Self-assembly of a DNA dodecahedron from 20 trisoligonucleotides with C3h linkers

Article abstract of DOI:10.1002/anie.200702682

(Figure Presented). Can 20 trisoligonucleotides with 20 x 3 individual sequences be programmed to self-assemble into a DNA dodecahedron? The answer is yes if one starts from a new generation of trisoligonucleotides based on C _3h-symmetric linkers with proper flexibility. The resulting dodecahedron has C₁ symmetry and may facilitate the construction of multimodular scaffolds in the future.

Full text of DOI:10.1002/anie.200702682

Communications
DOI: 10.1002/anie.200702682
DNA Nanostructures
Self-Assembly of a DNA Dodecahedron from 20 Trisoligonucleotides
with C3h Linkers**
Jan Zimmermann, Martin P. J. Cebulla, Sven Mönninghoff, and Günter von Kiedrowski*
Since Seemanꢀs pioneering work,[1a–c] DNA has been recog-
nized as a building material for programmable hollow 3D
nanoobjects. DNA nanoconstruction benefits from the struc-
tural rigidity ofshort DNA double strands, scalability, good
accessibility ofsynthetic and chemically modified DNA, and
the option for enzymatic amplification and processing. Four
strategies for the construction of nanoobjects such as
polyhedra exist so far. Strategy I is vertex-centered and goes
from noncovalent junctions to covalent objects: Noncovalent
three-way junctions are assembled from three linear oligo-
nucleotides. Each arm contains a sticky end sequence which is
hybridized to its complement in another junction and then
covalently connected by DNA ligases.[1] Strategy II is the
reverse ofstrategy I and makes use oftrisoligonucleotides, in
other words covalent junctions are noncovalently assembled
to give the target nanoobjects.[2d,e] Non-natural modes of
copying[2a] and amplifying[2b] were proposed to enable the
that reflects the basic symmetry of a virus was forseen for
strategy III,[1d] but so far this has not been achieved by any
strategy.
Previously prepared trisoligonucleotides with three differ-
ent arms were based on asymmetric linker constructs,[2a,e,f,6] so
that in principle a set of three different sequences could be
connected in three different ways. Linker scaffolds with C3h
symmetry are thought to be advantageous because all vertices
are expected to be subject to the same conformational
constraints. Moreover, diastereomeric mixtures obtained by
the utilization ofcommercially available racemic linker
amidites[6b,7] are avoided here.
Until now trisoligonucleotides with C3h linkers were
synthesized by chemical copying ofconnectivity, that is the
usage ofa 3 ’-connected trisoligonucleotide template for the
trislinking ofsuitable 5 ’-functionalized linear oligonucleoti-
des.[2a] C3h-symmetric linker molecules have also been used to
synthesize branched oligonucleotides having mixed sequence
directions[8a] or equal sequence direction but two[8b] or three[8c]
identical sequences. Very recently, a synthesis ofbranched
oligonucleotides with a C3h-symmetric linker has been
described, which is similar to our work presented here.[8d]
Scheme 1 shows the principle ofour approach. Target
trisoligonucleotides with three different sequences are assem-
bled on a DNA synthesizer by employing a trislinker amidite
orthogonally protected with 4,4’-dimethoxytrityl (DMT) and
[2c]
replication ofjunctions and nanoconstructs from the latter.
Strategy III is a face-centered approach, employing as many
oligonucleotides as there are faces on the object while each
oligonucleotide is composed ofas many segments as there are
edges surrounding the faces.[3a–d] Strategy IV first defines the
longest path through the object by connecting all vertices
using a very long DNA single strand; a set ofshorter
oligonucleotides generates suitable rigid motifs such as
double crossovers, while additional connectivities are
[4]
expressed by means ofparanemic crossover motsi.f
Recently the assembly oftriangular prisms, cubes, pentameric
and hexameric prisms, heteroprisms, and biprisms was
reported. A set ofsingle-stranded linear and cyclic DNA
building blocks was used, and in the latter case rigid organic
linker molecules were used as vertices.[5]
Herein we report on a new generation oftrisoligonucleo-
tides and their employment in benchmark experiments to
evaluate strategy II. We selected a dodecahedron, as poly-
hedra with a smaller number ofvertices have been described
already.[1–5] The feasibility of constructing a dodecahedron
[*] J. Zimmermann, M. P. J. Cebulla, S. Mönninghoff,
Prof. Dr. G. von Kiedrowski
Scheme 1. A) Trisoligonucleotide synthesis: 1) The first strand is con-
structed in the 5’!3’ direction using “reverse” 5’-nucleoside amidites.
2) Trislinker amidite 4 is coupled and subsequently detritylated. 3) The
second strand is synthesized in the 3’!5’ direction using standard 3’-
nucleoside amidites. 4) After Pd-catalyzed removal of the AOC protect-
ing group, the third strand is assembled, again in the 3’!5’ direction.
B) Synthesis of the trislinker amidite: a) 0.85 equiv allyloxycarbonyl
chloride (AOC-Cl), 0.85 equiv pyridine, THF; b) 0.85 equiv 4,4’-dime-
thoxytrityl chloride (DMT-Cl), pyridine; c) 1.3 equiv 2-cyanoethyl
N,N,N’,N’-tetraisopropylaminophosphorodiamidite, 1.3 equiv diisopro-
pylammonium tetrazolide, dichloromethane. Yields calculated after
chromatographic purification.
Lehrstuhl für Organische Chemie I, Bioorganische Chemie
Ruhr-Universität Bochum
Universitätsstrasse 150, NC 2/173, 44780 Bochum (Germany)
Fax: (+49)234-32-14355
E-mail: kiedro@ruhr-uni-bochum.de
Homepage: http://www.ruhr-uni-bochum.de/oc1
[**] This work was supported by the integrated project PACE (EU-IST-
FET). We thank M. Wüstefeld for automated DNA syntheses and
protocol development for trisoligonucleotides.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
3626
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3626 –3630

Angewandte
Chemie
Table 1: Trisoligonucleotide sequences assigned to each of the 20
vertices of the dodecahedron; connectivities as in Figure 1.
allyloxycarbonyl groups (AOC). The first strand is synthe-
sized using reverse amidites (in the 5’–3’ direction).[9a,b]
Following the coupling oftrislinker 4, which is derived from
the C3h alcohol 1, the second and third strands are subse-
quently synthesized after detritylation and Pd-catalyzed
removal ofthe AOC protecting group, respectively
(Scheme 1A).
Sequence 5’–3’
Vertex
v1
Sequence 5’–3’
Vertex
v11
CGTACTGCCTCTTTC
GCATGACGGAGAAAG
CACTAGCAGCTAGAC
CTCGAACCTGAACTG
GAGACACGGAAGATC
GGCGAAACGTTATAC
GTCTAGCTGCTAGTG
CCTCACGAATCTAGC
CAGCATATCCGCTAC
v2
v3
v4
v5
v6
v7
v8
v9
v10
GTAGCGGATATGCTG
GTATAACGTTTCGCC
GAACGAAGGATTCGG
v12
v13
v14
v15
v16
v17
v18
v19
v20
Trislinker amidite 4 is obtained from the trisalcohol 1,3,5-
tris(hydroxypropyl)benzene (1) by successive treatment with
a) allyloxycarbonyl chloride in the presence ofpyridine to
yield bisalcohol 2, then with b) 4,4’-dimethoxytrityl chloride
in pyridine to give alcohol 3, and finally with c) Bannwarthꢀs
reagent in the presence ofdiisopropylammonium tetrazolide
to give phosphoramidite 4 (Scheme 1B).
GCTAGATTCGTGAGG
CGCATATGGAAAGCC
CGGATCTTGCTAGAC
CCGAATCCTTCGTTC
GTTCAGTCTTAGCCG
GTCGCGTATCTATGG
GGCTTTCCATATGCG
CTGAGTCGCTTACAC
CTGAGCTGTCGATAC
CTTTCTCCGTCATGC
CGGCTAAGACTGAAC
GAGGTCATAAGCGAG
The set oftrisoligonucleotides was synthesized on a 1.3-
mmol scale on a commercially available polystyrene support
to which the 5’-starter nucleoside had been attached by means
[9c,12]
ofa succinic ester linkage.
After cleavage from the
GAAAGAGGCAGTACG
GTATCGACAGCTCAG
CTACTCTGATCGCAC
CTCTAGGAGATGCAG
CTCGCTTATGACCTC
GAACAGATCGCTTGG
support, the trisoligonucleotides were purified by polyacryl-
amide gel electrophoresis (PAGE), extracted, desalted by size
exclusion chromatography (NAP-25 columns), quantified by
UV spectroscopy, and characterized by MALDI-MS. Yields
were between 30 and 100 nmol after purification.
GTGCGATCAGAGTAG
CTGCATCTCCTAGAG
CATTGCTGTATTGCG
CCAAGCGATCTGTTC
GCATTCCGCAAAAAG
GTCTAAGGTCTTGCG
A regular dodecahedron consists oftwelve regular
pentagonal faces, three meeting at each vertex to give 30
edges and 20 vertices. To create a “stick model” ofa
dodecahedron, a set of20 trisoligonucleotides is required:
The centers ofthe trisoligonucleotides represent the vertices,
the edges are formed by hybridization of complementary
strands ofproximate trisoligonucleotides. A pool of30
independent 15mer double-stranded DNA sequences with
narrow melting temperatures was designed using the DNA
sequence generator developed by Feldkamp et al.[10] (con-
ditions: Tm between 52.3 and 56.28C, neutral pH, 50 mm
NaCl). Double-stranded sequences were then designated as
the edges ofthe dodecahedron, leading to a Schlegel
representation for vertex enumeration (Figure 1) and a
table oftrisoligonucleotides deifning the respective vertex
connectivities (Table 1).
CGCAATACAGCAATG
CTCGTTCAAGACAGG
CGTCTGATGTGTACG
CCATAGATACGCGAC
CTTTTTGCGGAATGC
GATCACTCTTGCGTG
GTGTAAGCGACTCAG
CCTGTCTTGAACGAG
GAACGTTAATACGGC
GATCTTCCGTGTCTC
CACGCAAGAGTGATC
GATGACTCTGTAGCG
GCCGTATTAACGTTC
CTTACGAATCCGGTG
CACACCTTACTAGCG
CGCTAGTAAGGTGTG
CGCTACAGAGTCATC
GTTGTAGAGGAAGCG
GTCTAGCAAGATCCG
CACCGGATTCGTAAG
CAGTTCAGGTTCGAG
CGTACACATCAGACG
CGCAAGACCTTAGAC
CGCTTCCTCTACAAC
For the self-assembly of the dodecahedron all 20 trisoli-
gonucleotides were combined in equimolar quantities (typical
experiment: 0.5 mm per trisoligonucleotide, 10 mm HEPES
buffer (pH 7.4), 100 mm NaCl). Best results were achieved
when submicromolar concentrations ofeach trisoligonucleo-
tide were used along with a temperature program that
provides a melting step (908C, 5 min), an annealing step at
458C (approximately 108C below the melting point ofthe
double strands) for at least 30 min, and a final cooling step
down to 48C. Native agarose gel electrophoresis ofthese
annealing mixtures showed one single discrete band with an
electrophoretic mobility comparable to that oflinear double-
stranded DNA of400 (3% gel) to 450 base pairs (2% gel);
this is in agreement with the actual number of450 base pairs
for the completely assembled dodecahedron (Figure 2, see
also the Supporting Information).
Treatment ofthe noncovalent product with mung bean
nuclease did not lead to any change ofelectrophoretic
mobility, while the digestion ofassembly subsets resulted in
such changes (Figure 2). This is to be expected because the
fully assembled dodecahedron contains only double-stranded
interconnections, while subsets necessarily contain digestible
single-stranded arms. As observed in earlier studies on a
trisoligonucleotide tetrahedron, the fully assembled product
Figure 1. Schlegel representation with numbers of trisoligonucleotides
assigned to the vertices (see Table 1). Each 3’ end is attached to
trislinker 1 by means of a phosphodiester linkage.
Angew. Chem. Int. Ed. 2008, 47, 3626 –3630
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3627

Communications
appears as a single band while fragments often show two or
more bands in the gel.[2e] We are tempted to explain the latter
as an indication ofcooperative structural integrity ofthe fully
assembled object.
Atomic force microscopy of the self-assembled product in
the liquid phase on mica showed quite uniform particles with
a diameter ofapproximately 20 nm, which is in agreement
with our expectations (Figure 3). The limited height ofabout
4 nm indicates that the dodecahedron has sufficient flexibility
towards compression and adsorption. Future applications
may benefit from the induced deformability of such “soft
balls”.
To demonstrate the usability ofthe dodecahedron as a
scaffolding device, we assembled dodecahedra by employing
one to six trisoligonucleotides in which one arm was extended
to bear an overhang sequence. These objects were then
exposed to solutions containing the corresponding comple-
mentary 5’-fluorescein-labeled oligonucleotides. Figure 4
shows the agarose gel ofthe assembled objects. As expected,
the fluorescence intensity increases with the number of labels
Figure 2. Native agarose gel (2%) of the dodecahedron and assembly
subsets a) before and b) after treatment with mung bean nuclease
(digestion of single-stranded DNA). Lane M: 100-bp DNA ladder.
Lanes 1a,b: v14–v17; lanes 2a,b: v13–v17; lanes 3a,b: v1, v2, v12–v17;
lanes 4a,b: v1, v2, v11–v18; lanes 5a,b: v1, v2, v5, v6, v11–v18;
lanes 6a,b: v1, v6, v12–v18, v20; lanes 7a,b: v1–v20.
Figure 4. Native agarose gel (3% agarose, 1”TBE buffer, 6.7 Vcmꢀ1
60 min at 58C) of dodecahedral assemblies containing different
numbers of overhang sequences hybridized with fluorescein-labeled
complementary oligonucleotides (see Table S8 in the Supporting
Information), imaged A) before and B) after treatment with SYBR gold
nucleic acid stain.
,
Figure 3. AFM image of v1–v20 showing discrete objects having a size
of about 20 nm when adsorbed on a mica surface (tapping mode,
liquid phase, 10 mm HEPES buffer, pH 7, 100 mm NaCl, 1 mm NiCl2).
3628
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3626 –3630

Angewandte
Chemie
126.07 126.13 126.46 (H-CAr), 126.95 128.10 128.17 (CAr DMT), 130.0
bound to the scaffold (Figure 4A), while the concomitant
decrease ofelectrophoretic mobility is barely detectable.
Similar experiments were carried out with recently devel-
oped[11] gold clusters as labels (see the Supporting Informa-
tion).
We have shown that the self-assembly of 3’-trisoligonu-
cleotides having a C3h-linker core generates the target
dodecahedra as the main products. Sequence overhangs
hybridize with hybrid molecules composed ofa complemen-
tary sequence tag and modular function such as a dye or a
nanoscaled cluster. Up to six positions have been successfully
addressed so far, giving reason to believe that scaffolded
multimodularity ofhighly complex assemblies is within the
scope ofthis approach. Conceivable applications ofr such
constructs are widespread and range from the trapping and
functionalization of size-matched nanoscale objects to the
construction ofmultimodular machines.
=
(C2, C2’, C6, C6’ DMT), 132.69 (CH2CH CH2), 136.51 (C1, C1’
PhOCH3 DMT), 141.09 142.03 142.52 (CH2CAr), 145.68 (C1 Ph
DMT), 154.83 (O(CO)O), 158.41 ppm (C4, C4’ PhOCH3 DMT);
MALDI-TOF MS m/z calcd for C40H46O7Na ([M+Na]+): 661.80;
found: 661.97; ESI MS m/z calcd for [M+Na]+: 661.8; found: 661.3.
4: A solution of 3 (1.00 g, 1.57 mmol) 3, diisopropylammonium
tetrazolide (0.31 g, 1.86 mmol), and 2-cyanoethyl N,N,N’,N’-tetraiso-
propylaminophosphorodiamidite (0.59 mL, 0.56 g, 1.86 mmol) in
20 mL ofabsolute dichloromethane was allowed to react at room
temperature for 2 h in an argon atmosphere. The crude product was
chromatographed on silica gel with cyclohexane/ethyl acetate (2:1,
0.5% triethylamine) as eluent to give 0.87 g (1.03 mmol, 66%) of
1
3
compound 4. H NMR: (400 MHz, CDCl3): d = 1.19–1.24 (2 d, J =
7.04 Hz, 6.84 Hz, 12H, NCH(CH3)2), 1.88–2.02 (m, 6H,
CH2CH2CH2O), 2.63–2.70 (m, 8H, CH2CH2CH2O and CH2CN),
3.13 (t, 2H, 3J = 6.32 Hz, CH2ODMT), 3.58–3.92 (m, 6H,
OCH2CH2CN and CH(CH3)2 and CH2CH2CH2OP), 3.82 (s, 6H,
3
3
OCH3), 4,18 (t, J = 6.48 Hz, 2H, CH2OAOC), 4.66 (dt, J = 5.8 Hz,
J = 1.28 Hz, 2H, CH2CH CH2), 5.27–5.42 (2 dm, 3Jtrans = 17.2 Hz,
5
=
3Jcis = 10.4 Hz, 2H, CH CH2), 5.92–6.04 (m, 1H, CH CH2), 6.81–6.87
(m, 7H, ArH), 7.20–7.48 ppm (m, 9H, ArH). 13C NMR: (100.6 MHz,
CDCl3): d = 20.69 20.76 (CH2CN), 24.90 24.97 25.04 (CH(CH3)2),
30.70 32.20 33.21 33.28 (CH2CH2O), 32.25 32.51 32.96
(CH2CH2CH2O), 43.36 43.48 (CH(CH3)2), 55.54 (OCH3), 58.61
58.80 (CH2CH2CH2OP), 63.24 (CH2ODMT), 63.33 63.50
=
=
Experimental Section
Compound 1 was prepared according to Ref. [12].
2: A solution of 1 (8.0 g, 31.7 mmol) and pyridine (2.2 mL, 2.1 g,
26.5 mmol) in 40 mL ofanhydrous tetrahydrouf ran under an argon
atmosphere was stirred and cooled to 08C. A solution ofallyloxy-
carbonyl chloride (2.85 mL, 3.24 g, 26.9 mmol) in 20 mL ofdry
tetrahydrofuran was added dropwise. The reaction mixture was
allowed to warm up to room temperature while it was stirred for 3 h.
The mixture was filtered and concentrated to dryness. The crude
product was purified by flash chromatography on silica gel (cyclo-
hexane/ethyl acetate 3:1) to give 3.97 g (11.8 mmol, 44%) of
compound 2. 1H NMR: (400 MHz, [D6]DMSO): d = 1.66–1.73 (m,
4H, CH2CH2OH), 1.85–1.92 (m, 2H, CH2CH2OAOC), 2.52–2.59 (m,
6H, CH2CH2CH2OH and CH2CH2CH2OAOC), 3.41 (mc, 4H,
CH2OH), 4.08 (t, 3J = 6.48 Hz, 2H, CH2OAOC), 4.41 (t, 3J =
=
(OCH2CH2CN), 67.87 (CH2OAOC), 68.70 (CH2CH CH2), 86.10
(Cq DMT), 113.32 113.35 (C3, C3’, C5, C5’ DMT), 117.95 (CN), 119.20
(CH2CH CH2), 126.31 126.41 126.71 (H-CAr), 126.93 128.13 128.59
(CAr DMT), 130.38 (C2, C2’, C6, C6’ DMT), 132.04 (CH2CH CH2),
=
=
137.07 (C1, C1’ PhOCH3 DMT), 141.32 142.28 142.91 (CH2CAr),
145.74 (C1 Ph DMT), 155.42 (O(CO)O), 158.71 ppm (C-4, C-4’
PhOCH3 DMT). 31P NMR: (162 MHz, CDCl3, phosphoric acid): d =
148.71 ppm. ESI MS m/z calcd for [M+H]+: 840.04; found: 839.4.
Polystryrene-based solid support for reverse DNA synthesis:
Solid supports were synthesized by coupling 5’-O-succinate-3’-O-
DMT nucleosides (synthesized according to Ref. [13]) to Custom
Primer Support 200 Amino (GE Healthcare); 3’-O-succinate-5’-O-
DMT nucleosides were synthesized according to Ref. [9b,c]. Coupling
ofthe corresponding succinic acid esters to the solid support (200
mmol amino functions per gram) was achieved by combining
equimolar quantities of3 ’-O-succinate-5’-O-DMT nucleosides and
amino-modified solid support with 1.3 equivalents of HBTU and
2 equivalents oftriethylamine in anhydrous DMF (4 mL per gram of
solid support) and gentle rotation of the reaction frit for 16 h at room
temperature. Unreacted amino functions were capped by suspending
1.00 g ofthe modified solid support in a solution of25 mg ofDMAP
and 0.5 mL ofacetic acid anhydride in 7.5 mL ofpyridine. The
suspension was gently agitated by rotating the reaction frit for 16 h at
room temperature. The resulting solid supports had loadings between
50 and 60 mmol per gram.
Trisoligonucleotide synthesis: Reverse amidites (ChemGene)
were allowed to react for 5 min on a DNA synthesizer (Gene
Assembler Plus) employing 5-benzylmercaptotetrazole (emp Biotec)
as activator. Standard protocols for detritylation, capping, and
oxidation were applied for all three strands. Coupling of linker 4
was performed in two consecutive injections (5 min each). Coupling
times in the second strand were 1.5 min, while the third strand
required 3 ” 15 min for the first amidite and 3 ” 2 min for the
following. Allyl deprotection was carried out by circulating a solution
of17.1 mg ofbis(diphenylphosphino)ethane, 24.7 mg ofbis(dibenzy-
5.08 Hz, 2H, CH2OH), 4.59 (dt, 3J = 5.56 Hz, 5J = 1.5 Hz, 2H,
3
CH2CH CH2), 5.23–5.35 (2 dm, Jtrans = 17.2 Hz, 3Jcis = 10.41 Hz, 2H,
=
=
=
CH CH2), 5.93 (mc, 1H, CH CH2), 6.83 (s, 2H, ArH), 6.84 ppm (s,
1H, ArH). 13C NMR: (100.6 MHz, [D6]DMSO): d = 30.26
(CH2CH2OAOC),
31.59
(CH2CH2CH2OAOC),
32.04
(CH2CH2CH2OH), 34.80 (CH2CH2OH), 60.63 (CH2OH), 67.47
=
=
(CH2OAOC), 68.14 (CH2CH CH2), 118.69 (CH2CH CH2), 126.08
=
126.48 (H-CAr), 132.72 (CH2CH CH2), 141.15 142.55 (CH2CAr),
154.83 ppm (O(CO)O); MALDI-TOF MS m/z calcd for
C19H28O5Na ([M+Na]+): 359.42; found: 359.43; ESI MS m/z calcd
for [M+H]+: 337.44; found: 337.1.
3: A solution of4,4 ’-dimethoxytrityl chloride (3.38 g, 10 mmol) in
20 mL ofanhydrous pyridine was added dropwise to a solution of 2
(3.97 g, 11.8 mmol) in 30 mL ofanhydrous pyridine over a period of
30 min. The reaction mixture was stirred for 16 h at room temper-
ature. The crude product was chromatographed on silica gel with
cyclohexane/ethyl acetate (2:1, 0.5% triethylamine) as eluent to give
2.57 g ofcompound
3
(4.0 mmol, 40%). 1H NMR: (400 MHz,
[D6]DMSO): d = 1.62–1.71 (m, 2H, CH2CH2OH), 1.75–1.90 (m, 4H,
CH2CH2O), 2.47–2.67 (m, 6H, CH2CH2CH2), 2.97 (t, 3J = 6.2 Hz, 2H,
CH2ODMT), 3.40 (m, 2H, CH2OH), 3.73 (s, 6H, OCH3), 4.04 (t, 3J =
3
7.0 Hz, 2H, CH2OAOC), 4.41 (t, J = 5.0 Hz, 1H, CH2OH), 4.60 (dt,
3J = 5.56 Hz, 5J = 1.52 Hz, 2H, CH2CH CH2), 5.21–5.36 (2 dm,
=
3Jtrans = 17.3 Hz, 3Jcis = 10.36 Hz, 2H, CH CH2), 5.93 (mc, 1H, CH
CH2), 6.74–7.39 ppm (m, 16H, ArH). 13C NMR: (100.6 MHz,
[D6]DMSO): d = 30.24 (CH2CH2OAOC), 31.56 (CH2CH2ODMT),
31.59 (CH2CH2CH2OAOC), 32.02 32.23 (CH2CH2CH2O), 34.76
(CH2CH2OH), 55.43 and 55.40 (OCH3), 60.63 (CH2OH), 62.52
=
=
lideneacetone)palladium (Acros organics), and 10.7 mL ofpyrrolidine
ꢀ1
in 10.0 mL ofacetonitrile for 15 min at 0.5 mLmin
.
Trisoligonucleotide purification and characterization by MALDI-
TOF MS: After cleavage from solid support (conc. ammonia, 558C,
16 h) products were purified by preparative denaturing PAGE, 12%
(450 V, 4 h, standard TBE buffer), extracted from the gel, and
=
(CH2ODMT), 67.44 (CH2OAOC), 68.11 (CH2CH CH2), 85.65 (Cq
=
DMT), 113.32 (C-3, C-3’, C-5, C-5’ DMT), 118.64 (CH2CH CH2),
Angew. Chem. Int. Ed. 2008, 47, 3626 –3630
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 3629

Communications
desalted using NAP25 columns (Amersham Biosciences). MALDI
MS analysis (Bruker Daltonics) was performed using 3-hydroxypico-
linic acid (3-HPA) as the matrix.
[3] a) R. P. Goodman, I. A. T. Schaap, C. F. Tardin, C. M. Erben,
R. M. Berry, C. F. Schmidt, A. J. Turberfield, Science 2005, 310,
1661 – 1665; b) R. P. Goodman, R. M. Berry, A. J. Turberfield,
Chem. Commun. 2004, 1372 – 1373; c) C. M. Erben, R. P. Good-
man, A. J. Turberfield, Angew. Chem. 2006, 118, 7574 – 7577;
Angew. Chem. Int. Ed. 2006, 45, 7414 – 7417; d) C. M. Erben,
R. P. Goodman, A. J. Turberfield, J. Am. Chem. Soc. 2007, 129,
6992 – 6993.
[4] W. M. Shih, J. D. Quispe, G. F. Joyce, Nature 2004, 427, 618 – 621.
[5] F. A. Aldaye, H. F. Sleiman, J. Am. Chem. Soc. 2007, 129, 13376 –
13377.
[6] a) M. Chandra, S. Keller, C. Gloeckner, B. Bornemann, A. Marx,
Chem. Eur. J. 2007, 13, 3558 – 3564; b) Asymmetric branching
amidites are commercially available for example, at Glen
Research (asymmetric doubler amidite, 5-Me-dC Brancher
Phosphoramidite) or Biosearch Technologies (5’-DMT-T C6-
O-Lev).
Self-assembly studies: Trisoligonucleotides were combined stoi-
chiometrically in HEPES buffer (10 mm, pH 7.4, 100 mm NaCl) to a
final concentration of 0.5 mm. This solution was heated to 908C
(5 min), cooled to 458C (0.18Csꢀ1, 40 min) and finally cooled to 48C
(0.18Csꢀ1). Digestion ofsingle-stranded DNA with mung bean
nuclease (Promega) was performed for 30 min at 308C in acetate
buffer (50 mm, pH 5, 1 mm Zn2+) applying 10 units ofmung bean
nuclease per mg ofDNA. AFM images (liquid phase, tapping mode)
were recorded with silicon nitride tips (AU NM-10, tip radius 10 nm,
Veeco) using a Solver Pro AFM (NT-MDT) equipped with an
AU028NTF adjustment unit and processed using the Image Analysis
2.2.0 software. DNA nanostructures were fixed to mica in 10 mm
HEPES, pH 7.4, 1 mm NiCl2 and 100 mm NaCl.
[7] M. S. Shchepinov, E. M. Southern, Russ. J. Bioorg. Chem. 1998,
24, 794.
Received: June 19, 2007
Revised: February 11, 2008
Published online: March 28, 2008
[8] a) K. V. Gothelf, A. Thomsen, N. Nielsen, E. Cló, R. S. Brown, J.
Am. Chem. Soc. 2004, 126, 1044 – 1046; b) T. Kuroda, Y. Sakurai,
Y. Suzuki, A. O. Nakamura, M. Kuwahara, H. Ozaki, H. Sawai,
Chem. Asian J. 2006, 1, 575 – 580; c) F. A. Aldaye, H. F. Sleiman,
J. Am. Chem. Soc. 2007, 129, 10070 – 10071; d) J. Tumpane, P.
Sandin, R. Kumar, V. E. C. Powers, E. P. Lundberg, N. Gale, P.
Baglioni, J.-M. Lehn, B. Albinsson, P. Lincoln, L. M. Wilhelms-
son, T. Brown, B. NordØn, Chem. Phys. Lett. 2007, 440, 125 – 129.
[9] a) M. Koga, M. F. Moore, S. L. Beaucage, J. Org. Chem. 1991, 56,
3757 – 3759; b) M. Koga, A. Wilk, M. F. Moore, C. L. Scremin, L.
Zhou, S. L. Beaucage, J. Org. Chem. 1995, 60, 1520 – 1530;
c) M. D. Matteucci, M. H. Caruthers, Tetrahedron Lett. 1980, 21,
3243 – 3246.
[10] U. Feldkamp, S. Saghafi, W. Banzhaf, H. Rauhe, “DNA Se-
quence Generator: A Program for the Construction of DNA Se-
quences”, can be found under http://www.cs.mun.ca/ ~ banzhaf/
molcomp.html#DNASequenceGenerator.
[11] W. M. Pankau, S. Mꢁnninghoff, G. von Kiedrowski, Angew.
Chem. 2006, 118, 1923 – 1926; Angew. Chem. Int. Ed. 2006, 45,
1889 – 1891.
Keywords: DNA structures · nanostructures · self-assembly ·
.
supramolecular chemistry · trisoligonucleotides
[1] a) N. C. Seeman, J. Theor. Biol. 1982, 99, 237 – 247; b) J. H. Chen,
N. C. Seeman, Nature 1991, 350, 631- 633; c) Y. W. Zhang, N. C.
Seeman, J. Am. Chem. Soc. 1994, 116, 1661 – 1669; d) N. C.
Seeman, Angew. Chem. 1998, 110, 3408 – 3428; Angew. Chem.
Int. Ed. 1998, 37, 3220 – 3238.
[2] a) L. H. Eckardt, K. Naumann, W. M. Pankau, M. Rein, M.
Schweitzer, N. Windhab, G. von Kiedrowski, Nature 2002, 420,
286; b) A. Luther, R. Brandsch, G. von Kiedrowski, Nature 1998,
396, 245 – 248; c) G. von Kiedrowski, L.-H. Eckardt, K. Nau-
mann, W. M. Pankau, M. Reimold, M. Rein, Pure Appl. Chem.
2003, 75, 609 – 619; d) M. Scheffler, A. Dorenbeck, S. Jordan, M.
Wüstefeld, G. von Kiedrowski, Angew. Chem. 1999, 111, 3513 –
3518; Angew. Chem. Int. Ed. 1999, 38, 3311 – 3315; e) A.
Dorenbeck, PhD Thesis, Ruhr-Universität Bochum 2000;
http://www-brs.ub.ruhr-uni-bochum.de/netahtml/HSS/Diss/Dor-
enbeckAxel/diss.pdf; f) L. H. Eckardt, PhD Thesis, Ruhr-Uni-
versität Bochum, 2006; http://www-brs.ub.ruhr-uni-bochum.de/
netahtml/HSS/Diss/EckardtLarsHenning/diss.pdf.
[12] Y. Yamagiwa, Y. Koreishi, S. Kiyozumi, M. Kobayashi, T.
Kamikawa, M. Tsukino, H. Goi, M. Yamamoto, M. Munakata,
Bull. Chem. Soc. Jpn. 1996, 69, 3317 – 3323.
[13] H. Roesch, A. Froehlich, J. F. Ramalho-Ortigao, M. Montenarh,
H. Seliger, Eur. Pat. EP 464638, 1992.
3630
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3626 –3630