Phenylboronic Acid Appended Pyrene-Based Low-Molecular-Weight Injectable Hydrogel: Glucose-Stimulated Insulin Release

Article abstract of DOI:10.1002/chem.201501170

A pyrene-containing phenylboronic acid (PBA) functionalized low-molecular-weight hydrogelator was synthesized with the aim to develop glucose-sensitive insulin release. The gelator showed the solvent imbibing ability in aqueous buffer solutions of pH valu

Full text of DOI:10.1002/chem.201501170

DOI: 10.1002/chem.201501170
Full Paper
&
Hydrogels |Hot Paper|
Phenylboronic Acid Appended Pyrene-Based Low-Molecular-
Weight Injectable Hydrogel: Glucose-Stimulated Insulin Release
Deep Mandal, Subhra Kanti Mandal, Moumita Ghosh, and Prasanta Kumar Das*[a]
Abstract: A pyrene-containing phenylboronic acid (PBA)
functionalized low-molecular-weight hydrogelator was syn-
thesized with the aim to develop glucose-sensitive insulin
release. The gelator showed the solvent imbibing ability in
aqueous buffer solutions of pH values, ranging from 8–12,
whereas the sodium salt of the gelator formed a hydrogel at
physiological pH 7.4 with a minimum gelation concentration
(MGC) of 5 mgmLÀ1. The aggregation behavior of this ther-
moreversible hydrogel was studied by using microscopic
and spectroscopic techniques, including transmission elec-
tron microscopy, FTIR, UV/Vis, luminescence, and CD spec-
troscopy. These investigations revealed that hydrogen bond-
ing, p–p stacking, and van der Waals interactions are the
key factors for the self-assembled gelation. The diol-sensitive
PBA part and the pyrene unit in the gelator were judiciously
used in fluorimetric sensing of minute amounts of glucose
at physiological pH. The morphological change of the gel
due to addition of glucose was investigated by scanning
electron microscopy, which denoted the glucose-responsive
swelling of the hydrogel. A rheological study indicated the
loss of the rigidity of the native gel in the presence of glu-
cose. Hence, the glucose-induced swelling of the hydrogel
was exploited in the controlled release of insulin from the
hydrogel. The insulin-loaded hydrogel showed thixotropic
self-recovery property, which hoisted it as an injectable soft
composite. Encouragingly, the gelator was found to be com-
patible with HeLa cells.
Introduction
sulin from soft materials (e.g., gels, vesicles, etc.) is a subject of
huge importance. In recent times, there is a growing impetus
in the development of stimuli-responsive gels that undergo
physicochemical changes in response to external stimuli such
as pH value, temperature, ionic strengths, biomolecules, and
so on.[5–7] These unique properties of gels were comprehen-
sively employed in molecular recognition, sensors, actuators,
drug delivery, and other domains.[3,4,8–11] The stimuli sensitivity
can be developed in the gelator scaffold by introducing task
specific functionalities.
In the twenty first century, diabetes is one of the most
common chronic disorders associated with glucose. In human
body, insulin carries glucose from food to cells.[1] In type 1 dia-
betes the insulin production from the pancreas is hampered,
ensuing a high glucose level in the bloodstream.[2] Conse-
quently, external insulin needs to be injected to sustain
a proper blood glucose level. Glucose responsive materials
have significant potential in the advancement of the insulin
regulatory systems that are needed for hyperglycemic thera-
pies.[3,4] In this perspective, a glucose-stimulated release of in-
In this regard, phenylboronic acid (PBA) is known to be an
intelligent saccharide receptor due to its unique ability to form
reversible covalent bonds with diols.[12] Most of the reported
PBA derivatives demonstrate glucose sensitivity primarily
based on fluorescent detection in solution phase at alkaline
pH values (i.e., pH 8–12).[13,14] To this end, James and coworkers
had reported a variety of boronic acid based molecular recep-
tors for selective saccharide sensing in an aqueous methanolic
buffer of pH 8.2 and a carbonate buffer of pH 10.[15,16] However,
none of the reported molecules are capable of forming a self-
assembled hydrogel. There are reports on glucose-responsive
polymeric gels that regulate either chemically modified insulin
or release insulin at higher pH-value ranges (i.e., pH 8.5–
10).[17,18] In comparison to polymeric gels, low-molecular-
weight gels (LMWGs) could be more befitting as they are re-
versible in nature and more sensitive to the changes of their
adjacent environment.[19] In addition, if the insulin regulatory
low-molecular-weight gels can be made to be injectable
through its thixotropic property, it would be highly befitting in
[a] D. Mandal, S. K. Mandal, M. Ghosh, Prof. P. K. Das
Department of Biological Chemistry
Indian Association for the Cultivation of Science
Jadavpur, Kolkata, 700032 (India)
E-mail: bcpkd@iacs.res.in
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201501170. It contains a table for the MGCs
of compound 1 at different pH values, a representation of I/I0 versus glu-
cose concentration at different peaks, a synthetic scheme, the 1H NMR spec-
troscopic and mass spectrometric analysis of compounds 1 and 1a, UV/Vis
spectra of compound 1a, the quenching of fluorescence through PET due
1
to incorporation of PBA group, H NMR spectra of compound 1a with the
glucose-containing supernatant solution on top of the gel, fluorescence
spectra of the glucose-containing supernatant solution, a calibration curve
of gelator 1a, photographs of gel 1a irradiated with a UV torch in the ab-
sence and presence of glucose after swelling, an absorption spectrum of
standard insulin, a calibration curve of insulin, the step-strain time-depen-
dent rheological analysis, and a fluorescence micrograph of the HeLa cells
in the presence of compound 1a.
Chem. Eur. J. 2015, 21, 12042 – 12052
12042
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
treating diabetic patients. Till date, no low-molecular-weight
hydrogel has been reported that demonstrates glucose sens-
ing as well as glucose-stimulated release of insulin at physio-
logical pH. Hence, we aim to design a glucose-sensitive injecta-
ble LMHG that can detect minute amount of glucose at phys-
iological pH and subsequently can be used as glucose-stimu-
lated insulin regulatory vehicle. There are several detection
methods for monitoring the saccharide content including UV/
Vis absorbance, fluorescence, circular dichroism, electrochemis-
try, equilibrium dialysis, Raman spectroscopy, nuclear magnetic
Figure 1. A) Structure of the investigated hydrogelators. B) Photograph of
the hydrogel. C) TEM image of the hydrogel 1a.
resonance
spectroscopy,
holographic detection,
and
others.[20–22] Among these techniques, one of the most promi-
nent and powerful optical read out for saccharide sensing is
fluorescent detection as it deals with very minute analyte con-
centrations and excellent sensitivity.[20] Thus, the development
of a fluorescent LMHG with a glyco-responsive PBA functionali-
ty would be helpful for the fluorimetric detection of glucose.
In the present study, we developed the pyrene-containing
PBA-functionalized low-molecular-weight hydrogelator 1 (Fig-
ure 1A) that can form a gel in aqueous buffer of pH values
ranging from 8–12. The sodium salt of the gelator, that is, com-
pound 1a, showed an efficient hydrogelation ability in a phos-
phate-buffered saline (PBS) solution of pH 7.4. Interestingly,
compound 1a can sense minute amount of glucose (0.1 mm)
at physiological pH (i.e., pH 7.4) through fluorescence detec-
tion. Upon addition of glucose (0.1–0.6 mm), the intrinsic fluo-
rescent intensity of the gelator was enhanced twofold. Impor-
tantly, the gel fibers swelled with an increase in the fiber size
in the presence of glucose. The loosening of the fibrillar net-
work owing to the swelling of this LMHG in the presence of
glucose, assisted in the controlled release of insulin entrapped
within the self-assembled fibrillar network (SAFIN) of the hy-
drogel. The most advantageous property of the insulin-loaded
hydrogel 1a was its thixotropic self-recovery property. This
unique thixotropic property of the gel–insulin composite made
it injectable and hoisted an opportunity to use it in diabetic
treatment. Moreover, the newly developed hydrogelator 1a
was found to be highly compatible towards HeLa cells, which
essentially enhanced its potential applications in biomedicine.
acid (l-phenylalanine) and a 2,2’-(ethylenedioxy)bis(ethylamine)
linker (Scheme S1 in the Supporting Information). The aroma-
ticity of the pyrene is expected to provide the necessary p–p
stacking and hydrophobic interaction required for self-assem-
bled gelation.[25] Intermolecular hydrogen bonding originated
from amide bonds of the amino acid linker might help in self-
aggregation.[25] The ethyleneoxy spacer was integrated mainly
to provide the necessary hydrophilicity in the gelator back-
bone. Aryl boronates have an intrinsic affinity towards dihy-
droxyl groups (such as glucose), which is helpful in the devel-
opment of glucose sensors. Earlier study reported that boronic
acid is a sufficiently biocompatible group with low immunoge-
nicity.[32] Hence, PBA was included at the terminal position to
bring in glycosensitivity and hydrophilicity in the structure of
the gelator.
Gelation behavior and characterization of the hydrogel
Compound 1 showed the gelation ability in aqueous buffer so-
lution of varying pH values (i.e., pH 8.5–12.0, carbonate buffer)
with the minimum gelation concentrations (MGCs) of 3.7–
6.6 mgmLÀ1 (Table S1 in the Supporting Information). However,
compound 1 could not form a hydrogel at physiological pH
(i.e., pH 7.4). Notably, the Na salt of compound 1, that is, com-
pound 1a, showed efficient hydrogelation in 25 mm phos-
phate buffer at pH 7.4 (yellow colored gel, Figure 1B) having
a MGC of 5 mgmLÀ1. Hydrogels formed at different pH values
were thermoreversible in nature, which means that they
melted by slow heating and again turned to gel on cooling.
The gels were stable at room temperature for several months.
The temperature at which the gel converts into a solution is
referred to as gel melting temperature or gel-to-sol transition
temperature (Tgel).[25] In the case for the LMHG 1a, the Tgel was
recorded to be 68oC at the MGC (i.e., 5 mgmLÀ1). Considering
the envisaged applications of the hydrogel in glucose sensing
and insulin release, our study mainly deals with the LMHG 1a
as it can form a gel at the physiologically relevant pH value
(i.e., pH 7.4).
Results and Discussion
The development of soft material (such as gels, vesicles, etc.)
based insulin regulatory vehicle that can simultaneously recog-
nize glucose and release insulin in response to an external glu-
cose stimuli is in high demand. Moreover, the development of
such stimuli-responsive soft materials with sufficient bio-com-
patibility is a challenging task particularly when we envisage
its applications in biomedicine including hyperglycemic thera-
pies. LMHGs comprised of fluorophore-tagged glyco-sensitive
motifs could be suitable to achieve the said objective. The es-
sential parameter that in general influence the self-assembled
gelation are hydrophobic interactions, p–p stacking, hydrogen
bonding, and most importantly, the optimum hydrophilic–lipo-
philic balance (HLB) in the gelator scaffold.[23–31] Considering
these factors, gelator 1 (Figure 1A) was synthesized with
a pyrene moiety coupled with phenylboronic acid by an amino
The morphological and physicochemical properties of the
hydrogel 1a were investigated by microscopic and spectro-
scopic techniques. A transmission electron microscopic (TEM)
image of the dried hydrogel confirmed the formation of an en-
tangled fibrillar network of compound 1a. The gel fibers had
an average diameter of 15–20 nm with a length of several mi-
crometers (Figure 1C). The influence of different non-covalent
Chem. Eur. J. 2015, 21, 12042 – 12052
www.chemeurj.org
12043
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
interactions operating in the self-aggregation of the gelators
was investigated by using FTIR spectroscopy. FTIR spectra were
recorded for compound 1a in the non-self-assembled state in
chloroform (CHCl3) as well as in the gel state in D2O (Fig-
ure 2A). The FTIR spectra of compound 1a in the non-gelating
solvent CHCl3 showed the transmittance peaks at approximate-
ly n˜ =3312, 1641, and 1509 cmÀ1 that were originated due to
taken place during the self-assembly of compound 1a in aque-
ous buffer. The aggregates that exhibit bathochormically shift-
ed J-band in their absorption spectrum, are called J-aggre-
gates.[33,34] It was originated due to edge-by-edge or side-to-
side assemblies.
Fluorescent spectroscopy is another important technique to
study the molecular interactions in self-assembled gelation.
The fluorescence spectra of compound 1a at very low concen-
tration (0.05 mgmLÀ1, i.e., 100 times lower than the MGC),
showed a strong emission peak at l=396 nm along with two
other peaks at l=376 and 417 nm upon excitation at l=
340 nm (Figure 2B). The characteristic of the emission spec-
trum, which is akin to the molecular pyrene, indicates that the
gelator molecule is in its non-self-assembled state at the exper-
imental concentration. With increase in the gelator concentra-
tion to 0.5 mgmLÀ1 a new excimer peak generated at a higher
wavelength (l=476 nm). At the MGC (5 mgmLÀ1) the fluores-
cence intensity of the excimer peak significantly enhanced and
further red shifted to l=519 nm. The emission intensity, due
to the molecular pyrene of compound 1a, substantially de-
creased with an increase in the gelator concentration. The ap-
pearance of the excimer peak indicates the participation of p–
p interaction between the adjacent pyrenyl moieties during
the self-assembly of compound 1a.[15,25] The generation of the
excimer peak at a concentration of approximately 0.5 mgmLÀ1
also pointed out that the self-assembly process might have
been initiated around this concentration. With further increase
in the gelator concentration, the intensity of the excimer peak
notably enhanced due to the participation of more gelators in
the self-aggregation that led to an enhancement in the p–p in-
teractions between the pyrenyl groups. Consequently, the frac-
tion of non-self-assembled molecules decreased, which result-
ed in a drastic drop in the emission peak intensity of molecular
pyrene.
Circular dichroism (CD) spectroscopy provides information
about the nature of the supramolecular aggregation depen-
ding on the molecular structures of the gelators. The CD spec-
tra of compound 1a have a negative cotton effect with double
minima at l=218 and 240 nm (Figure 3A). The peaks became
more prominent with an increase in the concentration of com-
pound 1a. The double minima in the CD spectra denote an a-
helical nature of the supramolecular chiral aggregates (not by
the positions but by its pattern).[25] From temperature-depen-
dent CD spectra it was observed that the CD peaks lost their
intensity with a rise in temperature up to 808C. Notably, the
peak intensities were regained upon cooling back to 208C (Fig-
ure 3B).
Figure 2. A) FTIR spectra of compound 1a in CHCl3. B) Fluorescence spectra
of compound 1a with varying concentrations.
nNÀH (amide A), amide nC (amide I), and nNÀH (amide II) vibra-
=
O
tions, respectively. These peaks were shifted to approximately
n˜ =3353–3532 (broad band), 1610, and 1538 cmÀ1, respectively,
upon gelation in D2O. This shift in the stretching and bending
frequencies suggests the involvement of intermolecular hydro-
gen bonding between carbonyl and amide NH groups during
the self-assembled gelation.[25] According to the UV/Vis spectra
of compound 1a, a bathochromic shift was noted when the
solvent was moved from the non-gelating CHCl3 to a gelating
one (i.e., PBS solution) (Figure S1 in the Supporting Informa-
tion). It is characteristic for a J-type aggregation, which had
Hence, the hydrophobic and p–p interactions originated
from the pyrenyl moiety, the intermolecular hydrogen bonding
between the amide linkage, and possibly the required HLB at-
tained by the presence of the hydrophilic boronic acid aided
the self-assembled hydrogelation of the PBA-tethered gelator,
1a.
Chem. Eur. J. 2015, 21, 12042 – 12052
www.chemeurj.org
12044
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 4. A) and C) Fluorescence spectra of compound 1a at different con-
centrations (A=0.05 and C=0.5 mgmLÀ1) with increasing glucose (glu) con-
centration. B) and D) Correlation curves (I/I0 vs. glucose concentration) for
the incremental change in the peak intensity at l=376 nm with fixed gela-
tor concentration of 0.05 (B) and 0.5 mgmLÀ1 (D). E) UV/Vis spectra of com-
pound 1a with varying glucose concentration. F) Photographs of a solution
of compound 1a in the absence and presence of glucose upon irradiating
with a UV torch.
Figure 3. A) Concentration-dependent CD spectra of compound 1a. B) Tem-
perature-dependent CD spectra of compound 1a (1 mgmLÀ1).
Glucose sensing
At this point, we wanted to make use of the PBA linked at the
terminal of gelator 1a for glucose sensing. The intrinsic fluo-
rescent character of the pyrene moiety in the gelator was uti-
lized for the fluorimetric detection of d-glucose. Fluorescence
spectra of compound 1a (0.05 mgmLÀ1, at the non-self-assem-
bled state) showed that the emission intensity of the pyrene
notably enhanced with the addition of very little amount of
glucose (0.1 mm, Figure 4A). The emission intensity of the ge-
lator steadily increased with an increase of the glucose concen-
tration. The maximum improvement was two-fold at 0.6 mm
and became saturated with further increase of the glucose
concentration. Importantly, minute amount of glucose can be
easily sensed through fluorescence detection by this newly de-
veloped gelator 1a at very low concentration. At the self-ag-
gregated state of compound 1a (0.5 mgmLÀ1), the intensity of
the excimer peak at l=476 nm gradually decreased, whereas
the molecular emission peak intensity (at l=376 and 417 nm)
owing to the pyrene moiety steadily improved with an in-
crease of the glucose concentration from 0.1 to 1.2 mm (Fig-
ure 4C). The highest incremental change in the fluorescence
intensity was noted for the peak appeared at l=376 nm. A
correlation between the fluorescent intensity (I/I0 vs. glucose
concentration, where I0 =fluorescence intensity of compound
1a in the absence of glucose, I=fluorescence intensity of com-
pound 1a in the presence of varying glucose concentrations)
of the peak appeared at l=376 nm and the glucose concen-
tration was given for a gelator concentration of 0.5 and
0.05 mgmLÀ1 (Figures 4B and D). A continuous enhancement
in the emission intensity profile was noted with an increase of
the glucose concentration. Also a correlation table was drawn
to compare the highest increment in peak intensity (I/I0) at dif-
ferent peaks (i.e., l=376, 417, and 476 nm, Table S2 in the
Supporting Information) at different glucose concentrations. In
the UV/Vis spectra of the gelator 1a (0.025 mgmLÀ1), the ab-
sorbance intensity of the pyrene gradually enhanced (Fig-
ure 4E) with an increase of the amount of glucose (0.1–
0.5 mm). Even the change in the bright blue fluorescence of
compound 1a in the absence and presence of glucose can be
differentiated when the solution is irradiated with a UV torch
(Figure 4F). The presence of glucose can be easily sensed visu-
ally from the bright fluorescence of the gelator compared to
that of the native gel. Besides glucose, other saccharides like
galactose and fructose that are present in blood may also in-
terfere with the fluorescence property of compound 1a. Sur-
prisingly, we found an almost unaltered fluorescence behavior
of the gelator 1a upon addition of galactose up to 15 mm.
Chem. Eur. J. 2015, 21, 12042 – 12052
www.chemeurj.org
12045
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
The fluorescence intensity of compound 1a merely enhanced
by 1.2-fold in the presence of 10 mm d-fructose. Thus, the sen-
sitivity of gelator 1a towards galactose and fructose is very
poor. Moreover, the concentrations of galactose and fructose
in blood are too low (0.1 mm) to interfere with the glucose
sensing by the fluorescence property of compound 1a.[43]
The probable reason behind the enhanced emission intensi-
ty of compound 1a in the presence of glucose could be as fol-
lows. In aqueous medium, the PBA-tethered gelator remains in
equilibrium with its neutral/trigonal from (i.e., compound 1)
and the anionic/tetrahedral form (i.e., compound 1a)
(Scheme 1).[20,35–37] The direction of the equilibrium depends on
(compound 1a) in a DMSO/water (1:3, v/v) mixture. We found
that the emission intensity of the pyrene moiety (compound
vi) was notably decreased when the PBA group was covalently
attached to it (resulting in compound 1a) (Figure S2 in the
Supporting Information). Hence, it can be affirmed that the
molecular emission behavior of pyrene gets suppressed owing
to this PET process (Scheme 1) resulting in low fluorescence. In
presence of glucose, compound 1a forms a cyclic boronate-
1,2-diol adduct (compound 1b, Scheme 1) where the overall
equilibrium is shifted to the stable anionic boronate ester form
(Scheme 1). The sp3-hybridizes boron atom of compound 1b
can no longer serve as an electron-acceptor atom. This trans-
formation of the boron atom from sp2 to sp3 obviously weak-
ens the PET process and consequently, a notable enhancement
in the fluorescence intensity of pyrene in both molecular and
self-assembled state was noted (Figures 4A and C).[20,35,38]
Moreover, in the presence of glucose the overall HLB of the ge-
lator 1b might have been altered, which would affect the
nature of self-aggregation. The altered HLB might have hin-
dered the p–p stacking interaction between the pyrene moiet-
ies. As a result, the excimer peak (Figure 4C) that appeared at
the self-aggregated state of compound 1a (0.5 mgmLÀ1) was
gradually decreased with an increase of the glucose concentra-
tion.
At this point, we were intrigued to find out the morphologi-
cal changes taking place to the self-assembled fibrillar network
(SAFIN) of the LMHG upon addition of glucose. According to
a scanning electron microscopic (SEM) study, the images
showed that the native gel had a fiber diameter of 15–20 nm
(Figure 5A). The fiber diameters gradually increased up to 40–
50, 58–70, and 80–120 nm in presence of 6, 12, and 18 mm
glucose, respectively for compound 1b (Figures 5B–D). These
size expansions of the gel fibers strongly support a glucose-in-
duced swelling behavior of the gel (Figure 5E). Presumably,
the addition of glucose altered the nature of the self-assembly
and resulted in the swelling of the hydrogel. The degree of
swelling (or mass swelling) is the most important parameter in
swelling studies. The percentage degree of swelling (% DS)
was calculated from the following Equation (1):
Scheme 1. Schematic representation of the PET process and the fluores-
cence enhancement of the gelator upon addition of glucose through the
suppression of the PET.
% DS ¼ ðMtÀM0Þ 100 M0
ð1Þ
where Mt is the mass in [mg] of the swollen gel at a time t and
M0 is the mass in [g] of the initial xerogel.[39] According to the
Figure 5F, it was clearly understood that the degree of swelling
gradually enhanced with time and became saturated when it
attained equilibrium swelling condition. This increment in
% DS was found to be more prominent with an increase of the
glucose concentration (6, 12, and 18 mm) compared to PBS so-
lution of pH 7.4.
the pKa value of PBA and the pH value of the solution. The pKa
value of PBA is approximately 8.9, which get reduced to ap-
proximately 8.0 upon linking up with an electron-withdrawing
group like amide.[36,37] At physiological pH (i.e., pH 7.4), there
will be an equilibrium between the two forms (i.e., compounds
1 and 1a) with a shift towards the neutral/trigonal boronic
acid form 1. The sp2-hybridized boron atom in neutral boronic
acid (i.e., compound 1) behaves as an electron-acceptor
moiety.[20,38] Excitation of the gelator at l=340 nm, possibly
led to a photoinduced electron transfer (PET) between the ex-
cited pyrene donor to the boron acceptor. To understand the
PET quenching of the pyrene due to the presence of PBA, we
have performed a fluorescence experiment of the gelator
moiety before (compound vi in Scheme S1 in the Supporting
Information) and after covalently attaching the PBA group
At this instant, there could be a possibility of gel dissolu-
tion/dilution due to the addition of aqueous buffer/glucose so-
lution on top of the gel during the swelling of the gel. To in-
1
vestigate this issue, we have carried out H NMR, fluorescence,
and UV absorption spectroscopy experiments (Figures S3, S4,
and S5 in the Supporting Information respectively) of the su-
pernatant solution containing 18 mm glucose (on top of the
Chem. Eur. J. 2015, 21, 12042 – 12052
www.chemeurj.org
12046
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Rheological behavior
Now, it would be intriguing to find out how the stiffness of
the gel changed upon addition of glucose to the soft materi-
als. Rheological studies provide information on the fluidity and
rigidity of viscoelastic materials such as gels. Two major param-
eters related to the viscoelasticity are the storage modulus (G’)
and the loss modulus (G’’). The storage modulus G’ refers to
the ability of a deformed material to restore its native form
whereas the loss modulus G’’ represents the flow behavior of
the material under applied stress. For viscoelastic materials like
gels G’>G’’ (G’ and G’’ꢀw0, w=angular frequency), and in the
sol state G’’>G’ (G’ꢀw2 and G’’ꢀw).[24–26] The stiffness of the
LMHG 1a in the absence and presence of glucose was deter-
mined from a rheological study (Figure 6). In a typical oscillato-
ry frequency sweep experiment of the native LMHG 1a
(5 mgmLÀ1) at a fixed strain (g, 0.01%), the storage modulus
Figure 5. SEM images of compound 1a in the presence of different glucose
concentration: A) 0, B) 6, C) 12, and D) 18 mm. E) Plot of the average gel
fiber diameters at different glucose concentrations. F) Plot of the % degree
of swelling with time at different glucose concentrations.
Figure 6. Rheology of compound 1a (5 mgmLÀ1) with increasing glucose
(glu) concentrations.
1
gel). In the H NMR experiment after 24 h of incubation no de-
tectable signal of the gelator was observed in the NMR spectra
of the supernatant solution in D2O (Figure S3 in the Supporting
Information). Similarly, in the fluorescence spectroscopic ex-
periment very poor intensity of compound 1a was observed
(Figure S4 in the Supporting Information) between 6 and 24 h
of the gel swelling experiment. The observed fluorescence in-
tensity of compound 1a in the supernatant solution is remark-
ably lower compared to that of the gel. The amount of gelator
coming out through dissolution was quantified by UV/Vis ab-
sorption by using a standard calibration curve (Figure S5 in the
Supporting Information). An amount of 1 mL of PBS solution
(pH 7.4) and 18 mm glucose solution was poured separately
over 1 mL of the gel at the MGC (5 mgmLÀ1). After 24 h, the
absorbance value of the supernatant solution was recorded at
l=340 nm and the concentrations of the dissolved compound
1a were found to be only 0.015 and 0.024 mgmLÀ1, respective-
ly. This indicated that only 0.42–0.5% of the initial gelator
amount was lost due to dissolution during swelling, which is
insignificant with respect to the gel at the MGC (5 mgmLÀ1).
We have irradiated the filtered gel with a UV torch after the
swelling experiment (Figure S6 in the Supporting Information).
Here too notable enhancement of fluorescence of the swelled
gel was observed, whereas the fluorescence intensity of the su-
pernatant glucose solution (Figure S4 in the Supporting Infor-
mation) was unaltered.
(G’ꢀ17000 Pa) was noted to be higher than the loss modulus
(G’’ꢀ4000 Pa) over the entire angular frequency. This is a typi-
cal characteristic of a viscoelastic material. Surprisingly, addi-
tion of glucose to the gel causes reduction in the G’ value. In
the presence of a 6 mm glucose solution the G’ value was re-
duced to approximately 14000 Pa for gel 1a (5 mgmLÀ1). With
an increase of the glucose content (12 and 18 mm), the G’
value further decreased to 13200 and 11 000 Pa, respectively.
This gradual decrease of the G’ value with an increasing glu-
cose concentration clearly indicates the weakening of the stiff-
ness of the gel. The decrease of the stiffness of the gluconated
gel 1a, probably originated from the change in the packing of
the fibers or the change in the charged environment of the
fibers upon binding with glucose. The decrease of the stiffness
might also be a consequence of the glucose-triggered swelling
of the gel.
Glucose-stimulated release of insulin from the hydrogel
Hydrogels are finding significant promise as drug delivery vehi-
cle.[40,41] The interstitial space within the fibrillar network of the
hydrogel might help in the entrapment of a drug and release
it at a rate, which is dependent on the porosity of the hydrogel
correlated to the extent of the crosslinked SAFIN.[40,41] Hence,
drug diffusion is supposed to be tuned by the degree of swel-
Chem. Eur. J. 2015, 21, 12042 – 12052
www.chemeurj.org
12047
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ling of the hydrogel. To this end, glucose-stimulated swelling
of the LMHG 1a (as observed from our SEM study), encour-
aged us to explore its possible application for in vitro drug re-
lease through a diffusion pathway.[40,41] Insulin was loaded
within the LMHG 1a at pH 7.4. Maximum 0.138 mgmLÀ1 of in-
sulin can be loaded within the SAFIN without disconcerting
the stability of the gel at the MGC (5 mgmLÀ1). The drug load-
ing capacity of the hydrogeln1a can be improved by increas-
ing the gelator concentration. Insulin (0.05 mgmLÀ1) loaded
hydrogel (5 mgmLÀ1, MGC) composites were used to investi-
gate the glucose-stimulated release of insulin. The amount of
released insulin was estimated from the absorbance value of
the supernatant solution at l=276 nm (Figures S7 and S8 in
the Supporting Information).[42] In the presence of only PBS so-
lution (pH 7.4) the insulin diffusion from the SAFIN was merely
13% within 24 h. Aqueous glucose solutions (600 mL) of vary-
ing concentrations (6, 12, and 18 mm) were added on the top
of each preloaded insulin–hydrogel (1a) composite and incu-
bated for several hours. The amount of released insulin in-
creased with an increase in time and after 24 h it was recorded
to be approximately 35 and about 58% in the presence of 6
and 12 mm glucose, respectively (Figure 7A). The insulin re-
effect, hypoglycemia was caused due to a rapid release of insu-
lin from drug delivery systems. Herein, the developed insulin–
gel composite could uphold the release of insulin for a long
period of time (two days) and subsequently would reduce the
chances of hypoglycemia. Although the existing results are
based on in vitro studies, their characteristics for self-regulated
insulin delivery could be very significant for hyperglycemic
treatment.
Circular dichroism (CD) study of insulin
Generally, native insulin exhibits two negative bands at l=209
and 223 nm in the CD spectra. These were referred to the a-
and b-helical structure of insulin, respectively (Figure 7B).[41,44]
The ratio of the both bands ([F209/F223]) provides a qualitative
measure of the overall conformation of insulin. In this case the
[F209/F223] values of native and released insulin were found to
be almost similar (1.16 and 1.08, respectively). Thus, no signifi-
cant conformational change took place for the released insulin
from the hydrogel network compared to that of standard one.
Thixotropy and injectable behavior of the insulin-loaded
hydrogel
A simple time-dependent step-strain experiment was executed
to look into the thixotropic behavior of the insulin
(0.138 mgmLÀ1) loaded gel at its MGC (5 mgmLÀ1). The applied
angular frequency was fixed at 1 radsÀ1 throughout the experi-
ment. Initially, 0.1% strain were applied to the insulin–gel com-
posite, at which the G’ value showed a higher account than
the G’’ value indicating distinct gel characteristics (Figure 8A).
The applied strain was then rapidly enhanced from 0.1 to 50%
and kept static for approximately 150 seconds. It was found
that the G’ value drastically decreased to a very low value
(ꢀ81 Pa), which even appears below the G’’ value (ꢀ192 Pa).
The immediate drop of the G’ value below the G’’ value con-
firms that the gel behavior was completely destroyed and sol
behavior was prominent (Figure 8A). Finally, the high strain
was removed and a constant strain of 0.1% was again applied
to observe the gel recovery property. Encouragingly, the insu-
lin–gel composite recovered about 98% of its stiffness within
approximately 80 seconds, which indicated the gradual break
down and fast recovery of the cross-linked SAFIN resulting in
sol-to-gel transition (Figure S9 in the Supporting Information).
Furthermore, the same self-recovery experiment was per-
formed four times in a cycle and in each case the stiffness was
recovered about 95% within a time interval of approximately
80–110 seconds. Thus, the gel showed a distinct thixotropic
property, which can be successfully modulated by altering the
applied strain on the soft nanocomposite.
Figure 7. A) Cumulative release percentage of insulin. B) CD spectra of
native and released insulin.
lease from the hydrogel network accelerated at higher concen-
tration of glucose. Almost 80% of the insulin were released in
24 h in the presence of 18 mm glucose. Release of insulin en-
trapped within the SAFIN through diffusion might be a conse-
quence of the glucose-induced swelling of the hydrogel. The
change in the electrical microenvironment of the fibers due to
addition of glucose might also be a driving force behind the
release of entrapped materials. The insulin release proportion-
ately increased as the degree of swelling improved with an in-
crease of the glucose concentration.[43–46] Hence, the rate of in-
sulin diffusion from the hydrogel can be tuned by varying the
glucose concentration. Generally, the human body contains
plasma glucose concentrations of approximately 5.5–6.0 mm,
whereas for diabetic patients it is above 10 mm.[43] Interesting-
ly, at higher glucose concentration (i.e., above 10 mm), com-
pound 1a can release high amount of drug (58-80%, depen-
ding on the glucose concentrations). This regulatory release of
insulin could be well-suited with the needs of the patient, as
the gel started to release insulin in less than 1 h and pro-
longed up to 48 h.[43] Often it was seen that as a frequent side
This thixotropic property of the insulin–gel composite made
it suitable to be utilized as an injectable material.[24] To demon-
strate the injectable behavior of the soft nanocomposite, we
prepared the insulin-loaded hydrogel 1a at pH 7.4 directly in
a syringe (Figure 8B). The plunger was then pressed and the
insulin–gel nanocomposite started flowing through the needle
(Figure 8C). The solution was then collected in a glass vial (Fig-
Chem. Eur. J. 2015, 21, 12042 – 12052
www.chemeurj.org
12048
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 9. MTT-based % cell viability of HeLa cells with varying concentration
of hydrogelator 1a (the incubation period was 24 h).
cein indicated the existence of live cells in the presence of
compound 1a (200 mgmLÀ1, Figure S10 in the Supporting In-
formation). Furthermore, when the images were taken by
using the excitation filter BP530-550 nm and a band absorb-
ance filter covering wavelengths below l=570 nm, negligible
red fluorescence was evinced. This result further confirmed the
abundance of live cells and reduces the possibility of gelator-
induced toxicity to mammalian cells. Hence, the newly devel-
oped LMHG 1a is inferred to be adequately cell viable.
Figure 8. A) Time-dependent repetitive cycle of the step-strain analysis of
the insulin-loaded hydrogel 1a with 5 mgmLÀ1 gelator concentration. Pho-
tographs of B) the insulin-incorporated hydrogel of 1a in a syringe, C) the in-
sulin–gel composite flowing through the needle of a syringe, D) a solution
of the insulin–gel composite after syringe processing, and E) the insulin-
loaded hydrogel 1a.
Conclusion
ure 8D) and turned into a gel within a very short time of ap-
proximately 4 min (Figure 8E). Thus, the developed soft insu-
lin–gel composite has immense potential as an injectable sub-
stance in diverse biomedicinal applications.
In summary, a pyrene-based PBA-containing injectable LMHG
was developed and utilized to detect minute amounts of glu-
cose through fluorescence sensing at physiological pH. Impor-
tantly, glucose-stimulated swelling of the hydrogel facilitated
in vitro controlled release of entrapped insulin. Encouragingly,
the gelators showed sufficient cell viability towards mammali-
an HeLa cells. Hence, the LMHG was employed in glucose
sensing as well as glycol-sensitive insulin release at physiologi-
cal pH. The stimuli-sensitive LMHG will have notable impor-
tance in the development of insulin regulatory systems re-
quired for diabetic therapies.
Cell viability study
Hydrogel 1a would find practical relevance in future applica-
tions if it shows considerable biocompatibility towards mam-
malian cells. Thus, we investigated the cell viability of com-
pound 1a towards a human cell line, that is, HeLa cells, by
a MTT-based assay.[47] A gelator solution in PBS (pH 7.4) at
a concentration range of 5–200 mgmLÀ1 was incubated with
HeLa cells. The hydrogelator 1a showed significant cell viability
after an incubation period of 24 h. At a gelator concentration
of 5 mgmLÀ1, more than 94% of the cells were found alive
(Figure 9). At a gelator concentration of 10 mgmLÀ1, the cell vi-
ability was also noted to be 94%, which merely reduced to
90% upon increasing the concentration of compound 1a to
50 mgmLÀ1. At moderately higher concentration of compound
1a (100 and 200 mgmLÀ1), 90 and 88% cell viability were ob-
served, respectively. Even at very high concentration of com-
pound 1a, that is, 2000 mgmLÀ1 (that can be highest achieved
without disconcerting the stability of the cell culture media),
85% of the cells were found to be alive. The biocompatibility
was further tested by using the LIVE/DEAD viability kit for
mammalian cells.[47,48] The exclusive presence of only a green
fluorescence (excitation filter of BP460-495 nm and a band ab-
sorbance filter covering wavelengths below l=505 nm) ensu-
ing from the enhanced fluorescence of DNA-intercalated cal-
Experimental Section
Materials: l-Phenylalanine, tert-butyloxycarbonyl (Boc) anhydride,
dicyclohexylcarbodiimide (DCC), 4-N,N-(dimethyl)aminopyridine
(DMAP), 1-hydroxybenzotriazole (HOBT), N-hydroxysuccinimide
(NHS), 4-carboxyphenylboronic acid, and solvents were procured
from SRL, India. Trifluoroacetic acid (TFA) and sodium hydroxide
(NaOH) were procured from Spectrochem, India. Pyrenebutyric acid
and 2,2’-(ethylenedioxy)bis(ethylamine) were bought from Sigma.
All deuterated solvents for NMR and FTIR spectroscopic experi-
ments and 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium
bromide (MTT) were obtained from Sigma–Aldrich Chemical Co. In-
sulin (concentration=40 IUmLÀ1
) was purchased from Torrent
Pharmaceuticals Ltd. The LIVE/DEAD viability kit (l-3224) for mam-
malian cells used for the cell viability assays was procured from
Molecular Probes, Invitrogen Chemical Company. High glucose Dul-
becco’s modified eagle medium (DMEM), heat-inactivated fetal
bovine serum (FBS), and trypsin–ethylenediaminetetraacetic acid
Chem. Eur. J. 2015, 21, 12042 – 12052
www.chemeurj.org
12049
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
(EDTA) (0.05%) were purchased from Invitrogen. Thin layer chro-
matography was done on Merck precoated silica gel 60-F254 plates.
Circular dichroism (CD) spectra: CD spectra of aqueous solutions
of gelator 1a and insulin were recorded with varying concentra-
tions and varying the temperature in a quartz cuvette (1 mm path
length) on a JASCO J-815 spectropolarimeter.
1H NMR spectra were recorded by using an AVANCE 500 MHz
(Bruker) spectrometer. Mass spectrometric data were acquired by
electron spray ionization technique on a Q-tof-micro quadruple
mass spectrometer (Micromass). Elemental analyses were per-
formed by using a Perkin–Elmer 2400 CHN analyzer. Transmission
electron microscopy (TEM) images were taken on a JEOL JEM 2010
high-resolution microscope operating at 200 kV. Scanning electron
microscopy (SEM) study was done on a JEOL-6700F microscope. A
Perkin–Elmer Spectrum 100 FTIR Spectrometer was used to record
the FTIR spectra. UV/Vis absorption spectra of the gelator 1a and
insulin were recorded on a Perkin–Elmer Lambda25 spectropho-
tometer..
Measurement of the degree of swelling: To determine the swel-
ling parameters, the xerogel (dry gel) of average weight of 5 mg
was subjected to swell in aqueous PBS solution and glucose solu-
tions of different concentrations from 6–18 mm. A piece of the xe-
rogel was weighed and immersed in an excess of the swelling
medium (PBS solution and glucose with different concentrations).
At predetermined time intervals, the swollen gel was carefully re-
moved from the medium and gently dried with a paper tissue in
order to remove excess liquid. Subsequently, it was weighed very
carefully until a constant mass was attained. Generally, the swollen
gels are fragile in nature. Hence, they were kept on a grid boat
with a mesh size of 1 mm. This process allowed the gel to be
placed in water without hampering the stability. The percentage
degree of swelling (% DS) was determined by using Equation (1).
The % DS was plotted as a function of time.
Synthetic procedure: Gelators 1 and 1a were synthesized by
using the following method (Scheme S1 in the Supporting Infor-
mation). Briefly, pyrene butyric acid (i) was coupled with the
methyl ester of l-phenylalanine (ii) in dichloromethane by using
DCC (1.1 equiv), a catalytic amount of DMAP, and HOBT (1.1 equiv)
following the reported protocol. The coupled product iii was sub-
jected to hydrolysis with 1n NaOH solution followed by work up
with 1n HCl. The free acid terminal end of the l-amino acid iv was
further coupled with mono Boc-protected 2,2’-(ethylenedioxy)bis-
(ethylamine) to get the product v. Boc protection was removed by
stirring the compound with TFA. The purified product vi was ob-
tained by column chromatography by using 60–120 mesh silica gel
and 10% methanol in chloroform. On the other hand, the NHS-
linked 4-carboxyphenylboronic acid was prepared by using 4-car-
boxyphenylboronic acid, DCC (1.5 equiv), and NHS (1.1 equiv) in
dry DMF (8 mL) and stirring overnight under a N2 atmosphere. To
this solution, an activated 4-carboxyphenylboronic acid mixture,
compound vi. and dry pyridine were added. The solution was
stirred overnight and the DMF was distilled out under vacuum.
The residue mixture was then purified through column chromatog-
raphy by using 100–200 mesh silica gel and methanol (3%)/chloro-
form as the eluent to obtain pure compound 1. The terminal bor-
onic acid was converted to the corresponding sodium salt (i.e.,
compound 1a) by adding one equivalent 0.1n NaOH (standar-
dized) to the methanolic solution of the acid 1. After brief stirring,
the solvent was removed and lyophilized in Virtis4KBTXL-75 freeze
drier under vacuum to get the sodium salt 1a. 1H NMR spectro-
scopic and mass spectrometric analysis of the gelators 1 and 1a
are provided in the Supporting Information.
To investigate the extent of dissolution during swelling, initially
1 mL of the PBS solution (pH 7.4) and the glucose solution (18 mm)
were poured separately over 1 mL of the gel (prepared at its MGC
of 5 mgmLÀ1). These were kept undisturbed for 24 h. After filtra-
tion, fluorescence spectra of the supernatant solutions were ob-
served at a certain time interval. The UV/Vis absorbance value of
the supernatant solution was also recorded at l=340 nm and the
concentration of the dissolved compound 1a was determined
from the standard calibration curve. We have also analyzed the
1H NMR spectra of the supernatant solution. Here D2O was used as
the gelation solvent and supernatant solution keeping the whole
experimental procedure same.
Rheology: The rheological experiments were carried out in cone
and plate geometry (diameter=40 mm) on the rheometer plate by
using an Anton Paar, MCR 302. The native gel was scooped on the
rheometer plate so that there was no air gap with the cone. A fre-
quency sweep experiment was done as a function of angular fre-
quency (0.1–500 radsÀ1) at a fixed strain of 0.01% at 258C and the
storage modulus (G’) and the loss modulus (G“) were plotted
against the angular frequency (w).
Loading of insulin within hydrogel: For the preparation of an in-
sulin-loaded hydrogel, the required amount of insulin (35–100 mL)
(from 1.388 mgmLÀ1 or 40 IUmlÀ1 stock solution in saline water)
was added to 1 mL of the gelator 1a (5 mgmLÀ1) in PBS solution
(pH 7.4), and then the mixture was kept undisturbed until gel for-
mation.
Preparation of the hydrogel: The requisite amount of compound
1 was taken in a screw capped vial having an internal diameter
(i.d.) of 10 mm and slowly heated to dissolve in aqueous buffer sol-
utions of different pH values (i.e., pH 8.5–12). On the other hand,
the gelation efficiency of compound 1a was tested in PBS solution
of pH 7.4. The solution was then allowed to cool slowly (undistur-
bed) to room temperature. The gelation was checked by “stable to
inversion” of the aggregated material in the glass vial.
Release of entrapped insulin from hydrogel: Glucose solutions of
different concentrations (6, 12, and 18 mm) were prepared in PBS
solution of pH 7.4. These glucose solutions (600 mL) were added on
top of the insulin (0.05 mgmLÀ1)–gel 1a (5 mgmLÀ1) composite
and incubated for several hours. The supernatant solution was re-
moved each time, and the amount of released material was esti-
mated from the absorbance of insulin at l=276 nm plotted in
a standard calibration curve. The release rate was investigated in
triplicates at room temperature and averaged.
Determination of the gel-to-sol transition temperature (Tgel): The
gel-to-sol transition temperature (Tgel) was recorded by gradually
increasing the temperature (by using a rate of 28CminÀ1) of the
thermostatted oil bath in which the hydrogel-containing glass vial
(i.d. 10 mm) was placed. The temperature (Æ0.58C) at which the
Thixotropic property of the insulin-hydrogel composite: The
continuous step-strain experiment was performed by first breaking
the gel by application of a continuous strain of 0.1–50%. After the
complete breakdown of the gel, as denoted by G’’>G’, gel recov-
ery was attempted at a constant strain of 0.1%. The process was
repeated to check the reversibility of the restoration process. The
entire study was performed by keeping a constant angular fre-
gel liquefied and started to flow was referred to as Tgel
.
Fluorescence spectroscopy: The emission spectra of the gelator at
varying concentrations were recorded on a Varian Cary Eclipse lu-
minescence spectrometer. Solutions were excited at lex =340 nm
in the absence and presence of glucose (0.1–1.2 mm). The excita-
tion and emission slit width were both 5 nm.
quency of 1 radsÀ1
.
Chem. Eur. J. 2015, 21, 12042 – 12052
www.chemeurj.org
12050
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[2] S. Vilaró, M. Palacín, P. F. Pilch, X. Testar, A. Zorzano, Nature 1989, 342,
798–800.
MTT assay: Cell viability of the hydorogelator was assessed by the
microculture MTT reduction assay as previously reported.[49] The
assay involves the conversion of a soluble tetrazolium salt by the
mitochondrial dehydrogenase present in live cells to an insoluble
colored product, that is, formazan. The amount of formazan
formed is estimated spectrophotometrically after dissolution of the
product in DMSO. The amount of formazan produced, is propor-
tional to the number of live cells. The mammalian cells were
seeded at a density of 15000 cells per well in a 96-well microtiter
plate 18–24 h before the assay. A stock solution of the hydrogela-
tor 1a in PBS solution was prepared. The concentration of the ge-
lator in the microtiter plate was varied from 5 to 200 mgmLÀ1. The
cells were incubated with the gelator for 24 h at 37oC under 5%
CO2. The cells were further incubated for another four hours in
150 mL of the MTT stock solution (5 mgmLÀ1). The produced forma-
zan was dissolved in DMSO and the absorbance at l=570 nm was
measured by using a BioTek Elisa Reader. The number of surviving
cells was expressed as percent: viability=(A570(treated cells)Àback-
ground)/A570(untreated cells)Àbackground)”100).
Fluorescence microscopic study: The LIVE/DEAD Viability/Cytotox-
icity Kit for mammalian cells was used to examine the cell viability
under a fluorescence microscope. The kit contains a mixture of
two nucleic acid binding strains, calcein AM (component A) and
ethidium homodimer-1 (component B). The acetomethoxy deriva-
tive of calcein (calcein AM) has the ability to pass through the cell
membrane. The esterase enzyme present in live cells removes the
acetoxy group. This form of the compound then intercalates with
the DNA and resulted in an enhancement of fluorescence and
a bright green color is observed. The ethidium homodimer-
1 (EthD-1) can only pass through damaged cell membranes, thus it
gets incorporated only into dead cells. It shows red fluorescence
upon binding with DNA in dead cells. The kit was stored at À208C
in the dark and was taken out and thawed at room temperature
just prior to assay. An aliquot (4 mL) of the supplied 2 mm EthD-
1 stock solution (component B) was added to 2 mL of sterile, tissue
culture-grade PBS solution and the mixture was vortexed to ensure
thorough mixing. An aliquot (1 mL) of the supplied 4 mm calcei-
n AM stock solution (component A) was then added to the 2 mL
EthD-1 solution and vortexed. The resulting working solution (2 mm
calcein AM and 4 mm EthD-1) was then added directly to HeLa cells
treated with the gelator solutions (200 mgmLÀ1 for 24 h) and incu-
bated further for 3 h. After incubation the excess dye was removed
from the extracellular medium and the cells were washed with PBS
solution. Then the cells were observed under the Olympus IX51 in-
verted microscope by using an excitation filter of BP460–495 nm
and a band absorbance filter covering wavelengths below l=
505 nm and the excitation filter BP530–550 nm and a band absorb-
ance filter covering wavelengths below l=570 nm.
[3] R. Ma, L. Shi, Polym. Chem. 2014, 5, 1503–1518.
[4] D. Roy, B. S. Sumerlin, ACS Macro Lett. 2012, 1, 529–532.
[5] Y. Qiu, K. Park, Adv. Drug Delivery Rev. 2001, 53, 321–339.
[6] M. D. Segarra-Maset, V. J. Nebot, J. F. Miravet, B. Escuder, Chem. Soc. Rev.
2013, 42, 7086–7098.
[7] T. Miyata, T. Uragamia, K. Nakamae, Adv. Drug Delivery Rev. 2002, 54,
79–98.
[8] M. Ikeda, R. Ochia, I. Hamachi, Lab. Chip 2010, 10, 3325–3334.
[9] J. D. Decoppet, T. Moehl, S. S. Babkair, R. A. Alzubaydi, A. A. Ansari, S. S.
Habib, S. M. Zakeeruddin, H. W. Schmidt, M. Grätzel, J. Mater. Chem. A
2014, 2, 15972–15977.
[10] J. Huang, C. L. Hastings, G. P. Duffy, H. M. Kelly, J. Raeburn, D. J. Adams,
A. Heise, Biomacromolecules 2013, 14, 200–206.
[11] A. Bieser, J. Tiller, Supramol. Chem. 2008, 20, 363–367.
[12] T. D. James, K. R. A. S. Sandanayake, S. Shinkai, Angew. Chem. Int. Ed.
Engl. 1996, 35, 1910–1922; Angew. Chem. 1996, 108, 2038–2050.
[13] S. Arimori, C. J. Ward, T. D. James, Tetrahedron Lett. 2002, 43, 303–305.
[14] Z. Wu, X. Zhang, H. Guo, C. Li, D. Yu, J. Mater. Chem. 2012, 22, 22788–
22796.
[15] M. D. Phillips, T. M. Fyles, N. P. Barwell, T. D. James, Chem. Commun.
2009, 6557–6559.
[16] Y. J. Huang, W. J. Ouyang, X. Wu, Z. Li, J. S. Fossey, T. D. James, Y. B.
Jiang, J. Am. Chem. Soc. 2013, 135, 1700–1703.
[17] T. Hoeg-Jensen, S. Ridderberg, S. Havelund, L. Schäffer, P. Balschmidt, I.
Jonassen, P. Vedsø, P. H. Olesen, J. Markussen, J. Pept. Sci. 2005, 11,
339–346.
[18] K. Kataoka, H. Miyazaki, M. Bunya, T. Okano, Y. Sakurai, J. Am. Chem. Soc.
1998, 120, 12694–12695.
[19] D. J. Abdallah, R. G. Weiss, Adv. Mater. 2000, 12, 1237–1247.
[20] V. Tharmaraj, K. Pitchumani, RSC Adv. 2013, 3, 11566–11570.
[21] A. M. Horgan, A. J. Marshall, S. J. Kew, K. E. S. Dean, C. D. Creasey, S. Ka-
bilan, Biosens. Bioelectron. 2006, 21, 1838–1845.
[22] V. L. Alexeev, A. C. Sharma, A. V. Goponenko, S. Das, I. K. Lednev, C. S.
Wilcox, D. N. Finegold, S. A. Asher, Anal. Chem. 2003, 75, 2316–2323.
[23] N. Javid, S. Roy, M. Zelzer, Z. Yang, J. Sefcik, R. V. Ulijn, Biomacromole-
cules 2013, 14, 4368–4376.
[24] S. K. Mandal, S. Brahmachari, P. K. Das, ChemPlusChem 2014, 79, 1733–
1746.
[25] D. Mandal, T. Kar, P. K. Das, Chem. Eur. J. 2014, 20, 1349–1358.
[26] A. Bernet, M. Behr, R. Q. Albuquerque, M. Schmidt, J. Senker, H. W.
Schmidt, Prog. Colloid Polym. Sci. 2013, 140, 1–13.
[27] J. Boekhoven, J. M. Poolman, C. Maity, F. Li, L. van der Mee, C. B. Minken-
berg, E. Mendes, J. H. van Esch, R. Eelkema, Nat. Chem. 2013, 5, 433–
437.
[28] S. K. M. Nalluri, N. Shivarova, A. L. Kanibolotsky, M. Zelzer, S. Gupta,
P. W. J. M. Frederix, P. J. Skabara, H. Gleskova, R. V. Ulijn, Langmuir 2014,
30, 12429–12437.
[29] J. Raeburn, G. Pont, L. Chen, Y. Cesbron, R. Lffiꢀvy, D. J. Adams, Soft
Matter 2012, 8, 1168–1174.
[30] V. J. Nebot, J. J. Ojeda-Flores, J. Smets, S. Fernndez-Prieto, B. Escuder,
J. F. Miravet, Chem. Eur. J. 2014, 20, 14465–14472.
[31] J. M. Poolman, J. Boekhoven, A. Besselink, A. G. L. Olive, J. H. van Esch,
R. Eelkema, Nat. Protoc. 2014, 9, 977–988.
[32] Z. Wu, S. Zhang, X. Zhang, S. Shu, T. Chu, D. Yu, J. Pharm. Sci. 2011, 100,
2278–2286.
Acknowledgements
[33] Y. Hirano, Y. Tokuoka, N. Kawashima, Y. Ozaki, Vib. Spectrosc. 2007, 43,
86–96.
[34] A. D’Urso, M. E. Fragala, R. Purrello, Chem. Commun. 2012, 48, 8165–
8176.
[35] A. J. Tong, A. Yamauchi, T. Hayashita, Z. Y. Zhang, B. D. Smith, N. Tera-
mae, Anal. Chem. 2001, 73, 1530–1536.
P.K.D. is thankful to the Department of Science and Technology,
India (SR/S1/OC-25/2011) for financial assistance. D.M., S.K.M.,
and M.G. acknowledge the Council of Scientific and Industrial
Research, India for research fellowships.
[36] J. Yan, G. Springsteen, S. Deeter, B. Wang, Tetrahedron 2004, 60, 11205–
11209.
[37] A. Matsumoto, R. Yoshida, K. Kataoka, Biomacromolecules 2004, 5,
1038–1045.
Keywords: glucose · hydrogels · phenylboronic acid · self-
assembly · sensors
[38] L. N. Neupane, C. R. Lohani, J. Kim, K. H. Lee, Tetrahedron 2013, 69,
11057–11063.
[39] E. Karadag, D. Saraydin, Turk. J. Chem. 2002, 26, 863–875.
[1] A. M. Lowman, M. Morishita, M. Kajita, T. Nagai, N. A. Peppas, J. Pharm.
Sci. 1999, 88, 933–937.
Chem. Eur. J. 2015, 21, 12042 – 12052
www.chemeurj.org
12051
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[40] a) R. A. Siegel, Y. Gu, A. Baldi, B. Ziaie, Macromol. Symp. 2004, 207, 249–
256; b) A. Shome, S. Debnath, P. K. Das, Langmuir 2008, 24, 4280–4288.
[41] S. Payyappilly, S. Dhara, S. Chattopadhyay, J. Biomed. Mater. Res. 2014,
102, 1500–1509.
[47] S. Brahmachari, D. Das, P. K. Das, Chem. Commun. 2010, 46, 8386–8388.
[48] M. Ghosh, S. Brahmachari, P. K. Das, Macromol. Biosci. 2014, 14, 1795–
1806.
[49] M. B. Hansen, S. E. Nielsen, K. Berg, J. Immunol. Methods 1989, 119,
[42] B. Yilmaz, Y. Kadioglu, J. Pharmaceuticals 2012, 2, 21–29.
[43] W. Wu, N. Mitra, E. C. Y. Yan, S. Zhou, ACS Nano 2010, 4, 4831–4839.
[44] L. Sun, X. Zhang, C. Zheng, Z. Wu, X. Xia, C. Li, RSC Adv. 2012, 2, 9904–
9913.
203–210.
[45] Q. Wu, L. Wang, H. Yu, J. Wang, Z. Chen, Chem. Rev. 2011, 111, 7855–
7875.
[46] V. Lapeyre, I. Gosse, S. Chevreux, V. Ravaine, Biomacromolecules 2006, 7,
3356–3363.
Received: March 25, 2015
Revised: May 14, 2015
Published online on July 14, 2015
Chem. Eur. J. 2015, 21, 12042 – 12052
www.chemeurj.org
12052
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim