Two approaches for the engineering of homogeneous small-molecule hydrogels

Article abstract of DOI:10.1039/c3sm50324k

Small-molecule hydrogelators have been widely used to engineer supramolecular hydrogels for biomedical applications. Typically, a change of the solubility of small molecules in solvent is used to trigger the gelation process. This requires a switch of pH or solvent by mixing two different types of solutions. However, due to the intrinsic ragged free energy landscape that underlies the self-assembly process and the high viscosity of the solution that limits the diffusion, the hydrogels made by these methods are often limited by their inhomogeneity and irreproducible physical properties. It is therefore desirable to circumvent these drawbacks and produce homogeneous hydrogels. Conversely, only a few studies have been done towards this direction. In this article, we present two novel approaches to engineer homogeneous hydrogels. One is based on the nano-dispersed colloids to hydrogel transition and the other is based on the decomposition of potassium persulfate to mildly change the pH. These two methods allow kinetically controlling the self-assembly process and the resulting hydrogels are indeed more homogeneous and reproducible. Moreover, the structural and morphological characterizations suggest that the structures of the hydrogels prepared by different approaches are distinct from each other, leading to diverse macroscopic mechanical properties. These results suggest that besides the thermodynamics, the self-assembly kinetics also plays an important role in determining the properties of the final assembled hydrogels. We propose that it is possible to rationally tune the physical properties of the hydrogels by simply control the self-assembly kinetics without changing the structure of the small-molecule hydrogelators.

Full text of DOI:10.1039/c3sm50324k

Soft Matter
View Article Online
View Journal | View Issue
PAPER
Two approaches for the engineering of homogeneous
small-molecule hydrogels†
Cite this: Soft Matter, 2013, 9, 4672
a
a
a
b
a
a
*
*
Baiyong Ding,‡ Ying Li,‡ Meng Qin, Yin Ding, Yi Cao and Wei Wang
Small-molecule hydrogelators have been widely used to engineer supramolecular hydrogels for biomedical
applications. Typically, a change of the solubility of small molecules in solvent is used to trigger the gelation
process. This requires a switch of pH or solvent by mixing two different types of solutions. However, due to
the intrinsic ragged free energy landscape that underlies the self-assembly process and the high viscosity of
the solution that limits the diffusion, the hydrogels made by these methods are often limited by their
inhomogeneity and irreproducible physical properties. It is therefore desirable to circumvent these
drawbacks and produce homogeneous hydrogels. Conversely, only a few studies have been done
towards this direction. In this article, we present two novel approaches to engineer homogeneous
hydrogels. One is based on the nano-dispersed colloids to hydrogel transition and the other is based on
the decomposition of potassium persulfate to mildly change the pH. These two methods allow
kinetically controlling the self-assembly process and the resulting hydrogels are indeed more
homogeneous and reproducible. Moreover, the structural and morphological characterizations suggest
that the structures of the hydrogels prepared by different approaches are distinct from each other,
leading to diverse macroscopic mechanical properties. These results suggest that besides the
thermodynamics, the self-assembly kinetics also plays an important role in determining the properties of
the final assembled hydrogels. We propose that it is possible to rationally tune the physical properties of
the hydrogels by simply control the self-assembly kinetics without changing the structure of the small-
molecule hydrogelators.
Received 31st January 2013
Accepted 5th March 2013
DOI: 10.1039/c3sm50324k
www.rsc.org/softmatter
naturally occurring biomacromolecules (e.g. proteins),36 which
shows high delity and reproducibility, the hydrogels prepared
Introduction
Hydrogels constructed using small-molecule gelators have been
widely used as attractive biomaterials for a wide range of
applications, such as tissue engineering, drug delivery, wound
healing, cell therapy, etc.1–23 By tailoring their structures, various
small molecules can be rationally engineered to self-assemble
into macroscopic hydrogels through different intermolecular
interactions.24–35 Amphiphilic peptides are a representative class
of small-molecule gelators that have a hydrophobic moiety
attached to the N-terminus of a short peptide. Hydrophobic
interactions and p–p stacking provide the driving forces for the
assembly process while the charged or hydrophilic C-terminus
of the peptide maintains the hydration of the entire assembled
network. However, different from the self-assembly of many
from small-molecule gelators could show distinct physical
properties, depending on how they are prepared. For example, it
was found that the hydrogel formed by Fmoc-FF (Scheme 1), a
diphenylalanine with
a uorenylmethoxycarbonyl (Fmoc)
moiety linked to the N-terminus, can show distinct mechanical
and optical properties, depending on the preparation
methods.37–52 Moreover, the mechanical properties of hydrogels
prepared using the same approach could differ from batch to
batch.43,53 These undesired features greatly hinder the applica-
tions of small-molecule hydrogels. There are two major reasons
aNational Laboratory of Solid State Microstructure, Department of Physics, Nanjing
University, 22 Hankou Road, Nanjing, Jiangsu, 210093, P. R. China. E-mail:
yingliubc@gmail.com; wangwei@nju.edu.cn
bState Key Laboratory of Analytical Chemistry for Life Science, School of Chemistry and
Chemical Engineering, Nanjing University, 22 Hankou Road, Nanjing, Jiangsu,
210093, P. R. China
† Electronic supplementary information (ESI) available: Characterization of
Fmoc-FF and supplementary gures. See DOI: 10.1039/c3sm50324k
‡ These authors contributed equally to this work.
Scheme 1 Structure of Fmoc-FF.
4672 | Soft Matter, 2013, 9, 4672–4680
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
that may account for this: rst, the energy landscape for the self- based on a colloid–gel transition and the second method is
assembly process of small-molecule gelators is rugged and can based on the decomposition of potassium persulfate (K2S2O8)
be trapped into various local minima, as depicted in Fig. 1. Each triggered pH decrease. Both of these two methods could provide
self-assembly pathway may involve multiple intermediate states homogeneous hydrogel with reproducible physical and chem-
and the inter-conversion between these intermediate states is ical properties. Our characterization indicated that the struc-
difficult. Moreover, slight changes of the environmental tures of the Fmoc-FF gels prepared using these two new
conditions can lead to dramatic shis of the self-assembly methods are different from those prepared using other
pathways and result in different types of hydrogel struc- methods. Consequently, these hydrogels also demonstrated
tures.48,53–56 Therefore, the resulting hydrogels rely on both the distinct mechanical properties. Therefore, we propose that
thermodynamics and the kinetics of the assembly. Second and besides directly changing the peptide sequence,61 it is also
more importantly, the production of uniform hydrogels is possible to engineer hydrogels with different physical proper-
limited by the diffusion of peptides into water medium.43 The ties by simply controlling the self-assembly kinetics.
hydrogel is oen prepared by a sol–gel transition process, in
which small-molecule gelators are initially dissolved in water at
a high pH or in organic solvent. To trigger the self-assembly of
small-molecule gelators in solution, a pH adjustment4,39,42,53,57
Experimental
Chemicals
or solvent switch29,41,58–60 is used. Due to the fast self-assembly
speed (less than a second), the gelation can occur locally at the
place where two different solutions are mixed, leading to
inhomogeneous hydrogels. Hence, it is desirable to develop
methods that can produce homogeneous gel in a controllable
fashion.
2-Chlorotrityl chloride resin and Fmoc-Phe-OH were purchased
from GL Biochem (Shanghai) Ltd; 1,1,1,3,3,3-hexauoro-2-
propanol (HFIP) and dimethylsulfoxide (DMSO) were purchased
from Sigma-Aldrich. O-Benzotriazole-N,N,N0,N0-tetramethyl-
uronium-hexauoro-phosphate (HBTU) was obtained from
NovaBioChem. Other chemicals were of analytical grade and
obtained from either Sinopharm Chemical Reagent Co. Ltd or
Nanjing Chemical Reagent Co. Ltd. All chemicals were used
directly without further purication.
Towards this goal, several strategies have been developed.
Tang et al. proposed a stepwise pH-adjustment followed by a
heating and cooling process to prepare homogeneous Fmoc-FF
gel.53 Adams and coworkers introduced a strategy to slowly
decrease the pH by hydrolysis of glucono-d-lactone. They
recently also reported a method to lower the pH to trigger the
gelation of small-molecule gelators by using light triggered
decomposition of diphenyliodonium nitrate.44 These elegant
studies used physical or chemical ways to make the pH change
in the gel system more homogeneous and controllable. The
structure and mechanical properties of hydrogels prepared
using these methods were indeed more reproducible and
consistent. However, the methods for the preparation of
homogeneous and reproducible small-molecule hydrogels are
still very limited. To expand the toolbox for the engineering of
homogeneous hydrogels, here we present two novel approaches
that can also lead to homogeneous hydrogel. The rst method is
Synthesis of Fmoc-FF dipeptide
Fmoc-FF (Scheme 1) was synthesized according to the common
solid phase peptide synthesis protocol.62–64 Briey, the rst
amino acid Fmoc-Phe-OH was attached onto the resin by
treating 2-chlorotrityl chloride resin (3 mmol) with Fmoc-Phe-
OH (1.55 g, 4 mmol) and diisopropylethylamine (DIPEA, 2.8 mL,
16 mmol) in dichloromethane (DCM, 20 mL) at room temper-
ature for 1 h. Then the mixture was ltered. The resin was
further capped with 1 : 2 : 17 (v/v/v) DIPEA–MeOH–DCM (3 ꢀ
20 mL), thoroughly washed with DCM, dimethylformamide
(DMF) and DCM, and then dried in vacuo to remove the residual
volatiles. The bead-loading was measured to be ꢁ0.5 to
0.7 mmol gꢂ1. Then the Fmoc-Phe loaded beads were pre-
swollen in DMF for 0.5 h. The Fmoc group was rst removed at
room temperature by 20% piperidine–DMF (v/v, 3 ꢀ 5 mL,
5 min) and the beads were thoroughly washed with DCM and
DMF. Then 5 mL DMF solution of Fmoc-Phe-OH (4 eq.), HBTU
(4 eq.), and DIPEA (8 eq.) was added to the resin. The mixture
was agitated at room temperature for 2 h. The mixture was
ltered and the resin was thoroughly washed with DMF and
DCM. Then the bead was incubated with 20% HFIP in DCM (50
mL) at room temperature for 30 min to cleave the peptide from
the beads and ltered. The ltrate was then concentrated
in vacuo. The residue was washed with ether to yield the peptide
Fmoc-Phe-Phe-OH (Fmoc-FF). The obtained Fmoc-FF was
characterized by NMR, mass spectroscopy, and HPLC (Fig. S1†).
Fig. 1 Schematic illustration of the energy landscape underlying the self-
assembly of small-molecule gelators, which is distinct from the funnel-like free
energy landscape of protein folding.36 I1, I2, I3, I4, I5, and I6 represent intermediate
states along the self-assembly pathway.
Preparation of Fmoc-FF hydrogel
DMSO exchange approach (the DMSO method). The DMSO
stock solution of Fmoc-FF (100 mg mLꢂ1) was freshly prepared
This journal is ª The Royal Society of Chemistry 2013
Soft Matter, 2013, 9, 4672–4680 | 4673

View Article Online
Soft Matter
Paper
prior to use. The stock solution was then diluted with doubly Fmoc-FF (5 mg mLꢂ1) on Zetasizer ZS90 (Malvern) with a
distilled water (ddH2O) containing 1 equivalent Na2CO3 to a 633 nm laser source. Zeta potential measurements were also
nal concentration of 5 mg mLꢂ1 and quickly mixed by vor- carried out on Zetasizer ZS90 (Malvern) at room temperature.
texing. The hydrogel formed in a few seconds, and was aged at The colloid solution of Fmoc-FF (5 mg mLꢂ1) was directly used
room temperature for one day without disturbance before for zeta potential measurement without dilution. The reported
characterization.
value is the average of 3 independent measurements.
Dropwise addition of aqueous acid (the HCl method). Fmoc-
Rheology. Rheological measurements were carried out by
FF powder was rst dissolved in ddH2O (5 mg mLꢂ1) by adding using a Haake RheoStrss 6000 rheometer at 25 C. The hydro-
0.5 M NaOH to the solution until the pH reached ꢁ9. Then, the gels were oil-sealed to prevent evaporation. The measuring
pH was lowered by dropwise addition of 0.1 M HCl to the system consisted of a cone and plate sensor with a diameter of
system. Upon the addition of HCl, the viscosity of the solution 20 mm and cone angle of 1ꢃ.
ꢃ
increased gradually. The solution turned to an opaque hydrogel
Circular dichroism (CD). CD spectra were measured using a
at pH ꢁ7. In order to make the system as homogeneous as JASCO J-815 CD spectropolarimeter (Jasco, Japan). The hydrogel
possible, vigorous mixing by ultrasonic mixing or vortexing was samples (5 mg mLꢂ1) were directly measured in a quartz cell of
applied while adjusting the pH. The gels were aged at room 0.1 mm path length with a wavelength scan from 300 to 190 nm
temperature for one day without disturbance before at room temperature. All spectra were baseline corrected before
characterization.
A colloid-to-hydrogel transition (the colloid method). To
obtain the Fmoc-FF colloid (5 mg mLꢂ1), Fmoc-FF powder was
analysis.
The reported spectra were the average of 3 scans.
Fluorescence. All steady-state emission spectra were
dispersed in ddH2O at neutral pH by continuous high-speed measured with a JASCO spectrouorometer FP-6500. Emission
vortexing for more than 48 h until a uniform white emulsion spectra were recorded from 290 to 450 nm (excitation at
was formed. To fabricate the Fmoc-FF gel, one equivalent of 1 M 280 nm). The spectra were normalized with respect to the
Na2CO3 was mixed with the colloid. Clear gel was then formed maximum uorescent signal and the contributions from
in ꢁ10 to 20 min, and was aged at room temperature for one day solvent were subtracted.
without disturbance before characterization.
A chemical reaction triggered pH reduction (the K2S2O8
Results and discussion
Homogeneous hydrogels prepared using a colloid-to-hydrogel
method). Similar to the HCl method mentioned above, Fmoc-FF
(5 mg mLꢂ1) was rst dissolved in ddH2O at pH ꢁ9. Then 1
transition (the colloid method)
equivalent of K2S2O8 was added to the solution. Hydrogel could
ꢃ
form aer being heated to 50 C for several hours and cooled Colloid solution of amphiphilic peptides can be easily obtained
down to room temperature. The decomposition of K2S2O8 can by physically dispersing the peptide powder in water through
also proceed at room temperature to trigger the gelation despite vortexing. Due to the amphiphilic nature, Fmoc-FF can form
that the whole gelation process will take much longer (typically micelles in water to give white colloid solution (Fig. 2A). Such
ꢃ
4–5 days). Raising the temperature to 90 C and then cooling colloid solution is quite stable and can be stored for more than
down allows the hydrogel to be prepared in less than 5 min. In a week without precipitation. Zeta potential of the micelles
order to allow complete decomposition of the added K2S2O8, the formed by Fmoc-FF is ꢁꢂ35.5 ꢄ 0.5 mV at neutral pH, indi-
hydrogel was aged at room temperature for at least 4 days before cating that they are highly negatively charged. The AFM image
characterization. It seems that the history of the heating process shows that these micelles are in a short-rod shape, distinct from
does not affect the properties of the resulting hydrogel.
the typical self-assembled brillar structures of Fmoc-FF in
water (Fig. 2B). The diameter of the rod-like micelles is ꢁ15 nm
and the length of the micelles is ꢁ180 nm. The size of the
micelles is also conrmed using DLS, which shows a narrow
Hydrogel characterization
Atomic force microscopy (AFM). The microscopic morphol- distribution at ꢁ20 nm (Fig. 2C). Moreover, the secondary
ogies of the hydrogels were imaged at room temperature by a structure of Fmoc-FF in such colloid solution is mainly amor-
NanoWizard II AFM (JPK, Germany) operating in intermittent phous, distinct from that in the brils (Fig. S2†). Starting from
contact mode (conditions: scan rate, 1 Hz; number of pixels, such colloid solution, uniform hydrogels can be easily obtained
512 ꢀ 512). Silica cantilevers (AC160, Olympus, Japan) with by adding one equivalent of 1 M Na2CO3 to the solution
typical tip radii of ꢁ10 nm and resonance frequencies of ꢁ200 (Fig. 2D). Because the decomposition of such micelles is slower
to 500 kHz were used for imaging. The samples were prepared than the mixing of solutions, the solution can remain homo-
as follows: rst, ꢁ5 mL of the hydrogel was placed on the surface geneous aer Na2CO3 solution is added. However, only the
of a freshly peeled mica surface and allowed to adsorb for 5– Fmoc-FF in the surface layer of the colloid nanoparticles can be
10 min. Subsequently, the mica plate with the hydrogel sample gradually deprotonated (Na2CO3 is a weak base) and released
was quickly frozen in liquid nitrogen and dried in vacuum. from the colloids to self-assemble into hydrogel. This process
Then, the surface of the mica plate was ushed with nitrogen occurred slowly and close to equilibrium, due to relatively
gas to remove gel samples loosely attached.
strong hydrophobic interactions between the deprotonated
Dynamic light scattering (DLS) and zeta potential. The DLS Fmoc-FF and the colloid surface. If the added base was suffi-
experiments were carried out for the colloid solution of cient, the colloids would eventually completely decompose and
4674 | Soft Matter, 2013, 9, 4672–4680
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
concentrations of dissolved small-molecule gelators. Due to
these reasons, the self-assembly of the hydrogel in this
approach can be more homogeneous and reproducible. It is
worth mentioning that stable Fmoc-FF hydrogels can be
prepared at a pH greater than 9 using this approach, which was
thought to be impossible by typical pH or solvent switch
methods (Fig. S3†).
Homogeneous hydrogels prepared by gradually decreasing pH
through the decomposition of potassium persulfate (the
K2S2O8 method)
K2S2O8 can mildly decompose in water to give oxygen gas and
proton, leading to a decrease in pH (Fig. 3A). The decomposi-
tion of K2S2O8 has previously been used to lower the pH of
sodium alginate solution to trigger the formation of nano-
assemblies.65 The decomposition of K2S2O8 is slow and can be
controlled by temperature. Compared with the hydrolysis of
glucono-d-lactone to lower pH introduced by Adams et al., the
amount of proton released by the decomposition of K2S2O8 can
be simply calculated from the stoichiometry of the chemical
reaction.43 This is because this reaction does not stop in equi-
librium but complete with 100% conversion. Moreover, this
reaction can end up at any desired pH depending on the
amount of the added K2S2O8, not limited by the pKa of gluconic
acid (ꢁpH 3.6). However, because the C-terminal carboxyl group
of Fmoc-FF is a weak acid, the decrease of pH in Fmoc-FF
solution becomes slower as compared with that in water
(Fig. 3A). We monitored theꢃ gelation of Fmoc-FF by the
decomposition of K2S2O8 at 50 C using rheology. As shown in
Fig. 3B, when 1 equivalent of (ꢁ0.01 M) K2S2O8 was added to the
Fmoc-FF solution (5 mg Lꢂ1, ꢁpH 9), the storage modulus
increased slowly along the decomposition of K2S2O8 and
reached a plateau of 10 Pa at ꢁ600 min, at which the reaction is
almost complete. The increase of the storage modulus was in
concert with the decrease of pH, indicating that the formation
of gel structure is indeed due to the protonation of Fmoc-FF
Fig. 2 Preparation of homogeneous hydrogels using the colloid-to-hydrogel
transition (the colloid method). (A) Photograph of the nano-dispersed Fmoc-FF
colloid solution (5 mg mLꢂ1). (B) AFM image of Fmoc-FF colloid. (C) Dynamic light
scattering of Fmoc-FF colloid solution (5 mg mLꢂ1). (D) Photograph of Fmoc-FF molecules. Fig. 3C shows a picture of Fmoc-FF hydrogel
hydrogel after Na2CO3 added to the colloid solution. (E) Schematic illustration of
prepared using the decomposition of K2S2O8, which is also
transparent and homogeneous. A little bit of oxygen gas bubbles
the mechanism for the colloid-to-hydrogel transition process.
may be trapped in the hydrogel due to the decomposition of
K2S2O8. It is worth mentioning that this method can also be
yield a transparent hydrogel within ꢁ20 min. The colloids in used at room temperature despite that the whole gelation
this approach actually serve as a “trap” that controls the amount process will take much longer (typically 4–5 days). Raising the
ꢃ
of Fmoc-FF in the solution, because the decomposition of the temperature to 90 C and cooling down allows the hydrogel to
as-formed colloids is an obligatory step for the following self- be prepared in 5 min. The hydrogels prepared at different
assembly process (Fig. 2E). Moreover, in this colloid-to-hydrogel temperatures show similar brillar structures (Fig. S4†). It
transition process, small-molecule gelators are highly packed as seems that the lower the temperature is, the thicker the brils
colloids, which greatly reduces the viscosity of the system and are. The gelation process is largely “programmable” by simply
makes the inter-conversion between different states along the controlling the decomposition kinetics of K2S2O8. Moreover, we
self-assembly pathway much easier. As proposed by Kramer's proposed that any chemical reactions that produce proton
theory, the reaction kinetics is related to the viscosity of the could be used in a similar way to trigger the formation of
system as follows: kꢂ1 f h ꢀ exp(DEa/RT), where k is the rate hydrogel, wherein the reactants and the products should not
constant, h is the viscosity of the system, DEa is the activation interfere with the self-assembly of the peptides. The idea of this
energy, R is the gas constant, and T is the temperature. In approach is to make the self-assembly process take place slowly
contrast, in typical sol–gel transitions, the gelation process and close to equilibrium so that it is not kinetically trapped
usually takes place at high viscosity due to the presence of high into the local minima in the energy landscape. Because both
This journal is ª The Royal Society of Chemistry 2013
Soft Matter, 2013, 9, 4672–4680 | 4675

View Article Online
Soft Matter
Paper
Fig. 3 Preparation of homogeneous hydrogels using the decomposition of K2S2O8 (the K2S2O8 method). (A) Decomposition of K2S2O8 (ꢁ0.01 M) results in the
decrease of pH in water (red circle) and in the Fmoc-FF solution (5 mg mLꢂ1, blue square) at 50 ꢃC. (B) Monitoring the gelation process using the rheology. The storage
modulus (blue square), G0, increases gradually and becomes larger than loss modulus (red circle), G00, at ꢁ10 min, indicating the formation of hydrogel. Both G0 and G00
reach plateaus at ꢁ600 min, which is in concert with the complete decomposition of K2S2O8 shown in (A). (C) Photograph of Fmoc-FF hydrogel prepared using the
K2S2O8 method.
Fmoc-FF and K2S2O8 are dissolved in the solution, the protons in solution (the spectra were taken at a low concentration of
produced by the decomposition of K2S2O8 can be immediately 0.1 mg mLꢂ1, Fig. S6†). With a close comparison of the uo-
used to protonate Fmoc-FF and the system maintains homo- rescence spectra of the hydrogels obtained via the four different
geneous in the entire gelation process.
methods, we found that the hydrogel formed by the K2S2O8
method is of the longest emission wavelength of 329 nm, sug-
gesting that the p–p stacking structures are better formed in
this gel. It is worth noting that the hydrogel formed by the
DMSO method shows the lowest uorescence at ꢁ325 nm,
Structural properties of Fmoc-FF hydrogels prepared using
different methods
Having established two new methods to prepare homoge- indicating poor p–p stacking in the hydrogel. Moreover, the
neous hydrogels, we characterized the structural properties of hydrogels prepared using the HCl and the K2S2O8 methods
Fmoc-FF hydrogels prepared using these two methods as well show weak broad uorescence at ꢁ460 nm, which may be due to
as those prepared using the HCl method and the DMSO the stacking of the formed brils (Fig. 4C).
method (Fig. S5†). AFM images show that Fmoc-FF forms
However, this peak is weak for the hydrogels prepared using
nanoscale brils in the hydrogels prepared using all four the colloid and the DMSO methods. The diversity of uores-
methods (Fig. 4A–D). However, the sizes of the brils formed cence spectra indicates that the nano-brils formed in the four
by all four methods are different from each other, indicating different approaches may have distinct structures.
that the preparation procedures play an important role in the
We also used CD to investigate the structures of Fmoc-FF
self-assembly of Fmoc-FF. The brils formed by the colloid brils in the hydrogels prepared using these four different
method are much thinner (ꢁ20 nm) than those formed by the methods (Fig. 4F). The CD signal arises from the chirality in the
K2S2O8 method (ꢁ100 nm). We hypothesize that this might be self-assembled brils instead of that of the peptides. Therefore,
because the colloid Fmoc-FF could serve as the seeds for the different CD spectra provide a strong evidence of different bril
growth of brils. Since there are many such kinds of nuclei structures. A similar CD spectrum shape but different intensi-
presented in the system, the resulting brils are thin. ties may suggest that the same bril structures are formed to a
However, for the K2S2O8 method, no such seeds are present in different extent. To our surprise, the CD spectra of these
the solution. The nucleation process is slow and the growth of hydrogels are different from each other, indicating Fmoc-FF
the brils occurs close to equilibrium. Nonetheless, the may adopt different chain arrangements. The circular
diameters of the brils formed by the two new methods are dichroism (CD) signals at ꢁ230 nm result from n–p* transi-
more uniform than those prepared by the HCl and the DMSO tions, and that near 280 nm corresponds to p–p* transitions in
methods, suggesting that these hydrogels are homogeneous at the uorenyl groups. Ulijn and coworkers have suggested a p–p
the microscopic level.
interlocked b-sheet structure for Fmoc-FF brils formed in the
We then used uorescence to explore the p–p stacking in the hydrogel prepared using dropwise pH switch method.42 The CD
four hydrogels (Fig. 4E). Because p–p stacking can lower the spectrum they obtained is similar to that we measured using the
energy gap between LUMO and HOMO orbital, leading to HCl method. Similar CD spectra were also reported by Huang
the shi of uorescence emission to longer wavelengths, the et al. Slight variations in these reported CD spectra may arise
uorescence spectra can be used to probe the state of p–p from different Fmoc-FF concentrations used for CD measure-
stacking in the gel systems. The uorescence of Fmoc-FF in the ment (Fig. S7†). However, the CD spectra of the hydrogels
four different hydrogels is indeed shied to longer wavelengths prepared using the other three methods are distinct, indicating
in comparison with the uorescence of unstructured Fmoc-FF that the brillar structures of Fmoc-FF can have different
4676 | Soft Matter, 2013, 9, 4672–4680
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
Mechanical properties of Fmoc-FF hydrogels prepared using
different methods
Since the Fmoc-FF hydrogels which were prepared using the two
new methods reported herein exhibit different microstructures
from those prepared using traditional methods, we then tested
whether such structural changes affect the macroscopic
mechanical properties. We measured the viscoelastic mechan-
ical properties of Fmoc-FF hydrogels prepared using these four
methods. For all four hydrogels, they show higher storage
moduli (G0) than viscous moduli (G00) in a broad frequency
region, suggesting that they are indeed in a gel form (Fig. 5).
However, the storage moduli vary dramatically for the four
different hydrogels, ranging from 10 to 400 Pa. All four hydro-
gels can be destroyed at high sweep frequencies (Fig. 5A and B),
indicating that they are so gels. It is worth noting that such
rising of the moduli is not because the hydrogels are not
equilibrated. Similar mechanical features were observed in
hydrogel samples that have been equilibrated for more than a
week. However, it seems that the hydrogels prepared using the
two novel methods are soer than that prepared using the HCl
and the DMSO methods. The molecular mechanism underlying
the distinct mechanical properties is unknown. It was proposed
that the undissolved peptides trapped in the hydrogel might
make the resulting gel stiffer.53 Recent systematic studies indi-
cated that pH and solvent conditions both signicantly affect
the mechanical properties of Fmoc-FF hydrogel.54 In all four
different kinds of hydrogels reported in this paper, the pH
values are maintained higher than 7 by using the desired
amount of Na2CO3 in the hydrogels. Nonetheless, by simply
controlling the gelation kinetics, it is possible to achieve
hydrogels of distinct mechanical properties. Moreover, the
rigidity of the hydrogels is also not correlated to the size of the
Fmoc-FF brils. The hydrogels prepared using the colloid and
the K2S2O8 methods show similar mechanical properties yet
have distinct brillar structures. It seems that the mechanical
properties of the hydrogel are determined by many factors, such
as the size, rigidity, concentration, and cross-linking density of
the brils. These microscopic factors work together and lead to
similar mechanical properties of the two hydrogels of distinct
structures coincidently. Another feature worth mentioning is
that the mechanical properties of the hydrogels prepared using
the colloid and the K2S2O8 methods vary in less than 15% and
are extremely reproducible in different batches. However, the
mechanical properties of the hydrogels prepared by the HCl
method and the DMSO method can vary from batch to batch
(the change can be as large as one order of magnitude.). This
may be because in these two methods hydrogels are prepared in
processes far from equilibrium and are not homogeneous. To
further illustrate the reproducibility of the mechanical proper-
ties prepared using these two novel methods, we monitored the
destroy–recovery of all four different hydrogels by rheology. As
shown in Fig. 6, we rst destroyed the hydrogels using a high
vibration amplitude of 100%. Upon mechanical agitation, the G0
becomes lower than G00, indicating that the gels are fully
destroyed. Then the vibration amplitude was lowered to 1% to
allow the hydrogels to recover. For all four different hydrogels,
Fig.
4
Structural characterizations of Fmoc-FF hydrogels prepared using
different methods. AFM images of the hydrogels prepared by the colloid method
(A), K2S2O8 method (B), HCl method (C), and DMSO method (D). (E) Fluorescence
emission spectra of the hydrogels prepared using the four different methods
(excited at 280 nm). (F) Circular dichroism spectra of those four hydrogels.
chirality. Such structural difference originates from different
growth kinetics and it is possible to alter the bril structure by
controlling the formation conditions. This nding is consistent
with that reported by Mu et al.48 They found that Fmoc dipep-
tides can form alternative structures using both experimental
and computational studies. In parallel, Houton et al. found that
the molecular packing of naphthalene dipeptides in the crystal
structure differs from that in the gel/ber phase.66 However,
kinetically controlling the self-assembly to predened struc-
tures is challenging.
This journal is ª The Royal Society of Chemistry 2013
Soft Matter, 2013, 9, 4672–4680 | 4677

View Article Online
Soft Matter
Paper
Fig. 6 Destroy–recovery cycles of Fmoc-FF hydrogel (5 mg mLꢂ1) monitored
using rheology. All four hydrogels prepared by the colloid method (A), K2S2O8
method (B), HCl method (C), and DMSO method (D) have been subjected to four
destroy–recovery cycles in each test.
Fig. 5 Viscoelastic mechanical properties of Fmoc-FF hydrogel (5 mg mLꢂ1
)
prepared by the colloid method (A), K2S2O8 method (B), HCl method (C), and
DMSO method (D). Error bars indicate the standard deviation from different
measurements and the lower error bars are not displayed for clarity.
prepared by the two new methods rise gradually and reach a
plateau in ꢁ1000 s. However, in the same recovery time period,
the mechanical moduli of the hydrogels prepared using the HCl
and DMSO methods are not fully recovered. Moreover, we
the G0 increases immediately aer the mechanical agitation is
removed and becomes larger than G00, indicating that the
hydrogels are restored. Obviously, the recovery rates of the
hydrogels prepared using the colloid and the K2S2O8 methods noticed that the maximum G0 for each destroy–recovery cycle of
are much faster than those for the hydrogels prepared using the the hydrogels prepared using the colloid and the K2S2O8
HCl and DMSO methods. Both G0 and G00 of the hydrogels methods is reproducible, while that for the hydrogels prepared
4678 | Soft Matter, 2013, 9, 4672–4680
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
by the HCl and DMSO methods varies in a large range. It seems
that these two hydrogels somehow show a bit mechanical
fatigue upon agitation. As shown in Fig. 6, the recovered storage
moduli are typically lower than the previous cycles for the
hydrogels prepared by the K2S2O8 and DMSO methods. There-
8 J. P. Jung, J. V. Moyano and J. H. Collier, Integr. Biol., 2011, 3,
185–196.
9 J. S. Rudra, T. Sun, K. C. Bird, M. D. Daniels,
J. Z. Gasiorowski, A. S. Chong and J. H. Collier, ACS Nano,
2012, 6, 1557–1564.
fore, the two new methods presented in this paper may be 10 M. C. Branco, D. M. Sigano and J. P. Schneider, Curr. Opin.
advantageous for the control of mechanical properties of the
self-assembled hydrogels.
Chem. Biol., 2011, 15, 427–434.
11 X. Liu, W. Zhuang, B. Li, L. Wu, S. Wang, Y. Li and Y. Yang,
Chem. Commun., 2011, 47, 7215–7217.
12 H. Wang, J. Wei, C. Yang, H. Zhao, D. Li, Z. Yin and Z. Yang,
Biomaterials, 2012, 33, 5848–5853.
Conclusions
13 H. Wang and Z. Yang, Nanoscale, 2012, 4, 5259–5267.
14 H. Wang, Z. Yang and D. J. Adams, Mater. Today, 2012, 15,
500–507.
15 C. Yang, D. Li, Z. Liu, G. Hong, J. Zhang, D. Kong and
Z. Yang, J. Phys. Chem. B, 2012, 116, 633–638.
16 D. Li, J. Liu, L. Chu, J. Liu and Z. Yang, Chem. Commun.,
2012, 48, 6175–6177.
In this article, we presented two novel approaches to engineer
homogeneous Fmoc-FF hydrogels. One is based on the colloid-
to-hydrogel transition and the other is based on the decompo-
sition of K2S2O8. These two methods greatly increase the
homogeneity and reproducibility of the self-assembled hydro-
gels. Although the mechanism that underlies these two
methods is different, they share the same fundamental prin-
ciple: kinetically controlling the self-assembly process. Our
results demonstrate that the self-assembly kinetics can play an
important role in the properties of the resultant assembled
hydrogels, which has somehow been previously overlooked.
Moreover, it is also possible to rationally guide the self-assembly
process to a designed pathway. We anticipate that these two
approaches can be extended to other self-assembly systems and
used as versatile tools for the construction of homogeneous
hydrogels. Since many applications of the self-assembled
hydrogels rely on their physical properties, these two
approaches can also potentially expand the applications of the
same small-molecule gelators by kinetically controlling the
assembly processes towards distinct physical properties.
17 L. Mao, H. Wang, M. Tan, L. Ou, D. Kong and Z. Yang, Chem.
Commun., 2012, 48, 395–397.
18 H. Wang and Z. Yang, So Matter, 2012, 8, 2344–2347.
19 J. Gao, H. Wang, L. Wang, J. Wang, D. Kong and Z. Yang, J.
Am. Chem. Soc., 2009, 131, 11286–11287.
20 Z. Yang, L. Wang, J. Wang, P. Gao and B. Xu, J. Mater. Chem.,
2010, 20, 2128–2132.
21 B. Li, L. Tang, L. Qiang and K. Chen, So Matter, 2011, 7,
963–969.
22 J. D. Hartgerink, E. Beniash and S. I. Stupp, Science, 2001,
294, 1684–1688.
23 W. Zheng, J. Gao, L. Song, C. Chen, D. Guan, Z. Wang, Z. Li,
D. Kong and Z. Yang, J. Am. Chem. Soc., 2013, 135, 266–
271.
24 J. D. Hartgerink, E. Beniash and S. I. Stupp, Proc. Natl. Acad.
Sci. U. S. A., 2002, 99, 5133–5138.
25 C. A. E. Hauser and S. Zhang, Chem. Soc. Rev., 2010, 39, 2780–
2790.
26 Z. Luo and S. Zhang, Chem. Soc. Rev., 2012, 41, 4736–4754.
27 D. J. Adams, Macromol. Biosci., 2011, 11, 160–173.
28 T. Aida, E. W. Meijer and S. I. Stupp, Science, 2012, 335, 813–
817.
29 R. Huang, W. Qi, L. Feng, R. Su and Z. He, So Matter, 2011,
7, 6222–6230.
Acknowledgements
This work was supported by the National Basic Research Pro-
gramme of China (nos 2012CB921801), the National Natural
Science Foundation of China under grant nos 91127026,
31170813 and 11074115, the program for New Century Excellent
Talents in University, and the Priority Academic Program
Development of Jiangsu Higher Education Institutions.
30 D. Silva, A. Natalello, B. Sanii, R. Vasita, G. Saracino,
R. N. Zuckermann, S. M. Doglia and F. Gelain, Nanoscale,
2013, 5, 704–718.
References
1 L. A. Estroff and A. D. Hamilton, Chem. Rev., 2004, 104, 1201–
1218.
31 A. S. Veiga, C. Sinthuvanich, D. Gaspar, H. G. Franquelim,
M. A. Castanho and J. P. Schneider, Biomaterials, 2012, 33,
8907–8916.
2 N. M. Sangeetha and U. Maitra, Chem. Soc. Rev., 2005, 34,
821–836.
3 Y. Gao, F. Zhao, Q. Wang, Y. Zhang and B. Xu, Chem. Soc. 32 S. Marchesan, L. Waddington, C. D. Easton, D. A. Winkler,
Rev., 2010, 39, 3425–3433.
4 V. Jayawarna, S. M. Richardson, A. R. Hirst, N. W. Hodson,
L. Goodall, J. Forsythe and P. G. Hartley, Nanoscale, 2012,
4, 6752–6760.
A. Saiani, J. E. Gough and R. V. Ulijn, Acta Biomater., 2009, 33 R. P. Nagarkar, R. A. Hule, D. J. Pochan and J. P. Schneider, J.
5, 934–943.
Am. Chem. Soc., 2008, 130, 4466–4474.
5 F. Zhao, M. L. Ma and B. Xu, Chem. Soc. Rev., 2009, 38, 883–891. 34 Z. Qiu, H. Yu, J. Li, Y. Wang and Y. Zhang, Chem. Commun.,
6 J. B. Matson and S. I. Stupp, Chem. Commun., 2012, 48, 26–33. 2009, 3342–3344.
7 F. Gelain, A. Horii and S. Zhang, Macromol. Biosci., 2007, 7, 35 R. J. Swanekamp, J. T. DiMaio, C. J. Bowerman and
544–551.
B. L. Nilsson, J. Am. Chem. Soc., 2012, 134, 5556–5559.
This journal is ª The Royal Society of Chemistry 2013
Soft Matter, 2013, 9, 4672–4680 | 4679

View Article Online
Soft Matter
Paper
36 J. N. Onuchic, Z. Luthey-Schulten and P. G. Wolynes, Annu. 52 D. M. Ryan, T. M. Doran, S. B. Anderson and B. L. Nilsson,
Rev. Phys. Chem., 1997, 48, 545–600. Langmuir, 2011, 27, 4029–4039.
37 Y. Zhang, H. Gu, Z. Yang and B. Xu, J. Am. Chem. Soc., 2003, 53 C. Tang, A. M. Smith, R. F. Collins, R. V. Ulijn and A. Saiani,
125, 13680–13681. Langmuir, 2009, 25, 9447–9453.
38 Z. M. Yang, H. W. Gu, D. G. Fu, P. Gao, J. K. Lam and B. Xu, 54 J. Raeburn, G. Pont, L. Chen, Y. Cesbron, R. Levy and
Adv. Mater., 2004, 16, 1440–1444. D. J. Adams, So Matter, 2012, 8, 1168–1174.
39 V. Jayawarna, M. Ali, T. A. Jowitt, A. E. Miller, A. Saiani, 55 D. J. Adams, L. M. Mullen, M. Berta, L. Chen and W. J. Frith,
J. E. Gough and R. V. Ulijn, Adv. Mater., 2006, 18, 611–
614.
40 S. Toledano, R. J. Williams, V. Jayawarna and R. V. Ulijn, J.
Am. Chem. Soc., 2006, 128, 1070–1071.
41 A. Mahler, M. Reches, M. Rechter, S. Cohen and E. Gazit,
Adv. Mater., 2006, 18, 1365–1370.
42 A. M. Smith, R. J. Williams, C. Tang, P. Coppo, R. F. Collins,
So Matter, 2010, 6, 1971–1980.
56 L. Chen, J. Raeburn, S. Sutton, D. G. Spiller, J. Williams,
J. S. Sharp, P. C. Griffiths, R. K. Heenan, S. M. King,
A. Paul, S. Furzeland, D. Atkins and D. J. Adams, So
Matter, 2011, 7, 9721–9727.
57 J. Ryu, S.-W. Kim, K. Kang and C. B. Park, Adv. Mater., 2010,
22, 5537–5541.
M. L. Turner, A. Saiani and R. V. Ulijn, Adv. Mater., 2008, 20, 58 T. Liebmann, S. Rydholm, V. Akpe and H. Brismar, BMC
37–41. Biotechnol., 2007, 7, 88.
43 D. J. Adams, M. F. Butler, W. J. Frith, M. Kirkland, 59 R. Orbach, L. Adler-Abramovich, S. Zigerson, I. Mironi-
L. Mullen and P. Sanderson, So Matter, 2009, 5, 1856–
1862.
Harpaz, D. Seliktar and E. Gazit, Biomacromolecules, 2009,
10, 2646–2651.
44 J. Raeburn, T. O. McDonald and D. J. Adams, Chem. 60 N. Amdursky, R. Orbach, E. Gazit and D. Huppert, J. Phys.
Commun., 2012, 48, 9355–9357. Chem. C, 2009, 113, 19500–19505.
45 M. Ma, Y. Kuang, Y. Gao, Y. Zhang, P. Gao and B. Xu, J. Am. 61 R. Orbach, I. Mironi-Harpaz, L. Adler-Abramovich,
Chem. Soc., 2010, 132, 2719–2728.
46 N. Amdursky, E. Gazit and G. Rosenman, Adv. Mater., 2010,
22, 2311–2315.
47 C. Tang, R. V. Ulijn and A. Saiani, Langmuir, 2011, 27, 14438–
14449.
48 X. Mu, K. M. Eckes, M. M. Nguyen, L. J. Suggs and P. Ren,
Biomacromolecules, 2012, 13, 3562–3571.
49 D. M. Ryan, S. B. Anderson and B. L. Nilsson, So Matter,
2010, 6, 3220–3231.
50 D. M. Ryan, S. B. Anderson, F. T. Senguen, R. E. Youngman
and B. L. Nilsson, So Matter, 2010, 6, 475–479.
51 D. M. Ryan, T. M. Doran and B. L. Nilsson, Langmuir, 2011,
27, 11145–11156.
E. Mossou, E. P. Mitchell, V. T. Forsyth, E. Gazit and
D. Seliktar, Langmuir, 2012, 28, 2015–2022.
62 Z.-X. Tie, M. Qin, D.-W. Zou, Y. Cao and W. Wang, Chin. Phys.
Lett., 2011, 28, 028702.
63 D. Zou, Y. Cao, M. Qin, W. Dai and W. Wang, Chem.
Commun., 2011, 47, 7413–7415.
64 D.-W. Zou, Z.-X. Tie, M. Qin, C.-M. Lu and W. Wang, Chin.
Phys. Lett., 2009, 26, 088103.
65 Y. Cao, X. Shen, Y. Chen, J. Guo, Q. Chen and X. Jiang,
Biomacromolecules, 2005, 6, 2189–2196.
66 K. A. Houton, K. L. Morris, L. Chen, M. Schmidtmann,
J. T. Jones, L. C. Serpell, G. O. Lloyd and D. J. Adams,
Langmuir, 2012, 28, 9797–9806.
4680 | Soft Matter, 2013, 9, 4672–4680
This journal is ª The Royal Society of Chemistry 2013