Designed Intercalators for Modification of DNA Origami Surface Properties

Article abstract of DOI:10.1002/chem.201500086

The modification of the backbone properties of DNA origami nanostructures through noncovalent interactions with designed intercalators, based on acridine derivatized with side chains containing esterified fatty acids or oligo(ethylene glycol) residues is reported. Spectroscopic analyses indicate that these intercalators bind to DNA origami structures. Atomic force microscopy studies reveal that intercalator binding does not affect the structural intactness but leads to altered surface properties of the highly negatively charged nanostructures, as demonstrated by their interaction with solid mica or graphite supports. Moreover, the noncovalent interaction between the intercalators and the origami structures leads to alteration in cellular uptake, as shown by confocal microscopy studies using two different eukaryotic cell lines. Hence, the intercalator approach offers a potential means for tailoring the surface properties of DNA nanostructures. Into the fold: Designed acridine derivatives bearing esterified fatty acids or oligo-ethyleneglycol side chains (see figure) can be used to alter the surface properties of DNA origami nanostructures. Noncovalent binding of the intercalators changes the interaction of the origami with solid supports and the membrane of living cells. HOPG=highly oriented pyrolytic graphite.

Full text of DOI:10.1002/chem.201500086

DOI: 10.1002/chem.201500086
Full Paper
&
Self-Assembly
Designed Intercalators for Modification of DNA Origami Surface
Properties
Josipa Brglez, Pavel Nikolov, Alessandro Angelin, and Christof M. Niemeyer*[a]
Abstract: The modification of the backbone properties of
DNA origami nanostructures through noncovalent interac-
tions with designed intercalators, based on acridine derivat-
ized with side chains containing esterified fatty acids or oli-
go(ethylene glycol) residues is reported. Spectroscopic analy-
ses indicate that these intercalators bind to DNA origami
structures. Atomic force microscopy studies reveal that inter-
calator binding does not affect the structural intactness but
leads to altered surface properties of the highly negatively
charged nanostructures, as demonstrated by their interac-
tion with solid mica or graphite supports. Moreover, the
noncovalent interaction between the intercalators and the
origami structures leads to alteration in cellular uptake, as
shown by confocal microscopy studies using two different
eukaryotic cell lines. Hence, the intercalator approach offers
a potential means for tailoring the surface properties of DNA
nanostructures.
tings,[11] intercalators are widely applied in analytical and me-
dicinal chemistry,[12] and mechanistic insights into their mode
of action are steadily increased by modern biophysical tech-
niques, such as magnetic tweezers[13] or time-lapse AFM imag-
ing.[14] In the context of structural DNA nanotechnology, inter-
calators have been used as fluorophore units in dye arrays on
a 3D DNA-tetrahedron nanostructure,[15] or as cargo of origami-
based drug delivery vehicles for circumvention of drug resist-
ance.[16,17] Recently, Ke et al. employed a specific DNA origami
design strategy, the underwinding of double helices, to in-
crease the affinity for intercalators, and they used this en-
hanced affinity for binding a PEG-tris-acridine compound.[18]
While this work represents an elegant demonstration that non-
covalent interactions can be tuned by design of DNA nano-
structures, this approach is still in its infancy and the properties
of the resulting supramolecular assemblies have not yet been
explored. Moreover, it is yet unknown whether and how inter-
calators with different chemical properties, for example, hydro-
philicity versus lipophilicity, can be used to modify the surface
properties of DNA nanostructures.
Introduction
The so-called “scaffolded DNA origami” technique[1] employs
a long single-stranded DNA (ssDNA) scaffold which is folded
into arbitrary shape by aid of short synthetic “staple strand”
oligonucleotides. Arguably, this technique has revolutionized
our capabilities to access an unlimited variety of finite DNA
nanostructures,[2,3] which are currently being exploited, in par-
ticular, as molecular pegboards for the precise arrangement of
molecular and colloidal components. The latter can encompass
a diverse spectrum of components ranging from small mole-
cules and proteins over nanoparticles to nucleic acid-based
probes.[4–9] In spite of the variability of the origami-tethered li-
gands, the nature of the origami scaffold is largely restricted to
negatively charged (deoxy)ribonucleic acids, and thus particu-
lar aspects of applications of ligand-decorated origami con-
structs have to be adapted to the highly charged nature of the
nucleic acid scaffold.
We here report on the modification of the backbone proper-
ties of DNA nanostructures through noncovalent interactions
with designed intercalators derivatized with side chains con-
taining either esterified fatty acids or oligo(ethylene glycol) res-
idues. Intercalators are small molecule probes which can insert
between the stacked planar bases of double-stranded DNA
(dsDNA).[10] Owing to the fact that they can be used for the
control of DNA structure in bioanalytical and biomedical set-
Results and Discussion
We reasoned that the covalent tethering of an intercalating ac-
ridine unit with different side chains should lead to novel non-
covalent modifiers for DNA origami structures. To investigate
this hypothesis, we synthesized two acridine derivatives (1, 2
in Scheme 1) which contained a hexaethyleneglycol- or bis-
(hexadecyloxy)propyl-hexanoate side chain, respectively, from
the common precursor 9-chloroacridine (3; Scheme 1).
Both derivatives, hydrophilic ethyleneglycol-acridine (1), in
the following denoted as EG-Acr, and lipophilic bis-palmitoyl-
glycerol-acridine (2), denoted as L-Acr, were obtained as pure
compounds in satisfactory yields of 69 and 27%, respectively.
[a] J. Brglez, Dr. P. Nikolov, A. Angelin, Prof. Dr. C. M. Niemeyer
Karlsruhe Institute of Technology (KIT)
Institute for Biological Interfaces (IBG 1), Hermann-von-Helmholtz-Platz
76344 Eggenstein-Leopoldshafen (Germany)
E-mail: niemeyer@kit.edu
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201500086.
Chem. Eur. J. 2015, 21, 9440 – 9446
9440
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
binding modes suggesting at
least two different types of
DNA–intercalator
complexes.
Indeed, the high lipophilicity of
L-Acr led to low solubility in
aqueous media and dynamic
light scattering (DLS) measure-
ments indicated the formation
of micelles at concentrations
greater
than
approximately
0.4 mm (Figure S1 in the Support-
ing Information). Hence, we
reason that the multiple binding
types are associated with forma-
tion of higher-order structures.
Although limited amounts of
material precluded the determi-
nation of binding constants of
EG-Acr and L-Acr for DNA origa-
mi nanostructures, fluorometric
analysis provided clear evidence,
that both ligands indeed bind
the origami nanostructure (Fig-
ure S2
in
the
Supporting
Information).
We then investigated whether
binding of EG-Acr or L-Acr leads
to changes in the morphology
and affects adsorption behavior
of DNA origami structures. To
Scheme 1. Synthetic route to acridine-based origami ligands EG-Acr (1) and L-Acr (2); a) MsCl, TEA in dry DCM, 0–
48C, 24 h; NaN3, DMF, 708C, 3 h; b) PPh3 in dry THF, 08C, 10 h, then H2O, 10 h; c) 6+3, THF, 608C, 12 h; d) phenol,
1008C, 80 min, then 7, 1008C, 2 h; e) 8+9, DCC/DMAP, DMF, RT, 3 h.
The detailed description of experimental procedures and char-
acterization of compounds is given in the Supporting
Information.
this end, a rectangular DNA origami structure of approximately
91”54 nm2 was assembled using the circular single-stranded
5438 nucleotide plasmid ss109Z5.[23] AFM images of these ori-
gami plates adsorbed onto Mg2+-activated mica revealed that
the nanostructures were formed in almost quantitative yields
(left image in Figure 2A). Treatment of the origami structures
with lipophilic L-Acr or hydrophilic EG-Acr affected their capa-
bility to adsorb to the hydrophilic mica surface (middle and
right image in Figure 2A, respectively). Lowered amounts of
EG-Acr–, and in particular of L-Acr–origami, were adsorbed to
mica, thereby providing evidence for altered surface properties
of the noncovalently modified DNA nanostructures. Statistical
analysis of the AFM images indicated that lateral (x,y) dimen-
sions of the origami were not significantly altered upon bind-
ing of L- or EG-Acr ligands (Figure 2B).
However, height analysis revealed significant differences be-
tween unmodified und intercalator-modified origami struc-
tures. In particular, we determined a typical average height of
planar 2D origami of approximately 1 nm, which was lower
than the theoretical value 2 nm, due to typical flattening of
biomolecules by the AFM tip.[24] The height of unmodified ori-
gami was increased by a factor of 2.5 or 3 in the case of EG-
and L-Acr binding, respectively. Since model calculations indi-
cate a length of about 2.3 and 2.8 nm for EG- and L-Acr, re-
spectively (Figure S3 in the Supporting Information), the AFM
data suggest dense arrangements of the Acr-ligands protrud-
ing from both sides of the origami.
In a first set of experiments, binding capabilities of EG-Acr
and L-Acr for regular dsDNA was investigated through spectro-
scopic equilibrium-binding titration analyses to determine af-
finity constants (Figure 1). To this end, EG-Acr and L-Acr ligand
solutions were prepared in buffer and in DMSO, respectively,
and progressive absorbance changes were monitored for free
ligands and ligand–DNA complexes, upon addition of increas-
ing amounts of dsDNA (Figure 1A).
Indeed, addition of dsDNA led to decrease in absorbance
and the observed changes were converted into binding iso-
therms (Figure 1B) using a neighbor-exclusion model, allowing
the estimation of the dissociation constants for the ligands in-
teracting with dsDNA.[19] Dissociation constants were deter-
mined as 59 and 44 mm for EG-Acr and L-Acr, respectively,
which are close to data reported for structurally similar com-
pounds, 9-aminoacridine[20] or methylene blue.[21] Since typical
binding constants for intercalation complexes between similar
aromatic compounds and dsDNA range from 104 to 106 mÀ1,[22]
the data measured for L-Acr and EG-Acr, provided initial evi-
dence for intercalation of the acridine unit within the double
helix. Continuous variation binding analysis (Job’s plot) was
performed to determine the stoichiometry of the binding (Fig-
ure 1C). While a single maximum was observed in the Job’s
plot of EG-Acr, the curve obtained for L-Acr revealed multiple
Chem. Eur. J. 2015, 21, 9440 – 9446
www.chemeurj.org
9441
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
biotinylated staple strands (right
image, in Figure 2C). To explain
this phenomenon, we reason
that the STV proteins reduce the
overall negative charge of the
DNA origami, thereby stabilizing
the integrity of the structures. It
is evident, however, that more
detailed investigations are nec-
essary to unravel mechanistic
details of DNA nanostructure–
surface adsorption processes.
To further explore how inter-
calation affects the origami’s sur-
face properties, we studied their
uptake by eukaryotic cells. Two
cell lines, HeLa and L929, were
used in these experiments. In
addition to unmodified, L-Acr- or
EG-Acr-modified origami struc-
tures, an origami construct bear-
ing ten cholesterol ligands was
used as control. All origami
structures were labeled with
nine Cy5 molecules to enable
their fluorescent detection by
confocal laser scanning micros-
copy (CLSM). In agreement with
earlier studies,[16,17,26,27] DNA ori-
gami structures were stable in
cell culture medium for at least
24 h (Figure S4 in the Support-
ing Information). The presence
of the various origami samples
(290 pm final concentration) in
Figure 1. Spectroscopic characterization of EG-Acr (1; left panel) and L-Acr (2; right panel). A) Representative spec-
tra of the titration of acridine ligands with increasing amounts of dsDNA led to progressive decrease in absorb-
ance. B) The observed changes in absorbance (dsDNA/Deapp ”[m2 cm]) were plotted against dsDNA concentration
to generate binding isotherms for determination of affinity constants. C) Continuous variation binding analysis
(Job’s plot), used to determine binding stoichiometry, indicates a single maximum for EG-Acr (left panel), while
the curve obtained for L-Acr revealed multiple binding modes, suggesting at least two different types of DNA–in-
tercalator complexes (right panel).
To follow up on the hypothesis that EG- and L-Acr binding
affects the origami’s surface properties and thus its adsorption
behavior, we also investigated physisorption to the very hydro-
phobic substrate HOPG (highly-oriented pyrolytic graphite),
which is often used for scanning-tunneling microscopy studies
on the self-assembly of lipophilic small molecules.[25] Unmodi-
fied origami adsorbed to HOPG only in the presence of low
concentrations of detergents (Tween 20, 0.05% v/v) and the
adsorption to HOPG led to severe damage of the nanostruc-
tures (left image, in Figure 2C). We hypothesize that the dis-
ruption of the folded structure results from favored van der
Waals and stacking interactions between the DNA nucleobases
and the HOPG surface which induce dehybridization of double
helices and an unfolding of the nanostructures. While no ad-
sorption could be achieved for EG-Acr-modified origami, the L-
Acr-modified origami adsorbed to a significantly larger extent
on the HOPG (Figure 2C, middle). In this case, more intact ori-
gami structures were observed, suggesting that intercalation
of L-Acr increased the binding capability and also stabilized
the folding of the nanostructures. Interestingly, stabilization of
the origami structure was also observed when streptavidin
(STV) molecules were bound to one side of the origami using
culture medium had no harmful effects on the cells (Figure S5
in the Supporting Information). In particular, the viability of
L929 cells exposed to DNA origami was not changed during
a period of four days. Moreover, cells proliferated and their
viability was not affected by the DNA nanostructures. Further
analysis of cells was conducted by CLSM microscopy
(Figure 3).
The results showed clear evidence that loading of origami
with both L-Acr or EG-Acr led to an increased cell uptake, as
compared to unmodified origami. In fact, intercalation-based,
noncovalent modification was as effective as covalent modifi-
cation with cholesterol (Figure 3B). Notably, the increased
uptake of L-Acr- or EG-Acr-modified origami was consistently
observed for both L929 and HeLa cell lines (Figure 3B and C,
respectively, see also Figure S6 in the Supporting Information).
The z-stack reconstruction of 3D confocal images (Figure 3D
and Figure S7 in the Supporting Information) provided clear
evidence that origami was internalized by the cells. Further-
more, the combined MitoTracker Green staining showed a pat-
tern of partial co-localization of DNA origami with the mito-
chondrial compartment, confirming the internalization of both
unmodified and ligand-modified nanostructures (Figure S6).
Chem. Eur. J. 2015, 21, 9440 – 9446
www.chemeurj.org
9442
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Experimental Section
Synthesis of methoxyhexaethylene glycol amine (6)
Hexaethylenamine 6 was synthesized according to established pro-
cedures.[28,29] In brief, a solution of anhydrous HEG 4 (500 mg,
1.68 mmol) and TEA (0.28 mL, 2 mmol) in dry CH2Cl2 (30 cm3) was
cooled to 08C. A solution of methanesulfonyl chloride (0.15 mL,
2.02 mmol) in CH2Cl2 (3 mL) was slowly added under stirring and
the reaction was then maintained at 48C for 24 h. The solvent was
evaporated and the crude material was purified by silica gel
column chromatography using 0–15% gradient of MeOH in DCM
to give monomesylate (89%). The product (TLC in DCM/MeOH 9:1,
Rf =0.42) was used without further purification. Sodium azide
(182 mg, 2.8 mmol) was added to a solution of monomesylate
(559 mg, 1.4 mmol) in dry DMF (2 mL). The mixture was heated at
708C for 2.5 h and then allowed to cool to room temperature.
DMF (3 mL) was added to improve azide solubility. The reaction
mixture was co-evaporated with toluene at 508C, and the residue
was purified by chromatography using a 0–15% gradient of MeOH
in DCM giving 417 mg (92%) of methoxyhexaethylene glycol azide
5. ESI-MS: calcd for C13H27N3O6 [M+H]+ 322.19; found [M+H]+ 322,
[M+Na]+ 344.13. A solution of 5 (200 mg, 0.62 mmol) in dry THF
(2 mL) was cooled to 08C. Triphenyl phosphine was added
(184 mg, 0.69 mmol) and the mixture was allowed to reach room
temperature. The reaction was monitored by TLC (iPrOH/aqueous
NH3 (0.05%)/H2O 6:3:1). After 10 h water was added (30 mL) to hy-
drolyze the intermediate phosphorus adduct and the mixture was
stirred, overnight (<10 h). The reaction mixture was diluted with
water and washed with toluene. Evaporation of the aqueous layer
yielded 154 mg (84%) of methoxyhexaethylene glycolamine com-
pound 6, which was used in the next step without further purifica-
tion. ESI-MS: calcd for C13H29NO6 [M+H]+ 296.2; found [M+H]+
296.2. 1H NMR ([D]CDCl3): d=3.36 (s, 3H), 2.87 (t, 2H), 3.51–3.55
(m, 4H), 3.64 ppm (s, 18H); 13C NMR: d=41.6, 58.9, 70.2, 70.5, 71.8,
72.8 ppm.
Figure 2. AFM analyses of DNA origami structures. A) Representative AFM
images of unmodified DNA origami (left), L-Acr- (middle) or EG-Acr-modified
origami (right) adsorbed to hydrophilic mica substrates; scale bar: 100 nm.
Note that unmodified origami binds to the surface in high density, while
low and medium density surface coverages are observable for L-Acr- or EG-
Acr-modified origami. B) Statistical analysis of lateral dimensions and height
of the origami structures (Nꢀ100). C) Adsorption of unmodified DNA origa-
mi (left), L-Acr- (middle) and STV-modified (right) origami an hydrophobic
HOPG substrates. Note that L-Acr or STV modification appears to stabilize
the integrity of the DNA nanostructures.
Synthesis of ethyleneglycol-acridine (EG-Acr; 1)
The 9-chloroacridine 3 (Sigma–Aldrich, 226 mg, 1.05 mmol) was
dissolved in dry THF (2 mL) under Ar. Compound 6 (155 mg;
0.52 mmol) was dissolved in dry THF (3 mL) and then mixed with
the solution of 3. After stirring overnight at 608C, the solvent was
evaporated and the product was purified by silica gel chromatog-
raphy using 2–18% gradient of MeOH in DCM to obtain 1 in 69%
yield. ESI-MS: calcd for C26H36N2O6 [M+H]+ 473.26; found [M+H]+
Conclusion
In conclusion, we here demonstrate that designed intercalators
can be used for noncovalent modification of origami struc-
tures. In particular, we found that acridine derivatives
equipped with esterified lipid acids or ethyleneglycol side
chains bind to DNA origami structures and induce significantly
changed surface properties of the nanostructures, as docu-
mented by their interaction with solid supports (mica, HOPG)
and cellular membranes. Despite this clear experimental evi-
dence, the detailed mechanistic modes of action remain pres-
ently unclear and further studies are required to unravel details
of DNA nanostructure–surface interaction. Nonetheless, the re-
sults of this study suggest that the intercalator approach has
a high potential for tailoring surface properties of DNA nano-
structures in a noncovalent and thus transient fashion. We
therefore anticipate that this strategy might prove useful for
further exploitation of structural DNA nanotechnology.
1
473.33. H NMR ([D]CDCl3): d=3.36 (s, 3H), 3.52 (m, 3H), 3.59–3.62
(m, 15H), 3.69 (t, 2H) 3.8 (t, 2H), 7.25–7.29 (d, 2H), 7.55 (t, 2H),
8.18 (d, 2H), 8.47 (d, 2H), 9.4 (s, 1H), 14.2 ppm (s, 1H); 13C NMR:
d=48.1, 58.6, 70.1, 71.5, 119.1, 122.8, 132.6, 157.05 ppm.
Synthesis of 6-(acridin-9-ylamino)hexanoic acid (8)
The 6-(acridin-9-ylamino)hexanoic acid (8) was synthesized accord-
ing to published procedures.[30] In brief, a mixture of 9-chloroacri-
dine 3 (1 g, 4.68 mmol) and phenol (5 g, 0.05 mol) were stirred at
1008C for 80 min and then 6-aminocapronic acid (7; 676 mg,
5.15 mmol) was added. The reaction mixture was incubated at
1008C for 2 h. After cooling to RT, the solvent was removed in
vacuo and the crude product (1.5 g) was purified by silica gel flash
chromatography (Biotage, 50 g column) using 4–20% gradient of
MeOH (0.05% HOAc) in DCM in 25 column volumes (CV) to obtain
the final product 8 in 95% yield. ESI-MS calcd for C19H20N2O2
[M+H]+ 309.15; found [M+H]+ 309.2; [MÀH]À 307.2. 1H NMR
Chem. Eur. J. 2015, 21, 9440 – 9446
www.chemeurj.org
9443
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 3. Confocal laser scanning microscopy (CLSM) analysis of DNA origami uptake by L929 fibroblasts and HeLa cells. A) CLSM images of L929 cells treated
with unmodified DNA origami, L-Acr- or EG-Acr-modified origami. An origami construct bearing ten cholesterol ligands was used as control (bottom row of
images). All origami structures contained nine Cy5 molecules to enable fluorescent detection. Cell nuclei were counterstained with Hoechst 33342 (scale bars:
20 mm; conditions: 3”104 cellsmLÀ1, 2.5 nm DNA origami, one ligand per bp, 6 h incubation). B) Statistical analysis of fluorescence signals of Cy5-labeled ori-
gami. Values were calculated as mean ÆSD from at least two independent measurements. Statistical significance was evaluated by one-way ANOVA, followed
by Dunnett’s multiple comparison test (*p<0.0001). C) Statistical analysis of CLSM images obtained for uptake experiments with HeLa cells (for original data,
see Figure S5 in the Supporting Information). D) The z-stack projection of CLSM images, obtained from HeLa cells treated with L-Acr-loaded DNA origami.
([D6]DMSO): d=1.4 (m, 2H), 1.55 (m, 2H), 2.22 (t, 2H), 4.05 (t, 2H),
7.5 (t, 2H), 7.95 (m, 4H), 8.58 ppm (d, 2H); 13C NMR: d=24.12,
28.81, 33.54, 48.77, 119.25, 123.03, 125.88, 134.29, 156.68, 174.41,
214.22, 228.37 ppm.
(t, 2H), 4.12 (dd 1H), 4.22–4.26 (dd, 1H), 7.21 (t, 2H), 7.46 (t, 2H),
7.96 (d, 2H), 8.15 ppm (d, 2H); 13C NMR: d=14.10, 22.67, 24.44,
26.01, 26.36. 29.68, 31.89, 33.86, 44.86, 54.67, 63.98, 70.19, 70.63,
71.77, 76.45, 122.85, 124.3, 129.7, 132.3, 173.2 ppm.
Synthesis of bis-palmitoyl-glycerol-acridine (L-Acr; 2)
Equilibrium binding titration by UV/Vis spectroscopy
Preparation of L-Acr (2) was adopted from a known procedure.[31]
To a stirred solution of carboxylic acid 8 (22 mg, 0.065 mmol) in
1 mL anhydrous DMF, DMAP (0.85 mg, 0.017 mmol; 10% mol) and
1,2-di-O-hexadecly-sn-glycerol 9 (77 mg, 0.14 mmol, Bachem) were
subsequently added. The reaction mixture was cooled to 08C. DCC
(16 mg, 0.08 mmol) was added and the reaction mixture was
stirred for 20 min at 08C and 3 h at 208C. Precipitated urea was fil-
tered off and the filtrate was evaporated in vacuo. The residue was
dissolved in DCM and, if necessary, filtered again to remove any
additional precipitated urea. The DCM solution was washed twice
with 0.5m HCl, once with saturated NaHCO3 solution, and dried
over anhydrous MgSO4. After filtration, the solvent was removed in
vacuo and the ester 2 was purified by silica gel flash chromatogra-
phy (Biotage, 25 g column) using 10 CV of DCM to remove DMF
and then gradient of MeOH in DCM, 0–15%, 25 CV to obtain final
product 2 in 27% yield. ESI-MS calcd for C54H90N2O4 [M+H]+
To quantify binding affinity of EG-Acr (1) and L-Acr (2) toward
dsDNA, absorbance titration at different DNA concentrations was
performed keeping the concentration of the ligand constant. The
dsDNA used in these experiments was a 4761 bp bacterial plasmid
prepared as previously described.[23] Concentration of dsDNA stock
solution (8 mm, base-pair concentration) was determined by UV/
Vis measurements in TE buffer (20 mm Tris, 2 mm EDTA, 12.5 MgCl2
pH 7.6) using an online oligonucleotide calculator (http://www.ba-
sic.northwestern.edu/biotools/oligocalc.html#helpMW).
Spectro-
photometric titrations of ligands 1 and 2 with dsDNA plasmid
were performed at 258C by adding increasing amounts (1–5 mL) of
dsDNA solution [base pair concentration of 8 mm] to a freshly pre-
pared ligand solution (either 50 mm of 1 dissolved in 1”TEMg (TE
buffer supplemented with 12.5 mm MgCl2) or 50 mm of 2 in DMSO.
After addition of dsDNA, the mixture was allowed to equilibrate for
5 min before measurements were made. Representative spectra
are shown in Figure 1A. Affinity constants were determined from
changes in absorbance, according to the equation derived from
1
831.69; found 831.73 [MÀH]À 829.8. H NMR ([D]CDCl3): d=0.87 (t,
3H), 1.24 (m, 52H), 1.54 (m, 6H), 1.77 (q, 2H), 2.02 (q, 2H), 2.41 (t,
2H), 3.41 (t, 4H), 3.4–3.47 (m, 4H), 3.55 (t, 2H), 3.62 (m, 1H), 4.01
Chem. Eur. J. 2015, 21, 9440 – 9446
www.chemeurj.org
9444
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
neighbor-exclusion models, as previously reported by McGhee and
von Hippel (see below).[19,21]
the centrifugal filter was washed once with 500 mL of 1”TEMg at
2200g to equilibrate the membrane. The DNA origami solution
was concentrated by stepwise addition of 100 mL of origami solu-
tion with addition of 600 mL of 1”TEMg buffer and centrifuged for
5 min at 2200g at 48C. The concentrated solution was eventually
washed five times with 600 mL of 1”TEMg. Subsequently, the solu-
tion was collected by centrifugation for 1 min at 1000g to yield
100–200 mL of DNA origami solution. Quality control was per-
formed with AFM and agarose gel analysis using a 1.5% agarose
gel in 1”TBEMg buffer (40 mm Tris, 20 mm boric acid, 2 mm EDTA,
12.5 mm Mg acetate, pH 8.00; electrophoresis conditions: 80 V for
2 h at 48C). Final DNA origami concentrations were quantified by
qPCR and the origami structures were characterized by AFM, as de-
tailed in the Supporting Information. For loading DNA origami
structures with acridine ligands, the latter were dissolved in DMSO
to a final concentration of 1 mm. Typically, 15 mL [~59 nm] or 25 mL
[~68 nm] of DNA origami solution was incubated with 1.6 or 5 mL
of ligand to prepare samples of ligand per bp ratio of 0.2 for HeLa
cells and ligand per bp of 1 for L929 cells. Samples were incubated
with DNA origami solution for 3–4 h. Subsequently, ligand-loaded
DNA origami solutions were added to the HeLa or L929 cell cul-
tures in a final concentration of approximately 3.5 and 2.5 nm, re-
spectively, and incubated at 378C and 5% CO2, overnight or 6 h,
respectively. After incubation, cells were washed with PBS (+/+)
buffer and stained for further investigation using confocal laser
scanning microscopy (CLSM). All samples were prepared at least in
duplicate.
½DNAŠ
Deapp
1
De
1
DeKb
¼
½DNAŠþ
Deapp ¼ jea À efj; Deapp ¼ jeb À efj
The intrinsic binding constant, Kb, was calculated from a half-recip-
rocal plot of the changes in the apparent extinction coefficient, ea
of ligands versus DNA concentration. The ea was calculated from
Aobs/ligand where Aobs was the observed absorbance; ef and eb cor-
respond to the extinction coefficient for the free and fully bound
ligand, respectively. Thermodynamic dissociation constants, Kd,
were then calculated as the ratio of the intercept to slope.
Continuous variation binding analysis (Job plot) by fluores-
cence spectroscopy
To determine dsDNA–ligand binding stoichiometries, the method
of continuous variation was carried out similar as described in the
literature.[32] Equimolar stock solutions of dsDNA [50 mm base-pair
concentration] and acridine ligands [50 mm] were prepared sepa-
rately in 1”TE buffer. Then 13 samples were prepared where differ-
ent volumes of both dsDNA and ligand stock solution were mixed
to give a final volume of 200 mL. The total molar concentration of
the two binding partners (e.g., a dsDNA and ligand) was held con-
stant [50 mm], but their mole fractions were varied. To achieve
ligand mole fractions (%L) in a range of 0.05–1, the volume of L
was varied from 10–200 mL and the volume of dsDNA was varied
from 190–0 mL before mixing. Samples were allowed to equilibrate
at RT for 2–3 h before spectroscopic measurements were made. To
this end, each solution was filled in a cuvette and fluorescence in-
tensity of the ligands was measured at lEx/Em =385/418 nm (F1
values). Control samples containing ligand only and buffer in the
absence of dsDNA were measured to correct for dilution effects
(F2 values). Differences in fluorescence (DF=F2ÀF1) were plotted
against the mole fraction of acridine ligand, c(L).
Cell culture and microscopy
For cultivation of L929 fibroblasts and HeLa cells, 1 mL of an ap-
propriate DMSO stock solution was thawed at 378C for 1–2 min
and added to a tenfold volume of pre-warmed complete nutrient
growth medium containing 10% (v/v) FCS in a culture flask. After
5–6 h incubation in 5% CO2 atmosphere at 378C, the medium was
changed with 10 mL of complete nutrient growth medium (supple-
mented with 10% (v/v) FCS). After incubation, overnight, growth
medium (w/o FCS) was exchanged and cells were grown and split
after they reached 85–95% confluence. Before subculture, the cell
monolayer was washed with PBS (À/À) (Dulbecco’s Phosphate
Buffer Saline, without calcium and magnesium, Life Technologies,
Germany) using approximately half the volume of culture medium.
Prior to incubation with DNA origami, cells were seeded in m-slides
(chambered coverslip), eight wells (ibiTreat, tissue culture treated,
ibidi, Germany), at a cell density of 3”104 cellsmLÀ1 for L929 cells
and 1”104 cellsmLÀ1 for HeLa cells and pre-cultured in cell culture
medium for 6 h for HeLa cells and 48 h for L929 cells. Quan-
tum 101 (PAA)+1% penicillin/streptomycin was used as growth
medium for HeLa cells and DMEM (Invitrogen)+1% penicillin/
streptomycin+10% FCS for L929 cells. Cells were stained with
Hoechst 33342 for dual color detection of DNA uptake. To this
end, cells seeded in eight-well m-slides were washed with 400 mL
PBS (+/+) (Dulbecco’s Phosphate Buffer Saline, with calcium and
magnesium, Life Technologies, Germany). Medium (400 mL) was
added to the cells followed by 100 mL Hoechst 33342 solution
(10 mgmLÀ1 in DMSO, Invitrogen, Germany, diluted 1:1000). The
cells were incubated for 15 min at 378C and 5% CO2, washed with
PBS (+/+) for 5 min and left in PBS (+/+) for CLSM imaging. In
some cases, cells were stained with Mitotracker Green (Life Tech-
nologies, Germany) for co-localization studies prior to Hoechst
33342 staining. Before staining, cells were washed with 400 mL PBS
(+/+). Then 200 mL media was added followed by 1 mL
[50 mgmLÀ1, DMSO] Mitotracker Green (Invitrogen, Germany). Cells
were incubated for 30 min at 378C and 5% CO2 and washed (3”)
Assembly of DNA origami
Preparation of ssDNA scaffold strand was carried out as previously
described.[23] In brief, dsDNA plasmid 109Z5 (5438 bp) was trans-
formed into ssDNA by nicking with Nb.BbvCI and subsequent di-
gestion with T7 exonuclease. DNA origami was then assembled
from solutions containing the 109Z5 ssDNA scaffold strand [700–
1200 nm, in 1”TE, pH 8.2] and each of the staple strands [100 mm,
in water] in 1”TEMg in a total volume of 0.5–1 mL. The sequence
of the 109Z5 ssDNA scaffold strand and the full list of staple
strands are given in Tables S1 and S2 in the Supporting Informa-
tion, respectively. The assembly of origami structures bearing strep-
tavidin molecules (right image in Figure 2C) was achieved by ex-
changing six staples against biotinylated derivatives (Table S3 in
the Supporting Information). Origami structures used for cell
uptake studies were modified with nine Cy5-labeled staples
(Table S4 in the Supporting Information) and control origami struc-
tures bearing covalently attached cholesterol ligands were assem-
bled by using ten cholesterol-labeled staples (Table S5 in the Sup-
porting Information). The annealing was performed by decreasing
the temperature from 75 to 258C at À68CminÀ1 on a PCR cycler
(Mastercycler Pro, Eppendorf). After assembly, DNA origami struc-
tures were purified from excess staple strands using Amicon ultra-
0.5 centrifugal filter units (MWCO 100 kDa, Milipore). Prior to use,
Chem. Eur. J. 2015, 21, 9440 – 9446
www.chemeurj.org
9445
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[3] W. M. Shih, C. Lin, Curr. Opin. Struct. Biol. 2010, 20, 276.
[4] B. Saccà, C. M. Niemeyer, Angew. Chem. Int. Ed. 2012, 51, 58–66; Angew.
Chem. 2012, 124, 60–69; .
[5] A. M. Hung, H. Noh, J. N. Cha, Nanoscale 2010, 2, 2530.
[6] T. Tørring, N. V. Voigt, J. Nangreave, H. Yan, K. V. Gothelf, Chem. Soc. Rev.
2011, 40, 5636.
[7] F. C. Simmel, Curr. Opin. Biotechnol. 2012, 23, 516.
[8] A. Rajendran, M. Endo, H. Sugiyama, Angew. Chem. Int. Ed. 2012, 51,
874–890; Angew. Chem. 2012, 124, 898–915; .
with PBS (+/+). Images were acquired using an inverted confocal
microscope (LEICA confocal fluorescence microscope, TCS SP5,
Leica, Germany), equipped with Software Leica Application
Suite 2.0.2 (LAS-AF), pinhole: 1 Airy (111.5 mm), resolution: 8 bit and
20” or 60” water objective. Cy5-labeled DNA origami, Mitotracker
Green and Hoechst 33342 fluorescence images were acquired by
scanning in sequential mode, to avoid cross excitation. Laser inten-
sities did not overcome 30–40%. Confocal z-series images were ob-
tained with a 63” water objective. Cy5 excitation was achieved
with a HeNe633 (633 nm) laser source and Hoechst 33342 excita-
tion was achieved with an external UV (351 nm) laser source. Mito-
tracker Green excitation was achieved with an Argon (488 nm)
laser source. Images were processed with Leica Application suite,
v 2.3.5 imaging software. To quantify the signal intensity of Cy5-la-
beled DNA origami, the mean fluorescence intensity (MFI) of
a scanned area of 20 mm (approximately the size of one attached
cell) was measured in two independent experiments (20 cells per
experiment). Statistical analysis was performed using GraphPad
Prism 6 software. To analyze the viability of HeLa cells, live/dead
staining was performed. To this end, 1”104 cells were seeded in
eight-well m-slides and propidium iodide (Life technologies, Germa-
ny) was used to stain dead cells while CalceinAm (Life technolo-
gies, Germany) stained the viable cells. Specifically, cells were
washed 1” with PBS (+/+) after sub-culturing. Calcein-AM
[10 mm] and propidium iodide [1 mm] solutions were prepared in
fresh medium. Cells were incubated in 400 mL of these solutions
for 30 min at 378C, 5% CO2 and washed afterwards 3”1 min with
PBS (+/+). Fresh medium was then added and cells were analyzed
by inverted fluorescence microscopy, equipped with AxioVision
software version 4.7 (Axio Imager, Zeiss, Germany).
[9] Q. Liu, C. Song, Z. G. Wang, N. Li, B. Ding, Methods 2014, 67, 205.
[10] L. S. Lerman, J. Mol. Biol. 1961, 3, 18.
[11] A. D. Richards, A. Rodger, Chem. Soc. Rev. 2007, 36, 471.
[12] B. A. Neto, A. A. Lapis, Molecules 2009, 14, 1725.
[13] K. Gunther, M. Mertig, R. Seidel, Nucleic Acids Res. 2010, 38, 6526.
[14] L. Alonso-Sarduy, G. Longo, G. Dietler, S. Kasas, Nano Lett. 2013, 13,
5679.
[15] H. Özhalici-Ünal, B. A. Armitage, ACS Nano 2009, 3, 425.
[16] Q. Jiang, C. Song, J. Nangreave, X. Liu, L. Lin, D. Qiu, Z. G. Wang, G. Zou,
X. Liang, H. Yan, B. Ding, J. Am. Chem. Soc. 2012, 134, 13396.
[17] Y. X. Zhao, A. Shaw, X. Zeng, E. Benson, A. M. Nystrom, B. Hogberg, ACS
Nano 2012, 6, 8684.
[18] Y. Ke, G. Bellot, N. V. Voigt, E. Fradkov, W. M. Shih, Chem. Sci. 2012, 3,
2587.
[19] J. D. McGhee, P. H. von Hippel, J. Mol. Biol. 1974, 86, 469.
[20] A. D. Papaphilis, Y. H. Shaw, Biochim. Biophys. Acta 1977, 476, 122.
[21] L. Z. Zhang, G. Q. Tang, J. Photochem. Photobiol. B 2004, 74, 119.
[22] H. Ihmels, D. Otto, Top. Curr. Chem. 2005, 258, 161.
[23] M. Erkelenz, D. M. Bauer, R. Meyer, C. Gatsogiannis, S. Raunser, B. Sacca,
C. M. Niemeyer, Small 2014, 10, 73.
[24] F. Moreno-Herrero, J. Colchero, A. M. Baro, Ultramicroscopy 2003, 96,
167.
[25] S. De Feyter, F. C. De Schryver, J. Phys. Chem. B 2005, 109, 4290.
[26] Q. Mei, X. Wei, F. Su, Y. Liu, C. Youngbull, R. Johnson, S. Lindsay, H. Yan,
D. Meldrum, Nano Lett. 2011, 11, 1477.
[27] J. Hahn, S. F. Wickham, W. M. Shih, S. D. Perrault, ACS Nano 2014, 8,
8765.
[28] S. Svedhem, C. A. Hollander, J. Shi, P. Konradsson, B. Liedberg, S. C.
Svensson, J. Org. Chem. 2001, 66, 4494.
Acknowledgements
This work was supported by Deutsche Forschungsgemein-
schaft (NI 399/10). We thank Barbara Sacca for help with the
DNA origami design.
[29] M. Botta, S. Quici, G. Pozzi, G. Marzanni, R. Pagliarin, S. Barra, S. Geninat-
ti Crich, Org. Biomol. Chem. 2004, 2, 570.
[30] K. Fukui, M. Morimoto, H. Segawa, K. Tanaka, T. Shimidzu, Bioconjugate
Chem. 1996, 7, 349.
[31] B. Neises, W. Steglich, Angew. Chem. Int. Ed. Engl. 1978, 17, 522; Angew.
Chem. 1978, 90, 556.
[32] T. C. Jenkins, in Drug–DNA Interaction Protocols (Ed.: K. R. Fox), Humana,
Totowa, 1997, p.195.
Keywords: DNA nanostructures · intercalators · nucleic acids ·
self-assembly · surface binding
[1] P. W. Rothemund, Nature 2006, 440, 297.
[2] J. Nangreave, D. Han, Y. Liu, H. Yan, Curr. Opin. Chem. Biol. 2010, 14,
Received: January 8, 2015
608.
Published online on May 14, 2015
Chem. Eur. J. 2015, 21, 9440 – 9446
www.chemeurj.org
9446
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim