Full Paper
cleotides can be captured and released independent on their
length. Exceeding this limit, the double stranded DNA can be
trapped in the supramolecular complex, but photoinduced dis-
sociation does not occur.
Supporting Information). The highly negative surface potential
of bare CDV (À26 mV) increased to +3.2 mV when the vesicles
were decorated with 3, which is less positive than the surface
potential previously detected for CDV covered with linker 2’
(+11 mV). In contrast to 2’, linker 3 is lacking the tetra(ethyle-
neglycol) spacer. It is reasonable to assume that the spacer
allows the spermine groups of 2’ to reach away from the CDV
so that they are exposed to the aqueous medium and accessi-
ble for DNA. Evidently, the spacer causes a higher z-potential
and enables stronger interaction with the negatively charged
binding partner. The lack of a spacer in 3 prevents the effective
display of positive charges and the negative surface potential
of the CDV prevails, leading to repulsion of the polyanionic
DNA. Additionally, it should be considered that without the
spacer, the binding sites for DNA are displayed rather rigidly
and close to the vesicle surface. It can be assumed that this
allows binding of ssDNA only, since this target is shorter and
more flexible than the dsDNA.
These results can be explained by the fact that host–guest
complexation with CDV is only possible with trans-azobenzene.
Irradiation with UV-light causes photoisomerization into the
cis-isomer, which is not a suitable guest for CDV and therefore
dissociates from the vesicles. This leads to the above-men-
tioned loss of multivalency and thus to release of the ssDNA.
Nevertheless, a small amount of trans-azobenzene is still pres-
ent in the photostationary state. Although this amount is not
sufficient to effectively bind the shorter oligonucleotides, the
larger dsDNA can still complex several linker molecules on the
same vesicle, forming a stable multivalent complex. Our find-
ings thereby illustrate the crucial role of multivalency in this
system.
To get deeper insight into the system and optimize com-
plexation, we then focused our interest on the influence of the
linker molecule. Having shown that the aggregation and re-
lease of the tested single-stranded DNAs is comparable, we se-
lected 50-mer ssDNA and dsDNA for that purpose. Figure 2
Upon applying linker 4 instead of linker 2 or 3, no agglutina-
tion was observed for either ssDNA or dsDNA. The OD600 of the
ternary mixtures remained as low as 0.01 (Figure 2). The parti-
cle size of 100 nm, measured by DLS, corresponds to the free
CDV (Figure S7 in the Supporting Information). These results il-
lustrate the effect of the protein-binding moiety. In a ternary
mixture comprising CDV and 2’, both DNA samples form supra-
molecular complexes leading to an increase in OD600. Com-
pared to 2’, linker 4 comprises a guanidine unit instead of
a spermine. Under the tested conditions (pH 7.2), the guani-
dine is mostly unprotonated and thus not charged. This was
confirmed by the negative surface potential of À14.8 mV re-
corded for CDV after addition of 4 (Figure S8 in the Supporting
Information). As a result, this linker should bind to the phos-
phate backbone of DNA by hydrogen bonding with guanidine
rather than by electrostatic interaction. Interestingly, no aggre-
gation was observed, indicating that the DNA cannot bend
into the confined geometry that would be needed for multiva-
lent hydrogen bonding. These results indicate that the lack of
directionality in the electrostatic interaction between DNA and
linker molecules 2’ and 3 is crucial for the formation of a supra-
molecular complex.
Figure 2. Time-dependent measurement of OD600 showing the formation of
ternary complexes of DNA and CDV with linker molecules 2’–4; Concentra-
tions: [1]=30 mm; [2’]={40, 60} mm; [3]=[4]=60 mm; [50-mer DNA]={1.6,
3.2} mm, [dsDNA]=40 nm.
In the second part of this study, we expand the scope of
target biomolecules from DNA to proteins. As positively
charged linker molecule we selected 2, which gave the best re-
sults in the experiments with DNA. Linker 5 was synthesized
for these studies with the same azobenzene and spacer unit,
bearing a dipeptide of glutamic acid instead of the spermine.
In a first set of experiments, we tested bovine serum albumin
(BSA, pI=4.7), bearing a net negative charge of À15 in combi-
nation with cationic linker 2 and polycationic protamine sulfate
(PS, pI=13.3) with a total charge of +20 in combination with
anionic linker 5 (Figure 3).
An increase in OD600 from 0.05 to about 1.0 was detected
for ternary mixtures when protein and linker were oppositely
charged, whereas other combinations resulted in no aggrega-
tion. The rate and extent of aggregation are controlled by the
concentration of the linker: instant and complete agglutination
occurred at high concentrations, whereas at low concentration,
shows the OD600 of ternary mixtures of CDV and DNA with dif-
ferent linkers (2’–4). In the case of linker molecule 3, only
ssDNA showed aggregation and consequently an increase in
optical density from 0.01 to about 0.91, which is comparable
to the results with linker 2’ discussed above. Under the same
conditions, dsDNA did not display substantial aggregation:
OD600 increased to only 0.07. The aggregation was also fol-
lowed by DLS (Figure S7 in the Supporting Information). The
average particle diameter increased from about 100 nm for the
free vesicles to more than 1000 nm after addition of 50-mer
DNA. In contrast, the particle diameter was unaffected by addi-
tion of dsDNA. Finally, measurements of the zeta potential
were carried out (shown for 50-mer DNA in Figure S8 in the
&
&
Chem. Eur. J. 2015, 21, 1 – 8
4
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!