 it was
2190 | Org. Biomol. Chem., 2008, 6, 2187–2194
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
DNA hybridization-triggered split luciferase reconstitution
was investigated by mixing LucN–ODN-5d with each of ODN–
LucC. It was observed that the catalytic activity of LucN–ODN-
5d was enhanced three-times in the presence of ODN-3dꢀ–
LucC of which the ODN sequence is fully complementary to
ODN-5dꢀ (Fig. 6).26 ODN-3e–LucC and ODN-5e–LucC containing
a non-complementary sequence showed no effect on luciferase
activity. Importantly, ODN-5dꢀ–LucC, which has a complementary
sequence attached to the N-terminus of LucC but at its 5ꢀ-end,
did not induce the functional re-assembly due to the reverse
orientation of each fragment. These results demonstrate that
the DNA sequence- and orientation-directed control of protein
function can be achieved in the split protein–DNA hybrid
systems.
Experimental
Synthesis of bifunctional reagent 3
The starting precursor pentafluorophenyl S-benzyl thiosuccinate 1
was synthesized according to the previously reported procedure.19
To a solution of 1 (0.39 g, 1 mmol) in CH2Cl2 (5 ml), DIEA
(340 ll, 2 mmol) and sarcosine tert-butyl ester hydrochloride
(0.18 g, 1 mmol) were added. The reaction solution was stirred
for 4 h at room temperature. After evaporation, the residue was
re-dissolved in CH2Cl2 and washed three times with saturated
aqueous citric acid. The organic layer was dried over anhydrous
Na2SO4, filtered, and evaporated to obtain the compound 2 as
1
yellow oil in 61% yield. H NMR: (CDCl3) d 7.20–7.29 (m, 5H),
4.12 (s, 2H), 4.03 (s, 2H), 3.08 (s, 3H), 2.96 (t, 2H), 2.77 (t, 2H),
1.45 (s, 9H). MALDI-TOF-MS (matrix: dithranol): calcd for [M
+ Na]+ = 374.14, obsd 374.07.
The compound 2 (636 mg, 1.7 mmol) was dissolved in 20 mL of
50% TFA in CH2Cl2. After stirring for 3 h at room temperature, the
solvent was removed under reduced pressure. To a solution of the
resulting intermediate in DMF (10 mL), N-hydroxysuccinimide
(230 mg, 2 mmol), DIEA (340 ll, 2 mmol), and HBTU (758 mg,
2 mmol) were added. The mixture was stirred at room temperature
for 16 h. After evaporation, the residue was dissolved in CHCl3
and washed with saturated aqueous citric acid. The organic layer
was dried over Na2SO4, filtered, and evaporated. The crude residue
was purified by column chromatography on silica gel with ethyl
1
acetate as the eluent to give 3 as a clear oil (169 mg, 26%). H-
Fig. 6 Recovery of enzymatic activity of split luciferase fragments
by re-assembly of complementary DNA hybridization. Reagents and
conditions: 1 lM LucN–ODN-5d, 1 lM of each ssDNA-modified LucC
fragments.
NMR (CDCl3): d 7.24–7.29 (m, 5H), 4.51 (s, 2H), 4.14 (s, 2H), 3.13
(s, 2H), 2.97 (t, 2H), 2.87 (s, 4H), 2.76 (t, 2H). MALDI-TOF-MS
(matrix: dithranol): calcd for [M + Na]+ = 415.09, obsd 414.98.
N-terminus-specific modification of proteins by oligonucleotides
All oligonucleotides were purchased from Sigma and lyophilized
before use. The lyophilized ODN-a (93.3 nmol) or -b (87.8 nmol)
was dissolved in 200 ll of borate buffer (50 mM sodium borate,
pH 8.5). To the solution, a solution of 3 in DMF (200 ll,
20 mM) was added. After incubation at room temperature for
3 h, the reaction mixture was applied to a NAP-5 gel filtration
column (GE Healthcare Bioscience) to obtain TE-ODNs. TE-
ODNs were analyzed by HPLC and MALDI-TOF-MS (matrix:
3-hydroxypicolinic acid). HPLC analyses were carried out on a
lBondasphere C18 column (5 micron, 150 × 3.8 mm, Waters)
eluted with 0.1 M ammonium acetate buffer (pH 7.0) containing
0–90% acetonitrile with a linear gradient over 60 min at a flow rate
of 1 mL min−1, detected at 260 nm. TE-ODN-a: tR = 14.5 min.
MALDI-TOF-MS (matrix: 3-hydroxypicolinic acid): calcd for [M
+ H]+ = 4047.8, obsd 4047.8. TE-ODN-b; tR = 14.1 min, calcd for
[M + H]+ = 4156.8, obsd 4158.3.
Conclusion
We have described a simple method for conjugating oligonu-
cleotides specifically to the N-terminus of recombinant proteins.
The technique involves the thioesterification of amine-terminated
oligonucleotides using the bifunctional reagent 3 followed by
native chemical ligation with proteins containing a cysteine at their
N-terminus. Given the commercial availability of oligonucleotides
that have an amine group at the 5ꢀ- and 3ꢀ-ends as well as internal
positions, the present method should be generally applicable to
attach proteins to any positions within DNA strands. Further-
more, as a proof-of-principle study, we have constructed two
split protein fragment–DNA re-assembly systems that allow the
functional reconstitution in response to complementary protein–
DNA or DNA–DNA interactions. We have demonstrated that
the enzymatic activity of split protein can be precisely regulated
in a sequence- and orientation-directed manner. Therefore, in
combination with recent advances in DNA-based nanotechnology
such as DNA origami, DNA architectures, and DNA-fueled
molecular machines,1,2 the semisynthetic split protein fragment–
DNA conjugate may become a platform for the construction
of new nanometre-sized functional biomaterials in which the
DNA hybridization and protein function is cooperatively coupled.
Additionally, the ability to control the re-assembly and thus
function of split protein by DNA will provide a new tool for
protein engineering and bioanalytical chemistry.
Cys-EGFP was obtained as described elsewhere.20 TE-ODN-a
or -b (4 ll, 500 lM) was mixed with Cys-EGFP (6 ll, 37 lM) in
ligation buffer (20 mM Tris-HCl, 500 mM NaCl, 1 mM EDTA,
50 mM MESNA, pH 8.5) and incubated for 48 h at 4 ◦C. The
reaction mixtures were analyzed by SDS-PAGE under reducing
conditions.
Construction of a LucN-Zif268 expression plasmid
The gene encoding an N-terminal half [1–437] of firefly luciferase
was amplified by PCR using pGEM-luc (Promega) as a template.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2187–2194 | 2191
©

View Article Online
The primer sequences were 5ꢀ-CCC GAA TTC ATA TGG AAG
ACG CCA AAA ACA TAA AGA AAG GCC C-3ꢀ (LucN-R,
NdeI) and 5ꢀ-CCC AAC CAT GGG CGG TCA ACT ATG AAG
AAG TGT TCG-3ꢀ (LucN-MCS-F, NcoI). The amplified DNA
fragment and expression vector pTWIN1 (New England Biolabs)
were digested with NdeI and NcoI, purified with a spin column
kit (Qiagen), and ligated with Ligation High (Toyobo). The
plasmid, pTWIN1-LucN-MCS, was cloned and verified by DNA
sequencing. The zinc finger domain Zif268 gene was amplified
from cDNA of human Egr-1 by PCR using two PCR primers;
5ꢀ-CCC CCC ATG GCG CGG CCG CCC ATA TGG AAC
GCC CTT ACG CTT GCC CGG TGG AG-3ꢀ (Zif268-R, NotI)
and 5ꢀ-CCC CAA CTA GTG CAT CTC CCG TGA TGC ACA
TGT CCT TCT GCC GCA AGT GGA TCT TGG TAT GCC-3ꢀ
(Zif268-Met-F, SpeI). To suppress the unwanted in vivo processing
reaction, we added a single methionine residue to the C-terminus
of Zif268. The PCR product was digested with NotI and SpeI
and subcloned into pTWIN1-LucN-MCS. The plasmid, pLucN-
Zif268-GyrA, was verified by DNA sequencing.
The soluble fraction of cell lysate was loaded onto a chitin column
washed with over 10 column volumes of wash buffer. Intein
cleavage was induced by incubating the beads with 3 bed volumes
of buffer (20 mM Tris-HCl, 500 mM NaCl, 1 mM EDTA, pH 7.0).
On-column cleavage reaction proceeded at 4 ◦C for 2 days. The
purity and concentration of obtained Cys-LucC were analyzed
with a BCA assay.
N-terminus-specific modification of Cys-LucC with dsODN-c
The 5ꢀ-amino-modified oligonucleotide (ODN-c: 5ꢀ-amino-TTT
ACG CCC ACG CTT T-3ꢀ) was annealed to the 5ꢀ-
fluorescein-labeled complementary oligonucleotide (ODN-cꢀ-fl:
5ꢀ-fluorescein-AAA GCG TGG GCG TAA A-3ꢀ). The obtained
dsODN containing a Zif268 binding sequence (dsODN-c) was
then modified with 3 in borate buffer (50 mM sodium borate,
pH 8.5). The native chemical ligation reaction procedure between
TE-dsODN-c and LucC was as that previously described.11 The re-
action mixture was then applied to agarose gel electrophoresis for
purification. The ligation product, dsODN-c–LucC, was extracted
from the gel and concentrated using a Microcon centrifuge filter kit
(Millipore). The concentration of dsODN-c–LucC was determined
by analyzing Coomassie Brilliant Blue-stained SDS-gel using the
software Scion Image (Scion Corporation).
Bacterial expression, purification, and zinc reconstitution of
LucN-Zif268
E. coli strain BL21(DE3)pLysS transformed with pLucN-Zif268-
GyrA was grown overnight at 37 ◦C in 5 ml LB broth containing
50 lg ml−1 ampicilin. Overnight cultures were added to a 250 ml
TB medium containing 50 lg ml−1 ampicilin and 34 lg ml−1
chloramphenicol. At an OD660 of 0.6, protein expression was
induced by 1 mM IPTG. Cells were cultured at 16 ◦C overnight and
collectedbycentrifugation. Thecellpelletwas resuspendedinwash
buffer (20 mM Tris-HCl, 500 mM NaCl, 1 mM EDTA, pH 8.5)
followed by sonication. After centrifugation, the soluble fraction
was loaded onto a chitin column and thoroughly washed with over
10 column volumes of wash buffer. Intein cleavage was induced
by incubating the beads with 3 bed volumes of cleavage buffer
(20 mM Tris-HCl, 500 mM NaCl, 1 mM EDTA, 20 mM DTT,
pH 8.5). On-column cleavage reaction was allowed to proceed for
24 h at 4 ◦C, affording LucN-Zif268.
The obtained LucN-Zif268 was reconstituted with zinc ion in
zinc buffer (20 mM Tris-HCl, 100 mM NaCl, 0.5 mM ZnCl2,
pH 8.0) to obtain a holo-form of LucN-Zif268. Both the apo-
and holo-forms of LucN-Zif268 were diluted in PBS(−) prior to
all assays. The protein concentration was determined by BCA
assay.
Gel shift assay
The fluorescein-labelled dsODN-c (10 lM) was mixed with various
concentrations of either an apo- or a holo-form of LucN-Zif268 in
buffer22 (20 mM Tris-HCl, 100 mM NaCl, 10% Glycerol, 0.1%
TritonX-100, 5mM DTT, 0.1mg ml−1 BSA; for the holo-form
experiments, MgCl2 and ZnSO4 were added at final concentration
of 5 mM and 20 lM, respectively), and incubated for several
minutes at room temperature. The protein–DNA complexes
were separated from free DNA by electrophoresis in a 10%
polyacrylamide gel. The fluorescence image of the gel was obtained
using a UV transilluminator.
Protein–DNA-interaction-mediated split luciferase reassembly
For split luciferase reconstitution assays, LucN-Zif268 (4 lM) was
mixed with 2 lM of each dsODN-c–LucC, Cys-LucC, or free
dsODN-c in 50 lL of PBS(−). After a 30 min incubation at room
temperature, a 50 ll of substrate solution (50 mM phosphate,
5 mM MgSO4, 3 mM DTT, 10 mM ATP, 30 lM coenzyme A,
300 lM D-luciferin, pH 7.8) was added. Luminescence intensity
was measured for 5 s with a multiwell luminometer AB-2100
(ATTO) and normalized against that of the basal LucN activity.
Plasmid construction and expression of Cys-LucC
The gene encoding a C-terminal half [438–554] of firefly luciferase
was amplified by PCR. The sequence of primers were 5ꢀ-CCT
TCC ATG GGG TTG AAG TCT TTA ATT AAA TAC AAA
GG-3ꢀ (LucC-R, NcoI) and 5ꢀ-GGG AAT TCG GAT CCT TAC
AAT TTG GAC TTT CCG CCC TTC TTG GCC-3ꢀ(LucC-F,
PstI). The amplified fragment was digested with NcoI and PstI
and subcloned into expression vector pTWIN1-MBP1 (New Eng-
land Biolabs). The plasmid, pDnaB-CRA-LucC, was cloned and
sequenced.
Plasmid construction and expression of LucN-TE
The gene encoding an N-terminus half of firefly luciferase was
amplified by PCR using pGEM-luc as a template. Two primers,
(LucN-R and LucN-F: 5ꢀ-GGG CAA CTA GTG CAT CTC CCG
TGA TGC ACC GGT CAA CTA TGA AGA AGT GTT CGT
CTT CGT CCC-3ꢀ) were used. The amplified fragment was
digested with NdeI and SpeI and subcloned into the expression
vector pTWIN1-MBP1 (New England Biolabs). The obtained
plasmid, pTWIN1-LucN-TE, was verified by DNA sequencing.
E. coli strain BL21(DE3)pLysS was transformed with pDnaB-
CRA-LucC and subjected to protein expression as described above.
2192 | Org. Biomol. Chem., 2008, 6, 2187–2194
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
The expression procedure of LucN-TE was the almost same
as that of LucN-Zif268 described above, except that the intein
self-cleavage reaction on the column and elution were induced by
thioesterification buffer (20 mM Tris-HCl, 500 mM NaCl, 1 mM
EDTA, 50 mM MESNA, pH 8.5).
6 (a) S. Howorka, S. Cheley and H. Bayley, Nat. Biotechnol., 2001, 19,
636–639; (b) A. Saghatelian, K. M. Guckian, D. A. Thayer and M. R.
Ghadiri, J. Am. Chem. Soc., 2002, 125, 344–345.
7 (a) I. Burbulis, K. Yamaguchi, A. Gordon, R. Carlson and R. Brent,
Nat. Methods, 2005, 1, 31–37; (b) I. Burbulis, K. Yamaguchi, R. Yu, O.
Resnekov and R. Brent, Nat. Methods, 2007, 4, 1011–1013.
8 B. Choi and G. Zocci, J. Am. Chem. Soc., 2006, 128, 8541–8548.
9 (a) C. M. Niemeyer, T. Sano, C. L. Smith and C. R. Cantor, Nucleic
Acids Res., 1994, 22, 5530–5539; (b) R. Wacker and C. M. Niemeyer,
ChemBioChem, 2004, 5, 453–459; (c) J. Tominaga, Y. Kemori, Y.
Tanaka, T. Maruyama, N. Kamiya and M. Goto, Chem. Commun.,
2007, 401–403.
10 (a) C. M. Niemeyer, J. Koehler and C. Wuerdemann, ChemBioChem,
2002, 3, 242–245; (b) C. M. Niemeyer, W. Bu¨rger and J. Peplies, Angew.
Chem., Int. Ed., 1998, 37, 2265–2268; (c) B. P. Duckworth, Y. Chen,
J. W. Wollack, Y. Sham, J. D. Mueller, T. A. Taton and M. D. Distefano,
Angew. Chem., Int. Ed., 2007, 46, 8819–8822.
11 (a) S. Takeda, S. Tsukiji and T. Nagamune, Bioorg. Med. Chem. Lett.,
2004, 14, 2407–2410; (b) S. Takeda, S. Tsukiji and T. Nagamune,
TetrahedronLett., 2005, 46, 2235–2238.
12 (a) M. Lovrinovic, M. Spengler, C. Deutsch and C. M. Niemeyer,
Mol. BioSyst., 2005, 1, 64–69; (b) M. Lovrinovic and C. M.
Niemeyer, Biochem. Biophys. Res. Commun., 2005, 335, 943–948; (c) M.
Lovrinovic, L. Fruk, H. Schro¨der and C. M. Niemeyer, Chem.
Commun., 2007, 353–355; (d) M. Lovrinovic and C. M. Niemeyer,
ChemBioChem, 2007, 8, 61–67.
Preparation of LucN–ODN-5d by expressed protein ligation
Cysteine-tagged single-strand ODN-5d was prepared and purified
by gel chromatography according to our previously described
procedure.11 The obtained ODN was mixed with LucN-TE and
incubated in ligation buffer at 4 ◦C for 48 h. The purification of the
ligation product was carried out by anion-exchange chromatogra-
phy on a Poros HQ (Perspective Biosystems) and concentrated by
ultra-filtration using Amicon Ultra (MWCO 30kDa) (Millipore).
Next, the solution was subjected to gel filtration chromatography
on Superdex 75 HR 10/30 (GE Healthcare Bioscience) using
PBS(−) as an eluent. The concentration of obtained LucN–ODN-
5d was determined from the Coomassie Blue-stained SDS-gel.
Preparation of a ssODN–LucC series
13 Niemeyer and coworkers also used the expressed protein ligation
method for preparing peptide nucleic acid (PNA)–protein conjugates:
M. Lovrinovic, R. Seidel, R. Wacker, H. Schroeder, O. Seitz, M.
Engelhard, R. S. Goody and C. M. Niemeyer, Chem. Commun., 2003,
822–823.
Thioesterification of single-strand oligonucleotides, NCL with
Cys-LucC, and subsequent purification was carried out as de-
scribed above to obtain ODN-3dꢀ–, ODN-5dꢀ–, ODN-3e–, and
ODN-5e–LucC. The concentration of these conjugates was deter-
mined from a Coomassie Blue-stained SDS-gel.
14 T. W. Muir, D. Sondhi and P. A. Cole, Proc. Natl. Acad. Sci. U. S. A.,
1998, 95, 6705–6710.
15 P. E. Dawson, T. W. Muir, I. Clark-Lewis and S. B. H. Kent, Science,
1994, 266, 776–779.
16 (a) T. J. Magliery, C. G. M. Wilson, W. Pan, D. Mishler, I. Ghosh,
A. D. Hamilton and L. Regan, J. Am. Chem. Soc., 2005, 127, 146–157;
(b) A. Galarneau, M. Primeau, L.-E. Trudeau and S. W. Michnick, Nat.
Biotechnol., 2002, 20, 619–622; (c) J. N. Pelletier, F.-X. Campbell-Valois
and S. W. Michnick, Proc. Natl. Acad. Sci. U. S. A., 1998, 95, 12141–
12146; (d) I. Remy and S. W. Michnick, Mol. Cell. Biol., 2004, 24,
1493–1504; (e) S. Cabantous, T. C. Terwilliger and G. S. Waldo, Nat.
Biotechnol., 2005, 23, 102–107; (f) N. Johnsson and A. Varshavsky,
Proc. Natl. Acad. Sci. U. S. A., 1994, 91, 10340–10344; (g) T. Ozawa,
M. Takeuchi, A. Kaihara, M. Sato and Y. Umezawa, Anal. Chem.,
2001, 73, 5866–5874; (h) T. Ozawa, Y. Sako, M. Sato, T. Kitamura
and Y. Umezawa, Nat. Biotechnol., 2003, 21, 287–293; (i) T. Ozawa, A.
Kaihara, M. Sato, K. Tachihara and Y. Umezawa, Anal. Chem., 2001,
73, 2516–2521; (j) A. Kaihara, Y. Kawai, M. Sato, T. Ozawa and Y.
Umezawa, Anal. Chem., 2003, 75, 4176–4181; (k) R. Paulmurugan, Y.
Umezawa and S. S. Gambhir, Proc. Natl. Acad. Sci. U. S. A., 2002,
99, 15608–15613; (l) R. Paulmurugan and S. S. Gambhir, Anal. Chem.,
2003, 75, 1584–1589; (m) R. Paulmurugan and S. S. Gambhir, Anal.
Chem., 2005, 77, 1295–1302; (n) C.-D. Hu and T. K. Kerppola, Nat.
Biotechnol., 2003, 21, 539–545; (o) M. C. Wehr, R. Laage, U. Bolz, T. M.
Fischer, S. Gru¨newald, S. Scheek, A. Bach, K. Nave and M. J. Rossner,
Nat. Methods, 2006, 3, 985–993.
Re-assembly of split luciferase mediated by DNA–DNA
hybridization
In a 96-well plate, LucN–ODN-5d (1 lM) was mixed with an
equimolar of each ssODN–LucC in 100 lL of PBS(−) and
incubated for 30 min at room temperature. A 100 ll of substrate
solution was then added and the luminescence intensity was
measured for 5 s as described above.
Acknowledgements
This work was partially supported by NEDO (New Energy and
Industrial Technology Development Organization). S. Takeda
acknowledges the JSPS Research Fellowships. We thank RIKEN
BioResource Centre for providing human Egr-1 cDNA (RDB
1198).
17 T. J. Tolbert and C.-H. Wong, Angew. Chem., Int. Ed., 2002, 41, 2171–
2174.
References and notes
18 M.-Q. Xu and T. C. Evans, Methods, 2001, 24, 257–277.
19 D. A. Stetsenko and M. J. Gait, J. Org. Chem., 2000, 65, 4900–4908.
20 B. Lee, S. Takeda, K. Nakajima, J. Noh, J. Choi, M. Hara and T.
Nagamune, Biosens. Bioelectron., 2004, 19, 1169–1174.
21 The present method has also been successfully applied to the conjuga-
tion of oligonucleotides to two other proteins, dihydrofolate reductase
(DHFR) and Renilla luciferase, in yields ranging from 10 to 40%
(Fig. S2, ESI†). The ligation yield might be improved by increasing
the concentration of recombinant proteins or by using a thiophenol
additive: P. E. Dawson, M. J. Churchill, M. R. Ghadiri and S. B. H.
Kent, J. Am. Chem. Soc., 1997, 119, 4325–4329.
22 J.-S. Kim and C. O. Pabo, Proc. Natl. Acad. Sci. U. S. A., 1998, 95,
2812–2817.
23 E. Conti, N. P. Franks and P. Brick, Structure, 1996, 4, 287–298.
24 M. W. Southworth, K. Amaya, T. C. Evans, M.-Q. Xu and F. B. Perler,
BioTechniques, 1999, 27, 110–120.
1 (a) C. M. Niemeyer and D. Blohm, Angew. Chem., Int. Ed., 1999, 38,
2865–69; (b) N. C. Seeman, Annu. Rev. Biophys. Biomol. Struct., 1998,
27, 225–248; (c) N. C. Seeman, Trends Biotechnol., 1999, 17, 437–443;
(d) U. Feldkamp and C. M. Niemeyer, Angew. Chem., Int. Ed., 2006,
45, 1856–1876; (e) P. K. W. Rothemund, Nature, 2006, 440, 297–302.
2 (a) B. Yurke, A. J. Turberfield, A. P. Millis, Jr., F. C. Simmel and J. L.
Neumann, Nature, 2000, 406, 605–608; (b) J.-S. Shin and N. A. Pierce,
J. Am. Chem. Soc., 2004, 126, 10834–10835; (c) Y. Tian, Y. He, Y. Chen,
P. Yin and C. Mao, Angew. Chem., Int. Ed., 2005, 44, 4355–4358.
3 (a) D. S. Wilson and J. W. Szostak, Annu. Rev. Biochem., 1998, 68, 611–
647; (b) R. R. Breaker, Chem. Rev., 1997, 97, 371–390; (c) S. Tsukiji, K.
Ramaswamy and H. Suga, Pure Appl. Chem., 2004, 76, 1525–1536.
4 (a) C. M. Niemeyer, Trends Biotechnol., 2002, 20, 395–401; (b) C. M.
Niemeyer, Rev. Mol. Biotechnol., 2001, 82, 47–66.
5 D. R. Corey and P. G. Schultz, Science, 1987, 238, 1401–1403.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2187–2194 | 2193
©

View Article Online
25 T. Zako, K. Ayabe, T. Aburatani, N. Kamiya, A. Kitayama, H. Ueda
and T. Nagamune, Biochim. Biophys. Acta, 2003, 1649, 183–189.
26 To date, several reports have demonstrated the functional reconstitution
of split luciferases (ref. 16j–m), and some of them achieved over
1000-fold enhancement of catalytic activity upon protein fragment
complementation: for a representative example, see ref. 16m. However,
the increase of luciferase activity in the present system was quite low
(approximately 2–4 fold). In this regard, it should be noted that the
previous experiments using split luciferase reporters are performed
only in live cells or cell lysates, which contain various endogeneous
components, such as chaperones, that assist the correct folding and
assembly of proteins. In contrast, the present experiments were carried
out in a purified in vitro system without any chaperone-like components,
which might be a reason for the low activity recovery. To our
knowledge, no comparable data, i.e. split luciferase reconstitution in
a purified system, has been reported so far. Also, the catalytic activity
enhancement upon reconstitution might be improved by optimizing the
spacer length of each fragment and/or split sites.
2194 | Org. Biomol. Chem., 2008, 6, 2187–2194
This journal is
The Royal Society of Chemistry 2008
©