Self-Assembled Vesicle–Carbon Nanotube Conjugate Formation through a Boronate–Diol Covalent Linkage

Article abstract of DOI:10.1002/chem.201703452

A vesicle–single walled carbon nanotube (CNT) conjugate was developed by a boronic acid–diol covalent linkage between a self-assembled vesicle and dispersed CNT. Trimesic acid based phenylboronic acid appended triple-tailed amphiphiles (T1 and T1S) were synthesized that formed monolayered vesicles through H-aggregation in DMSO–water (2:1 v/v) and pure water, respectively. Aqueous CNT dispersion was prepared with cholesterol-based glucose-functionalized amphiphile (D1). These two supramolecular self-assemblies were covalently linked by using a boronic acid–diol interaction between a phenylboronic acid based T1S vesicle and 1,2-diol moieties of glucose tethered dispersing agent (D1) to develop a vesicle–CNT conjugate. Lewis acid–base chemistry was exploited to form this boronate–diol adduct between two supramolecular self-assemblies. The formation of vesicles, CNT dispersion, and the vesicle–CNT conjugate was characterized by microscopic and spectroscopic techniques. Anticancer drug doxorubicin was encapsulated within this T1S-vesicle–D1-CNT conjugate with a higher loading capacity compared to the individual cargo carrier (vesicle or CNT). This cytocompatible T1S-vesicle–D1-CNT conjugate successfully delivered loaded doxorubicin within a B16F10 melanoma cell and also exhibited better cellular transportation ability compared to the drug-loaded vesicle or CNT. This was further reflected in an enhanced killing efficiency of the cancer cells by the vesicle–CNT conjugate compared to the drug-loaded vesicle or CNT.

Full text of DOI:10.1002/chem.201703452

A Journal of
Accepted Article
Title: Self-Assembled Vesicle-Carbon Nanotube Conjugate Formation
through Boronate-Diol Covalent Linkage
Authors: Prasanta Kumar Das, Soumik Dinda, Deep Mandal, and
saheli Sarkar
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201703452
Link to VoR: http://dx.doi.org/10.1002/chem.201703452
Supported by

10.1002/chem.201703452
Chemistry - A European Journal
Self-Assembled Vesicle-Carbon Nanotube Conjugate Formation
through Boronate-Diol Covalent Linkage
Soumik Dinda, Deep Mandal, Saheli Sarkar and Prasanta Kumar Das*[a]
Department of Biological Chemistry, Indian Association for the Cultivation of Science,
Jadavpur, Kolkata – 700 032, India
[a] S. Dinda, D. Mandal, S. Sarkar and Prof. P. K. Das
Department of Biological Chemistry
Indian Association for the Cultivation of Science
Jadavpur, Kolkata – 700032, India
Fax:
+(91)-33-24732805
E-mail: bcpkd@iacs.res.in
*To whom correspondence should be addressed: bcpkd@iacs.res.in
1
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Abstract
A vesicle-single walled carbon nanotube (CNT) conjugate is developed by boronic acid-diol
covalent linkage between self-assembled vesicle and dispersed CNT. Trimesic acid based
phenylboronic acid appended triple tailed amphiphiles (T1 and T1S) were synthesized that
formed monolayered vesicles through H-aggregation in DMSO-water (2:1 v/v) and pure
water, respectively. Aqueous CNT dispersion was prepared with cholesterol based glucose
functionalized amphiphile (D1). These two supramolecular self-assemblies were covalently
linked using boronic acid-diol interaction between phenylboronic acid based T1S vesicle and
1,2-diol moieties of glucose tethered dispersing agent (D1) to develop vesicle-CNT
conjugate. Lewis acid-base chemistry was exploited to form this boronate-diol adduct
between two supramolecular self-assemblies. The formation of vesicles, CNT dispersion and
the vesicle-CNT conjugate was characterized by microscopic and spectroscopic techniques.
Anticancer drug doxorubicin was encapsulated within this T1S-vesicle-D1-CNT conjugate
with a higher loading capacity compared to individual cargo carrier (vesicle or CNT). This
cytocompatible T1S-vesicle-D1-CNT conjugate successfully delivered loaded doxorubicin
within B16F10 melanoma cell and also exhibited better cellular transporation ability
compared to drug loaded vesicle or CNT. This was further reflected in enhanced killing
efficiency of the cancer cells by vesicle-CNT conjugate compared to drug loaded vesicle or
CNT.
2
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Introduction
Over the past few decades, lipid vesicles gained enormous importance because of its ability
to mimic the properties of biological membranes and cellular organelles.[1,2] Vesicle, a unique
class of supramolecular self-assembly, is an well established drug delivery vehicle with
considerable cargo loading ability, excellent cytocompatibility and biodegradability.[3-5]
Concurrently, in recent years, single walled carbon nanotube (CNT) has also emerged as
succesful cellular transporter due to its needle like hollow shape and superior ability to
translocate through plasmatic membrane.[6] CNT, the pseudo one dimensional allotrope of
carbon, has been widely exploited in numerous scientific fields due to its intrinsic
mechanical, thermal and optical properties.[7,8] Both vesicles and CNTs have their individual
prominence in supramolecular self-assembly and biomedicine. Despite having such
extraordinary credentials, the major limitations of using CNTs as drug delivery vehicle are
poor cargo loading ability and inherent cytotoxicity to mammalian cells.[9] Moreover, the
hydrophobic surface of CNTs non-specifically binds with the nutrients (essential constituents
for cell growth) leading to indirect toxicity towards living cells.[9c] On the contrary, the major
drawback of vesicles is its fast elimination from blood circulation due to efficient uptake by
reticulo-endothelial system.[10] Some pegylated lipids (stealth liposomes) had overcome this
problem where polyethylene glycol (PEG) moiety acts as the steric stabilizer.[10] However,
this sterically stabilized vesicle often prevented while entering into the targeted cells.[3a]
To overcome these hurdles, necessity of developing a new soft-nanocomposite has
emerged with low cytotoxicity and improved drug loading ability. To this end, amalgamation
of self-assembled vesicles with dispersed CNT may lead to the development of vesicle-CNT
conjugate with desired features. Till date, very few attempts were made to develop CNT-
liposome conjugates. Regev and co-workers reported the formation of CNT-liposome
conjugate through covalent amide linkage between carboxylated multi-walled CNT and the
3
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
primary amine containing liposome.[11] CNT-liposome conjugate was also utilized as
molecular-transporter by utilizing the ligand-receptor interaction between liposome and
CNT.[7b] Electrostatic interaction had also been used to form single walled carbon nanotube-
liposome complex.[12] Also there are reports on the formation of CNT mediated fusion of lipid
vesicles, vesicle encapsulated multi-walled CNT.[13] However, the befitting role of the
vesicle-CNT conjugate over the individual cargo carrier (vesicle or CNT) is yet to be
emphasized.
Herein, the vesicle-CNT conjugate was constructed using covalent interaction
between the CNT dispersing agent and vesicular building blocks. Phenylboronic acid
appended trimesic acid based monolayered vesicles (T1S) were linked with the cholesterol
based glucose functionalized CNT dispersing agent (D1) using complementary boronic acid-
diol interaction chemistry (Scheme 1, Figure 1a). The lewis acid-base interaction was
exploited to make the covalent linkage between the building block of monlayered vesicles
and CNT dispersing agent to develop the vesicle-CNT conjugate. To the best of our
knowledge, this is the first ever approach where dispersed CNT surface has been decorated
with monolayered vesicles. Microscopic and spectroscopic techniques were utilized to
characterize this newly developed vesicle-CNT conjugate. The vesicle-CNT conjugate was
found to be stable and cytocompatible in biological milieu. Anticancer drug doxorubicin was
encapsulated within this vesicle-CNT conjugate with improved loading capacity compared to
its individual constituents. Importantly, vesicle CNT conjugate delivered the drug with
greater efficiency inside the B16F10 melanoma cells compared to drug loaded individual
vesicle or dispersed CNT resulting in higher killing of cancer cells by the conjugate.
Results and Discussion
A trimesic acid based triskelion amphiphile (T1) was synthesized having three phenylboronic
acid (PBA) moieties at the terminals (Figure 1a). C3-symmetrical trimesic acid was chosen as
4
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
a primary scaffold because of its tri-functionalized structure with aromatic moiety at its core
that could instigate ordered -stacking during self-assembly. The PBA moieties were
integrated through oxyethylene spacer at the terminals of T1 to impart hydrophilicity in the
amphiphile.[14,15] With the objective to develop vesicle-CNT conjugate, a cholesterol based
glucose functionalized CNT dispersing agent (D1, Figure 1a) was synthesized.[16] The 1,2-
diol moieties of glucose unit of D1 would be appropriate for linking with boronic acid of
vesicular building block via boronate-diol covalent interaction.[17] Compound-T1 was found
to be insoluble in water while it formed a translucent solution in DMSO-water (2:1 v/v)
binary solvent mixture with an indication of vesicle formation. The corresponding
transmission electron microscopic (TEM) image showed the formation of spherical vesicular
self-assemblies having average diameter of 150-200 nm (Figure 1b). Since T1 formed vesicle
in DMSO-water binary solvent mixture, we synthesized the sodium salt of T1 (T1S, Figure
1a) with the expectation that it will form vesicle in pure aqueous medium. Indeed, T1S
formed translucent solution in water with a preliminary indication of vesicle formation. The
respective TEM image also confirmed the formation of vesicular aggregates in water with
comparatively smaller diameter (50-100 nm, Figure 1c). Both vesicular self-assemblies
(originating from T1 and T1S) bear common morphological characteristics of thin wall with
a hollow core.
Critical Aggregation Concentration (CAC).
The critical aggregation concentration (CAC) of T1 and T1S was measured for both
amphiphiles by surface tension method in the respective solvent systems. The gradual
decrease of surface tension upon increasing the concentrations of T1 and T1S continued until
reaching the CAC values. After attainment of the CAC value, any further addition of
amphiphile molecules will just increase the number of aggregation without any substantial
5
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
change in surface tension values. The CAC values were found to be ~250 and ~300 µg/mL
for T1 and T1S, respectively (Figure S1 in the Supporting Information).
Microscopic Studies.
Field-emission scanning electron microscopic (FESEM) images showed the formation of
vesicular self-assemblies by T1 and T1S in respective solvents having average diameter of
150-200 nm and 50-100 nm, respectively (Figure 2a,b). In concurrence with the FESEM
images, atomic force microscopy (AFM) further ensured the perfectly spherical morphology
formed by the self aggregation of T1 and T1S in respective solvent systems (Figure 2c,d).
AFM images showed the average diameter of T1 vesicles were in the range of ~200 nm while
that was ~90 nm for T1S vesicles. Moreover, the cross-section analysis of the AFM images
(Figure 2c (p-q) and 2d (r-s)) provide an idea about the height of the vesicles which was ~30
nm and ~15 nm for T1 and T1S vesicles, respectively (Figure 2e and 2f) indicating that both
the vesicles were hollow and flattened in shape.[18]
Spectroscopic Studies.
Dynamic Light Scattering (DLS) and Zeta (ζ) Potential Measurements.
The dynamic light scattering (DLS study) revealed the mean hydrodynamic diameter of both
vesicles. In Figure 2g, size distribution profile for T1 delineates the average hydrodynamic
diameter for T1 vesicles is ~150 nm while it is quite smaller for T1S vesicles (~80 nm).
These spectroscopic data regarding the size of both the vesicular aggregates (T1 and T1S) is
in well agreement with the microscopic evidences.
The colloidal stability of T1 and T1S was estimated by zeta (ζ) potential
measurement. Very high ζ potential value in case of T1S (-22.91 mV) compared to T1 (-3.7
mV) is probably due to negatively charged boronate ion in the side chains of T1S. These
values indicate substantial stability of the corresponding vesicular aggregates.[19]
6
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
1
Solvent Dependent H NMR Study. To know the various interacting forces responsible for
1
the formation of supramolecular self-assemblies, the solvent dependent H-NMR experiment
was performed for both amphiphiles (T1 and T1S). The NMR signals of aromatic protons
gradually upfield shifted with reduced peak intensity upon transformation from non-self-
assembled state (T1 and T1S taken in DMSO-d6) to self-assembled state (T1 taken in
DMSO-d6-D2O (2:1 v/v) and T1S taken in D2O) of the amphiphiles (Figure 3a,b). In DMSO-
d6, the aromatic proton of central trimesic acid of T1 showed a sharp peak at  = 8.42 ppm
while the aromatic protons of PBA moieties showed distinguished peaks at the region  =
7.89 ppm and  =7.92-8.13 ppm and 6.88-6.95 ppm, respectively. In DMSO-d6, T1 was
present in molecularly dissolved state without any self-association. Upon increasing the D2O
content, the peaks stated above got shifted in the upfield region i.e. at  = 7.89 ppm and 
=7.39-7.62 ppm and 6.51-6.59 ppm, respectively accompaning with reduced peak intensities.
In a similar fashion, the NMR signals of aromatic protons of trimesic acid of T1S got upfield
shifted form  = 8.80 ppm to  = 8.28 ppm on moving form non-self-assembled state (in
DMSO-d6) to self-assembled state (in D2O). Notably, the aromatic protons of PBA moieties
exhibited upfield shifting of the NMR signals from  =7.81-8.07 ppm and 7.81-7.88 ppm to 
=7.41-7.62 ppm and 6.41-6.48 ppm, respectively. Thus, self-aggregation gets started and
facilitated during the increment of D2O content. This result delineates the active participation
of hydrophobic interaction and -stacking during vesicle formation during self-assembly of
T1 and T1S.[15b]
Steady State Fluorescence Anisotropy.
Steady state fluorescence anisotropy (r) measurement was carried out using 1,6-diphenyl-
1,3,5-hexatriene (DPH) as probe to find out the nature of microenvironment in self-
aggregated structure. The r-value increased from 0.08 to 0.19 and 0.07 to 0.15 (Table 1) for
DPH included T1 and T1S self-assemblies over a concentration range from 2 to 20 mg/mL,
7
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
respectively. The significantly higher r values (greater restriction in the movement of DPH in
the hydrophobic region) of DPH (0.19 for T1 and 0.15 for T1S) compared to that of micellar
aggregates of sodium dodecyl sulfate (r = 0.054) further confirm the formation of vesicular
aggregates by T1 and T1S.[20]
UV-vis Study.
The supramolecular aggregation pattern of T1 and T1S was investigated from UV-vis
spectroscopy in the presence of well known chromophore, 8-anilino-1-naphthalenesulfonic
acid (ANS) (1×10-5 M). ANS is prone to confine itself at the hydrophobic region of the
aggregated structure. It showed absorption maxima at 380 and 378 nm in non-self-assembled
state of T1 and T1S (both were taken in DMSO), respectively. With increasing the water
content the absorption maxima of ANS gradually blue shifted for both amphiphiles indicating
the initiation of self-aggregation of monomeric units. In self-aggregated state of T1 (DMSO-
H2O (2:1 v/v)) and T1S (water), the absorption maxima of ANS was shifted to 367 and 345,
respectively (Figure 3c,d). This blue shift in absorption maxima of ANS during self-
aggregation designates parallel plane-to-plane stacking of the amphiphiles forming a
sandwich type arrangement (H-type aggregation).[14b,21] However, no such change in
absorption and aggregation pattern is known to be reported for bilayered phosphocholine
(PC) liposomes.[14b] Thus, a monolayered ordered arrangement of T1 and T1S possibly took
place due to the hydrophilic−hydrophobic−hydrophilic molecular backbone of the
amphiphiles, which is different in nature from that of bilayered PC liposomes.[14b,22] The X-
ray diffraction (XRD) of T1 and T1S in self-aggregated state showed the broad peak at 2θ
~23° (Figure S2 in the Supporting Information), which corresponds to stacked
organization of amphiphiles.[23]
Development of Vesicle-CNT Conjugate.
8
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
At this point, we intend to amalgamate the newly developed vesicles with dispersed CNTs to
fabricate vesicle-CNT conjugate. Cholesterol based glucose tailored amphiphile (D1, Figure
1a) debundled the amphiphobic CNT with 63% efficiency (Figure S3a in the Supporting
Information) in water (see experimental section). The hydrophobic cholesteryl moiety of D1
possibly got attached non-covalently on the CNT surface and the hydrophilic glucose unit
facilitated the dispersion of hydrophobic CNTs in aqueous medium. This CNT dispersion
was found to be stable more than 3 months without any precipitation. The debundling of CNT
was confirmed from the corresponding TEM image (Figure 1d) where individual nanotubes
having a diameter of 5-7 nm and average length of 400-600 nm were well separated from
each other. Furthermore, Raman spectroscopic investigation was carried out to characterize
the aqueous dispersion of CNT by D1. The sharp G-band at 1590 cm-1 was observed in the
Raman spectrum due to the tangential C-C stretching transitions in CNT (Figure S4a in the
Supporting Information).[24] Moreover, other two peak at 1350 cm-1 and 2690 cm-1 were
found in the Raman spectrum corresponding to the D-band and its second-order G'-band
respectively (Figure S4a in the Supporting Information).[24] All the above bands corroborate
with the Raman spectrum of solid pristine CNT (Figure S4b in the Supporting Information).
The aqueous dispersion of CNT (D1) was mixed with the preformed vesicles of T1S to
fabricate vesicle-CNT conjugate (see experimental section). As to our expectation, the TEM
image of the mixed solution (T1S-vesicle + D1-CNT dispersion) confirmed the successful
formation of T1S-vesicle-D1-CNT conjugate (Figure 4a and Figure S5 in the Supporting
Information). The vesicles were found to be aligned over the CNT surface and the shape and
the size of the vesicles were found to be unchanged after conjugation with CNT. The vesicles
get attached on the CNT surface through lewis acid-base chemistry where boronate unit of
the vesicular building block acts as lewis acid and 1,2-diol of the glucose unit on the CNT
surface acts as lewis base.[17] The formation of cyclic boronate-diol adduct through covalent
9
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
linkage between building block of vesicles (T1S) and dispersing agent of CNT (D1) was
1
characterized by H-NMR and mass spectroscopy (Figure S6 and S7 in the Supporting
Information). This covalent linkage tied up both vesicle and CNT leading to formation of
T1S-vesicle-D1-CNT conjugate. Corresponding AFM images (Figure 4b) also supported the
attachment of vesicles on CNT surface confirming the successful conjugation of vesicle with
CNTs. Importantly, vesicle-CNT conjugate suspension was found to be stable in water
(Figure S3b in the Supporting Information). Thus, the conjugation between self-assembled
vesicle and dispersed CNT did not affect the structural integrity of the vesicle as well as the
non-covalent attachment of dispersing agent on CNT surface, otherwise CNT would have
been precipitated. The vesicle-CNT conjugate formation was further characterized by
thermogravimetric analysis (TGA). To this end, TGA plot of D1-CNT showed a sharp
decrease in weight between 200 and 300 °C corresponding to the decomposition of D1
(Figure 4c). Similarly, TGA plot of T1S-vesicle-D1-CNT conjugate exhibited sharp decrease
in weight between 100 and 200 °C corresponding to the decomposition of newly formed
boronate-diol cyclic conjugate (Figure 4d). Almost 35% and 38% residue remains after the
decomposition which is corresponding to the amount of CNT present in the sample in case of
D1-CNT and T1S-vesicle-D1-CNT conjugate, respectively (Figure 4c,d). Hence, ~65% of
the D1 dispersing agent gets attached to the walls of the CNT while ~62% of the cyclic
adduct (D1-T1S) was involved in the formation of vesicle-CNT conjugate.
To ensure the formation of vesicle-CNT conjugate between T1S-vesicle and D1-
CNT, control experiment was carried out where CNT was debundled with a well-known
dispersing agent SDBS (sodium dodecyl benzenesulfonate) and mixed with preformed
vesicular solution of T1S (Figure S8a in the Supporting Information). In another experiment,
monolayered vesicle of a triskelion amphiphile devoid of PBA moiety was mixed with
aqueous dispersion of D1-CNT by (Figure S8b in the Supporting Information).[14b] In both
10
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
cases, corresponding TEM images (Figure S8c, S8e and S8d, S8f in the Supporting
Information) did not provide any indication on the formation of vesicle-CNT conjugate.
Similarly, the AFM images did not exhibit any attachment of vesicles to the CNT surfaces
(Figure S8g and S8h in the Supporting Information). These control experiments further
confirmed that covalent bond formation between complementary boronate-diol moieties of
vesicle and CNT dispersing agent led to the formation of new vesicle-CNT conjugate.
The conjugation of vesicles on dispersed CNT surfaces was also investigated by
photoluminescence spectroscopy. We have studied the emission property of fluorescein
isothiocyanate (FITC) in T1S vesicle, T1S-vesicle-D1-CNT conjugate and water. FITC is
known to get localized at the hydrophobic core of the aggregated structure.[25] Concurrently,
CNT is also well known for quenching the emission intensity of the fluorescence probe
(FITC).[26] FITC exhibited emission maxima at 520 nm upon excitation at 495 nm (Figure
5a). In T1S vesicle, a significant increase in the FITC fluorescence intensity was noted
obviously due to its localization at the hydrophobic wall of the monolayered vesicles.
Notably, the fluorescent intensity of FITC gets quenched drastically in the T1S-vesicle-D1-
CNT conjugate (Figure 5a), which confirms the presence of CNT in close proximity of the
probe, located at the hydrophobic vesicular wall. Hence it further validates the formation of
T1S-vesicle-D1-CNT conjugate by the attachment T1S vesicle on the surface of dispersed
CNT, which is in concurrence with microscopic evidences.
Moreover, we measured the decay time of coumarin 153 (C153, excitation 405 nm)
in T1S vesicle, T1S-vesicle-D1-CNT conjugate as well as in water in the time-correlated
single-photon counting (TCSPC) experiment monitored at 550 nm (Figure 5b). The decay
time of free C153 in water was found to be 1.63 ns while it increased to 1.71 ns when the dye
was included within T1S vesicle. However, the decay time observed in T1S vesicle is
distinctly lower than that was reported for bilayered PC liposomes (4.65 ns), which further
11
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
indicates the formation of monolayered vesicles by T1S.[14b] This increment in decay time of
C153 within T1S vesicle was possibly due to the encapsulation of the dye in more
hydrophobic environment. Importantly, the decay time of C153 further increased to 1.93 ns
within T1S-vesicle-D1-CNT conjugate. Thus, T1S-vesicle-D1-CNT conjugate possessed
enhanced hydrophobic microenvironment compared to T1S vesicle presumably due to the
attachment of monolayered vesicles on the surface of hydrophobic CNT. All spectroscopic
and microscopic evidences ensured that the boronate-diol covalent interaction between the
vesicular amphiphiles and CNT dispersing agent led to the development of T1S-vesicle-D1-
CNT conjugate.
Doxorubicin Loading and Release.
After ensuring the successful formation of T1S-vesicle-D1-CNT conjugate, we aim to
investigate the cargo loading ability and cellular transportation efficacy of this conjugate. To
this end, anticancer drug doxorubicin was separately loaded in the T1S vesicle, aqueous CNT
dispersion (D1) as well as T1S-vesicle-D1-CNT conjugate (see experimental section). Free
doxorubicin was discarded from the loaded doxorubicin in vesicle and T1S-vesicle-D1-CNT
conjugate by size exclusion column chromatography using sephadex G-50. In case of CNT,
free doxorubicin was separated after precipitating the drug loaded CNT through
centrifugation (see experimental section). The eluted solution from the sephadex column
showed characteristic UV-vis absorbance peaks at λmax = 490 nm indicating the presence of
the doxorubicin within the drug carriers (Figure 6a). Drug encapsulated vesicle and T1S-
vesicle-D1-CNT solutions were further observed under fluorescence microscope, which
showed bright red emitting spherical aggregates (Figure 6b,c). This observation subsequently
confirmed the loading of doxorubicin within vesicle and T1S-vesicle-D1-CNT conjugate.
Encapsulation of the drug was further ensured by comparing the fluorescence intensity of free
doxorubicin and loaded doxorubicin in different drug carriers (vesicle, CNT and T1S-vesicle-
12
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
D1-CNT conjugates) at 590 nm (λex= 490 nm) (Figure S9 in the Supporting Information). The
emission intensity of loaded doxorubicin in each carrier (T1S-vesicle, D1-CNT and T1S-
vesicle-D1-CNT) was found to be very low compared to that of free doxorubicin at same
concentration (Figure S9 in the Supporting Information).[14b] Notably, encapsulated
doxorubicin regained its fluorescence intensity and became comparable to free doxorubicin
upon treating the drug loaded, T1S-vesicle with Triton X-100, D1-CNT with 1(N) HCl and
T1S-vesicle-D1-CNT conjugate with both Triton X-100 and 1(N) HCl (Figure S9 in the
Supporting Information). Triton X-100 assisted the rupturing of vesicular wall resulting in
release of the drug.[27] Protonation of the –NH2 group of doxorubicin in presence of HCl
resulted in the release of hydrophilic drug from CNT surface due to the lower affinity of the
drug towards the hydrophobic CNT surface.[28] Triton X-100 and 1(N) HCl both have
cumulative effect on doxorubicin loaded T1S-vesicle-D1-CNT conjugate for releasing the
drug. Photoluminiscence spectroscopy further confirmed the drug loading ability of different
cargo transporters.
Doxorubicin loading efficiency of the T1S-vesicle-D1-CNT conjugate was found to
be 67% whereas in case of individual T1S-vesicle and D1-CNT, it was 57% and 23%,
respectively (measured after treating the doxorubicin entrapped T1S-vesicle with Triton X-
100, D1-CNT with 1(N) HCl and T1S-vesicle-D1-CNT conjugate with both Triton X-100 and
1(N) HCl, followed by comparing the absorbance of released doxorubicin with standard
calibration curve, see experimental section, Figure S10 in the Supporting Information).[29]
T1S-vesicle is capable to accommodate considerable amount of doxorubicin in its inner
aqueous core, while CNT can upload doxorubicin to its side wall through hydrophobic
interaction. In T1S-vesicle-D1-CNT conjugate, combined presence of the aqueous core of
vesicle and large CNT surface resulted a synergistic contribution towards the notable
improvement in the drug loading efficiency compared to the individual constituents.
13
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Media Stability.
Prior to utilize this newly developed T1S-vesicle-D1-CNT conjugate as a delivery vehicle,
we tested the stability of this conjugate in biological milieu as well its cytocompatibility
against the mammalian cells. According to the visual appearance and suspension stability
index, vesicle-CNT conjugate ([CNT] = 25 g/mL) in Dulbecco’s Modified Eagles’ Medium
(DMEM) with varying fetal bovine serum (FBS) concentration was found to be more than
90% stable (Figure S11a in the Supporting Information). Moreover, the conjugate also
exhibited significant stability (>90%) up to 10 days in both DMEM media supplemented with
10% FBS as well as in physiological pH i.e. pH 7.4 PBS buffer (Figure S11b,c in the
Supporting Information). Ensuring high media stability of T1S-vesicle-D1-CNT, the
cytocompatibility of T1S vesicle, dispersed (D1) CNT and T1S-vesicle-D1-CNT conjugate
was investigated against the B16F10 mammalian cell line by MTT assay (Figure S12 in the
Supporting Information).[30] The T1S vesicle (100-500 µg/mL) showed a superior
cytocompatibility having 90-95% viable cells after 12h of incubation. In case of dispersed
(D1) CNT (5-25 µg/mL), moderate cell viability (~40-60%) was noted presumably due to the
inherent toxicity and adverse effect on the necessary cellular nutrients by CNT (Figure S12 in
the Supporting Information).[9] Interestingly T1S-vesicle-D1-CNT conjugates with similar
CNT concentration range (5-25 µg/mL), exhibited substantial cytocompatibility where ~70-
80% cells were found to be alive (Figure S12 in the Supporting Information). The inclusion
of highly cytocompatible vesicular aggregates on the surface of CNT significantly reduced
the toxic effect of CNT in the T1S-vesicle-D1-CNT conjugate and satisfied its suitability as
cellular transporter.
Internalization of Doxorubicin within Cancer Cells.
Herein, drug transportation ability by individual delivery vehicle was investigated upon
incubation of doxorubicin (5 g/mL) loaded T1S-vesicle, D1-CNT and T1S-vesicle-D1-CNT
14
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
conjugate, separately with B16F10 melanoma cells for 6 h. In each instance, bright red
fluorescence image of cells under the fluorescence microscope confirmed the successful
internalization of doxorubicin inside B16F10 melanoma cells (Figure 7a,b,d,e,g,h). The
acidic microenvironment of B16F10 cancer cells might have facilitated the protonation of the
–NH2 group of doxorubicin and successecsive release of drug from hydrophobic CNT
surface. The extent of doxorubicin internalization within B16F10 cells by different delivery
vehicle was assessed by flow cytometric analysis. The respective flow cytometric plot
exhibited fluorescence intensity in the order of 104–105 having mean fluorescence value of
29295, 39333 and 82860 for doxorubicin loaded T1S-vesicle, D1-CNT and T1S-vesicle-D1-
CNT conjugate, respectively (Figure 7c,f,i). It is evident that dispersed CNT has better cargo
transportation ability than the vesicle. This observation is in concurrence with previous report
that despite having higher loading capacity, vesicle has the poor internalization efficacy in
contrast to the CNT.[31,32] Importantly, with the objective of the present study, most efficient
drug delivery was exhibited by T1S-vesicle-D1-CNT conjugate having highest mean
fluorescence value of 82860 (Figure 7i). Thus, in comparison to the individual vesicle and
CNT, the newly developed T1S-vesicle-D1-CNT conjugate is more proficient towards
transporting the anti-cancer drug doxorubicin inside the B16F10 cancer cells primarily due to
the amalgamation of vesicles and CNT having high drug loading capacity and efficient cell
penetration ability, respectively.
Cancer Cell Killing.
The superior cellular transportation ability of T1S-vesicle-D1-CNT conjugate prompted us to
investigate the efficacy of the delivered doxorubicin towards killing the cancer cells. Herein,
we studied the % killing of B16F10 cells for doxorubicin loaded delivery vehicles by MTT
assay. B16F10 cells were treated for 12 h with doxorubicin loaded (5-25 g/mL) T1S-
vesicle, D1-CNT and T1S-vesicle-D1-CNT conjugate, separately (Figure 7j). In each
15
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
instance, the killing efficacy almost steadily increased with increase in concentration of
doxorubicin from 5 to 25 g/mL. Also, at a given concentration of doxorubicin, the % killing
of cancer cells distinctly improved from vesicle to CNT and being most efficient for T1S-
vesicle-D1-CNT conjugate. Encouragingly, drug loaded T1S-vesicle-D1-CNT conjugate
showed ~71% killing of B16F10 cells at [doxorubicin] = 25 g/mL, whereas drug loaded
native vesicle and dispersed CNT exhibited 44% and 49% killing of cancer cells, respectively
(Figure 7j). In case of free doxorubicin, only ∼21% killing of B16F10 cells was observed at
25 g/mL of the drug upon 12 h of incubation (Figure S13 in the Supporting Information).
This indicates the inefficient killing ability of free doxorubicin without any cargo transporter.
Cancer cell killing efficiency by doxorubicin loaded different delivery vehicles follows the
same trend that was observed in the flow cytometry (T1S-vesicle-D1-CNT conjugate is being
the most efficient). Hence, the amalgamation of vesicle and CNT through boronate-diol
covalent interaction leads to the development of a highly efficient delivery vehicle (T1S-
vesicle-D1-CNT conjugate) which has improved drug loading ability and notable proficiency
in cellular transportation.
Conclusions
In summary, the vesicle-CNT conjugate was developed with facile combination of
supramolecular self-assemblies of phenylboronic acid appended trimesic acid based
monolayered vesicles (T1S) and cholesterol based glucose functionalized CNT dispersion
(D1) using boronate-diol covalent linkage. Formation of this soft-nanocomposites was
thoroughly investigated using different microscopic as well as spectroscopic techniques.
Anticancer drug doxorubicin was encapsulated within this vesicle-CNT conjugate with
improved loading capacity compared to its individual constituents. This newly developed
conjugate efficiently delivered the drug inside the B16F10 melanoma cells in comparison to
only vesicle/CNT resulting in higher killing of cancer cells. The present study may instigate a
16
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
novel strategy towards developing smart delivery vehicle with regulated dose of loaded drug
and improved proficiency in cellular transportation.
Acknowledgements:
P.K.D. is thankful to CSIR, India (ADD, CSC0302) for financial assistance. S.D., D.M.
acknowledge CSIR, India and S.S thanks DST inspire for Research Fellowships. We
acknowledge Dr. Moumita Ghosh for helpful discussion.
1
Supporting Information Available: Experimental procedures, H-NMR and mass spectral
analysis of the amphiphiles, synthetic scheme, critical aggregation concentration (CAC) plot,
XRD plot, Raman spectra, 1H-NMR and mass spectra of boronate-diol adduct, visual
observations of D1-CNT and T1S-vesicle-D1-CNT conjugate in water, control experiment by
mixing of T1S-vesicle and CNT dispersed by SDBS and another triskelion vesicle and D1-
CNT and the corresponding TEM and AFM images, photoluminescence spectra of
doxorubicin loaded D1-CNT, T1S-vesicle, T1S-vesicle-D1-CNT conjugate and the release of
doxorubicin, UV-Vis spectra of released doxorubicin from T1S-vesicle-D1-CNT conjugate,
T1S-vesicle and D1-CNT, suspension stability index of T1S-vesicle-D1-CNT conjugate, cell
viability data, killing of B16F10 cells by free doxorubicin. Supporting information for this
article is available on the website under http://www.chemeurj.org/ or from the author.
References:
[1] a) D. A. Wilson, R. J. M. Nolte, J. C. M. van Hest, J. Am. Chem. Soc. 2012, 134,
9894−9897; b) P. Walde, K. Cosentino, H. Engel, ChemBioChem 2010, 11, 848-865;
[2] a) H. T. Jung, B. Coldren, J. A. Zasadzinski, D. J. Iampietro, E. W. Kaler, Proc. Natl.
Acad. Sci. 2001, 98, 1353-1357; b) Y. X Xu, G. T. Wang, X. Zhao, X. K. Jiang, Z. T. Li,
Langmuir 2009, 25, 2685-2688.
[3] a) A. A. Gabizon, Adv. Drug Deliv. Rev. 1995, 16, 285-294; b) P. Majumder, S. Bhunia,
J. Bhattacharyya, A. Chaudhuri, J. Control. Release 2014, 180, 100-108.
17
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
[4] a) D. Zucker, D. Marcus, Y. Barenholz, A. Goldblum, J. Control. Release 2009, 139, 73-
80; b) D. Pramanik, B. K. Majeti, G. Mondal, P. P. Karmali, R. Sistla, O. G. Ramprasad, G.
Srinivas, G. Pande, A. Chaudhuri, J. Med. Chem. 2008, 51, 7298-7302.
[5] a) K. Kono, Adv. Drug Deliv. Rev. 2001, 53, 307-319; b) M. H. Gaber, K. Hong, S. K.
Huang, D. Papahadjoponlos, Pharm. Res. 1995, 12, 1407-1416.
[6] a) D. Pantarotto, R. Singh, D. M. Carthy, M. Erhardt, J. P. Briand, M. Prato, K.
Kostarelos, A. Bianco, Angew. Chem. Int. Ed. 2004, 43, 5242-5246; Angew. Chem. 2004,
116, 5354-5358; b) D. Tasis, N. Tagmatarchis, A. Bianco, M. Prato, Chem. Rev. 2006, 106,
1105-1136; c) S. Dhar, Z. Liu, J. Thomale, H. J. Dai, S. J. Lippard, J. Am. Chem. Soc. 2008,
130, 11467-11476.
[7] a) S. Iijima, Nature 1991, 354, 56-58; b) E. Miyako, K. Kono, E. Yuba, C. Hosokawa, H.
Nagai, Y. Hagihara, Nat. Commun. 2012, 3, 1-9; c) S. Brahmachari, D. Das, A. Shome, P. K.
Das, Angew. Chem. Int. Ed. 2011, 50, 11243-11247; Angew. Chem. 2011, 123, 11439-11443;
[8] a) B. G. Cousins, A. K. Das, R. Sharma, Y. Li, J. P. McNamara, I. H. Hillier, I. A.
Kinloch, R. V. Ulijn, Small 2009, 5, 587-590; b) E. Miyako, T. Itho, Y. Nara, T. Hirotsu,
Chem. Eur. J. 2009, 15, 7520-7525.
[9] a) R. Wang, C. Mikoryak, S. Li, D. Bushdiecker, I. H. Musselman, P. Pantano, R. K.
Draper, Mol. Pharmaceutics 2011, 8, 135-1361; b) K. Kostarelos, Nat. Biotechnol. 2008, 26,
774-776; c) A. Casey, E. Herzog, F. M. Lyng, H. J. Byrne, G. Chambers, M. Davoren,
Toxicol. Lett. 2008, 179, 78-84.
[10] a) A. Gabizon, H. Shmeeda, Y. Barenholz, Clin. Pharmacokinet. 2003, 42, 419-436; b)
A. L. Klibanov, K. Maruyama, A. M. Beckerleg, V. P. Torchilin, L. Huang, Biochim.
Biophys. Acta, Biomembr. 1991, 2, 142-148.
[11] F. Karchemski, D. Zucker, Y. Barenholz, O. Regev, J. Control. Release 2012, 160, 339-
345.
18
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
[12] T. Kim, C. Do, S. Kang, M. Lee, S. Lim, S. Choi, Soft Matter 2012, 8, 9073-9078.
[13] a) R. M. Bhaskara, S. M. Linker, M. Vögele, J. Köfinger, G. Hummer, ACS Nano
2017, 11, 1273-1280; b) P. Zou, G. Shi, C. Y. Pan, J. Polym. Sci. A Polym. Chem. 2009, 47,
3669–3679.
[14] a) C. M. A. Leenders, M. B. Baker, I. A. B. Pijpers, R. P. M. Lafleur, L. Albertazzi, A.
R. A. Palmans, E. W. Meijer, Soft Matter 2016, 12, 2887-2893; b) S. Dinda, M. Ghosh, P. K.
Das, Langmuir 2016, 32, 6701-6712; c) A. Heeres, C. van der Pol, M. Stuart, A. Friggeri, B.
L. Feringa, J. van Esch, J. Am. Chem. Soc. 2003, 125, 14252-14253.
[15] a) C. M. A. Leenders, G. Jansen, M. M. M. Frissen, R. P. M. Lafleur, I. K. Voets, A. R.
A. Palmans, E. W. Meijer, Chem. Eur. J. 2016, 22, 4608–4615; b) P. Besenius, K. P. van den
Hout, H. M. H. G. Albers, T. F. A. de Greef, L. L. C.Olijve, T. M. Hermans, B. F. M. de
Waal, P. H. H. Bomans, N. A. J. M. Sommerdijk, G. Portale, A. R. A. Palmans, M. H. P van
Genderen,. J. A. J. M. Vekemans, E. W. Meijer, Chem. Eur. J. 2011, 17, 5193-5203.
[16] D. Mandal, S. Dinda, P. Choudhury, P. K. Das, Langmuir 2016, 32, 9780-9789.
[17] a) W. L. A. Brooks, B. S. Sumerlin, Chem. Rev. 2016, 116, 1375-1397; b) Y. Kubo, R.
Nishiyabu, T. D. James, Chem. Commun. 2015, 51, 2005-2020.
[18] I. Reviakine, A. Brisson, Langmuir 2000, 16, 1806-1815.
[19] E. Chibowski1, A. Szczes, Adsorption 2016, 22,755-765.
[20] D. Khatua, J. Dey, J. Phys. Chem. B 2007, 111, 124-130.
[21] Y. Dong, B. Xu, J. Zhang, X.Tan, L. Wang, J. Chen, H. Lv, S. Wen, B. Li, L. Ye, B.
Zou, W. Tian, Angew. Chem. Int. Ed. 2012, 51, 10782-10785; Angew. Chem. 2012, 124,
10940-10943.
[22] Y. Zhou, D. Yan, Chem. Commun. 2009, 1172-1188.
[23] W. Lyu, J. Feng, W. Yan, C. F. J. Faul, J. Mater. Chem. C 2015, 3, 11945-11952.
19
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
[24] M. S. Dresselhausa, G. Dresselhausc, A. Joriob, A.G. Souza Filhob, R. Saito, Carbon
2002, 40, 2043-2061.
[25] H. Zhu, M. J. McShane, J. Am. Chem.Soc. 2005, 127, 13448-13449.
[26] a) D. Mandal, M. Ghosh, S. Maiti, K. Das, P. K. Das, Colloids Surf., B. 2014, 113, 442
449; b) N. Nakayama-Ratchford, S. Bangsaruntip, X. Sun, K. Welsher, H. Dai, J. Am.
Chem.Soc. 2007, 129, 2448-2449.
[27] Y. Shai, D. Bach, A. Yanovsky, J. Biol. Chem. 1990, 265, 20202-20209.
[28] a) M. Ghosh, S. Brahmachari, P. K. Das, Macromol. Biosci. 2014, 14, 1795-1806; b) Z.
Chen, Z. Li, Y. Lin, M. Yin, J. Ren, X. Qu, Biomaterials 2013, 34, 1364-1371.
[29] S. Brahmachari, M. Ghosh, S. Dutta, P. K. Das, J. Mater. Chem. B 2014, 2, 1160-1173.
[30] M. B. Hansen, S. E. Nielsen, K. Berg, J. Immunol. Methods 1989,119, 203-210.
[31] F. Wang, J. Liu, Nanoscale 2013, 5, 12375-12382;
[32] O. Regev, Y. Barenholz, S. Peretz, D. Zucker, Y. B. Felsen, Future Med. Chem. 2013, 5,
503-505.
20
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Table 1. Steady-state fluorescence anisotropy (r) of DPH with varying concentrations of
−water (2:1, v/v) and
T1 in DMSO
T1S in water
Concentration/
r-values
mg/mL
T1
T1S
2
4
8
10
16
20
0.08
0.08
0.14
0.16
0.19
0.19
0.07
0.07
0.13
0.15
0.15
0.15
21
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Scheme 1. Schematic representation of the formation of supramolecular vesicle-CNT
conjugate through boronate-diol covalent linkage.
22
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Figure 1. a) Structure of synthesized triskelion amphiphiles and dispersing agent.
Negatively stained TEM images of b) T1 prepared in DMSO-water (2:1, v/v) c) T1S
prepared in water and d) dispersed CNT in water by D1.
23
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Figure 2. FESEM images of a) T1-vesicles prepared in DMSO-water (2:1, v/v) b) T1S-
vesicles prepared in water, AFM images of c) T1-vesicles prepared in DMSO-water (2:1,
v/v) (inset shows height profile along p-q ) and d) T1S-vesicles prepared in water (inset
shows height profile along r-s). e) Size distribution profile from DLS for T1 and T1S
vesicles.
24
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Figure 3. Solvent dependent 1H-NMR spectra of a) T1 and b) T1S showing the shift of
aromatic protons of trimesic acid (●) and phenylboronic acid (■,▲). UV-Vis spectra of
ANS doped c) T1 and d) T1S in different solvent systems.
25
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Figure 4. a) Negatively stained TEM b) AFM images of T1S-vesicle-D1-CNT
conjugate. Thermogravimetric analysis (TGA) plots of c) D1-CNT and d) T1S-vesicle-
D1-CNT conjugate.
26
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Figure 5. a) Fluorescence spectra of FITC in water, T1S-vesicle and T1S-vesicle-D1-
CNT conjugate in water b) Decay curves of free C153 in water and encapsulated in T1S-
vesicle and T1S-vesicle-D1-CNT conjugate prepared in water (excitation wavelength =
405 nm).
27
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Figure 6. a) UV-Vis spectra of loaded doxorubicin in T1S-vesicle-D1-CNT conjugate,
T1S-vesicle and fluorescence microscopic images of doxorubicin encapsulated b) T1S-
vesicle and c) T1S- vesicle-D1-CNT conjugate.
28
This article is protected by copyright. All rights reserved.

10.1002/chem.201703452
Chemistry - A European Journal
Figure 7. Bright field and fluorescence microscopic images of B16F10 cells after 6 h
incubation with doxorubicin loaded (a,b) T1S-vesicle, (d,e) D1-CNT, (g,h) T1S-vesicle-
D1-CNT conjugate. Corresponding flow cytometric histogram plots for (c) T1S-vesicle,
(f) D1-CNT, (i) T1S-vesicle-D1-CNT conjugate and in all the flow cytometry plots the
x-axis denotes the doxorubicin fluorescence intensity. The mean fluorescence values are
given in the inset. (j) % Killing of B16F10 cells incubated with doxorubicin loaded T1S-
vesicle, D1-CNT, T1S-vesicle-D1-CNT conjugate for 12 h with varying doxorubicin
concentration. Percent errors are within±5% in triplicate experiments.
29
This article is protected by copyright. All rights reserved.