DOI: 10.1002/chem.201100640

Versatile Low-Molecular-Weight Hydrogelators: AchievingMultiresponsiveness through a Modular Design

Lilia Milanesi,*[a] Christopher A. Hunter,[b] Nadejda Tzokova,[b]

Jonathan P. Waltho,[c] and Salvador Tomas*[a]

Introduction

Low-molecular-weight gelators (LMWGs) are small mole-cules that self-assemble into long fibers, resulting in the ge-lation of the solvent.[1–3] These fibers are held together bynoncovalent, weak interactions between the individualLMWG molecules.[1,2, 4–7] The gelating properties of LMWGscan be modulated by modifying the strength of these inter-actions, as demonstrated by the thermal reversibility of thegel-to-sol transition.[1–3, 7–11] The close relationship betweenthe structural integrity of the gel and the weak interactionsallows tuning of the macroscopic properties of LMWGsmore efficiently than polymeric (covalently assembled)gels.[1,12–15] LMWGs are therefore ideal candidates for thedevelopment of stimuli-responsive smart materials.[1,13,16,17]

Low-molecular-weight hydrogelators (LMWHs), that is,LMWGs that gelate water, are of special interest for the de-velopment of biocompatible materials.[5,8,13, 18,19] For example,LMWHs with applications in regenerative medicine, tissueengineering, drug delivery, biosensing, and catalysis havebeen reported.[12,13, 20–28] For many LMWHs the ability togelate water can be modulated as a response to one particu-lar external stimulus (other than a temperature change). Inthese LMWHs, responsiveness is commonly achieved by

Abstract: Multiresponsive low-molecu-lar-weight hydrogelators (LMWHs) areideal candidates for the developmentof smart, soft, nanotechnology materi-als. The synthesis is however very chal-lenging. On the one hand, de novodesign is hampered by our limited abil-ity to predict the assembly of smallmolecules in water. On the other hand,modification of pre-existing LMWHs islimited by the number of differentstimuli-sensitive chemical moieties thatcan be introduced into a small mole-cule without seriously disrupting theability to gelate water. Herein wereport the synthesis and characteriza-tion of multistimuli LMWHs, based ona modular design, composed of a hy-drophobic, disulfide, aromatic moiety, amaleimide linker, and a hydrophilic

section based on an amino acid, hereN-acetyl-l-cysteine (NAC). As mostLMWHs, these gelators experience re-versible gel-to-sol transition followingtemperature changes. Additionally, theNAC moiety allows reversible controlof the assembly of the gel by pHchanges. The reduction of the aromaticdisulfide triggers a gel-to-sol transitionthat, depending on the design of theparticular LMWH, can be reverted byreoxidation of the resulting thiol. Final-ly, the hydrolysis of the cyclic imidemoieties provides an additional triggerfor the gel-to-sol transition with a time-

scale that is appropriate for use indrug-delivery applications. The effi-cient response to the multiple externalstimuli, coupled to the modular designmakes these LMWHs an excellentstarting point for the development ofsmart nanomaterials with applicationsthat include controlled drug release.These hydrogelators, which were dis-covered by serendipity rather thandesign, suggest nonetheless a generalstrategy for the introduction of multi-ple stimuli-sensitive chemical moieties,to offset the introduction of hydrophil-ic moieties with additional hydrophobicones, in order to minimize the upset-ting of the critical hydrophobic–hydro-philic balance of the LMWH.

Keywords: disulfide · gels · male-imide · multiresponsiveness · self-assembly

[a] Dr. L. Milanesi, Dr. S. TomasInstitute of Structural and Molecular BiologyandDepartment of Biological SciencesSchool of Science, Birkbeck University of LondonLondon WC1E 7HX (UK)Fax: (+44) 20-7631-6246E-mail : l.milanesi@mail.cryst.bbk.ac.uk

s.tomas@bbk.ac.uk

[b] Prof. C. A. Hunter, Dr. N. TzokovaCentre for Chemical BiologyKrebs Institute for Biomolecular ScienceDepartment of ChemistryThe University of Sheffield, Sheffield S3 7HF (UK)

[c] Prof. J. P. WalthoDepartment of Molecular Biology and BiotechnologyThe University of Sheffield, Sheffield S10 2TN (UK)andFaculty of Life SciencesandManchester Interdisciplinary BiocentreThe University of Manchester, Manchester M1 7DN (UK)

Supporting information for this article is available on the WWWunder http://dx.doi.org/10.1002/chem.201100640.

Chem. Eur. J. 2011, 17, 9753 – 9761 � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 9753

FULL PAPER

adding chemical moieties that are sensitive to the externalstimulus of interest (e.g., to pH, light, ionic strength, enzy-matic reactions, or the addition of ligands) to a pre-existinggelating scaffold.[10, 13,14,29–33] The changes in these chemicalmoieties, brought about by the external stimulus, lead eitherto a gel-to-sol or a sol-to-gel transition that in some casesmay be reversibly controlled. Whereas many LMWHs re-sponsive to a single stimulus have been reported to date, ex-amples of irreversible or reversible multiresponsiveLMWHs are still rare, in spite of the superiority of thesematerials relative to single-responsive gelators.[13,22,28, 29,31, 32]

However, achieving multiresponsiveness is a challengingtask. LMWHs are amphiphilic molecules for which the abili-ty to assemble into gel-forming fibers depends to a largeextent on the hydrophobic–hydrophilic balance, finely tunedby the presence of intermolecular interactions, such as Hbonds.[1–3,11] Changes in this balance, brought about by theaddition of several stimuli-responsive moieties, may easilylead to loss of the ability of the molecule to assemble into ahydrogel.[11] The impact of these moieties may be reducedby offsetting the effect of an additional moiety, which is hy-drophilic in nature, with the presence of another one, whichis hydrophobic, each of them located in appropriate parts ofthe molecule. Herein we report multistimuli-responsiveLMWHs based on this approach (Scheme 1); these were dis-covered by serendipity during the synthesis of optical trig-gers for the study of protein folding.[34,35] A hydrophobic di-sulfide bond, located in the hydrophobic part of the mole-cule provides reversible responsiveness to redox stimu-li.[11,36, 37] Hydrophilic carboxylic-acid moieties, located in the

hydrophilic part of the molecule, provide reversible respon-siveness to changes in pH.[10,29,31, 32] Additionally, the pres-ence of hydrolysable cyclic imides provides a third mecha-nism for gel disassembly, albeit a nonreversible one.

Results and Discussion

Synthesis and characterization of the hydrogelators : Com-pounds 2 and 3 were prepared in a one-pot reaction fromthe bismaleimide derivative 1 by using a procedure previ-ously reported for the synthesis of protein-crosslinking re-agents (Scheme 1).[34, 35] Addition of NAC over 1 yielded amixture of 2 and 3 in a 5:3 ratio; this mixture was separatedby preparative reversed-phase HPLC (RP-HPLC). Thewater-gelating ability of 2 and 3 was readily observed whenevaporating the corresponding aqueous fractions containingthe gelators. Both 2 and 3 were obtained as a mixture of dia-stereoisomers due to the stereochemistry typical of this typeof Michael addition reactions. The mixture of diastereoiso-mers was approximately 1:1 for both compounds as shownby NMR and no attempts were made to separate each dia-stereoisomer for the purpose of this work (see the Support-ing Information, Figure 1S, 2S, and 3S). In terms of gellingefficiencies, pure diastereoisomers or pure enantiomers aretypically superior relative to the corresponding diastereoiso-meric mixture or racemate.[38,39] However, efficient gelationproperties can sometimes be attained from a diastereotopicmixture of compounds.[40]

The imide groups of 2 and 3 are susceptible to base-cata-lyzed hydrolysis in aqueous solution, a reaction observed inmany maleimide bioconjugates and maleimide-thiols ad-ducts in general.[41,42] The hydrolysis of each imide groupyields two additional carboxylic-acid groups that may affectthe solubility of 2 and 3 and the ability to gelate water(Scheme 1S in the Supporting Information). We thereforestudied the kinetic stability of 2 and 3 against hydrolysis bymonitoring changes in the UV/Vis spectra of solutions (i.e. ,below the critical gelation concentration (CGC)) of bothcompounds. As expected the rate constants are pH depen-dent (Table 1 and Figure 4S in the Supporting Information).The hydrolysis of the succinimide moiety is relatively slow

Scheme 1. NAC=N-acetyl-l-cysteine.

Table 1. Observed pseudo-first-order rate constants for hydrolysis of sol-utions of 2 and 3 and hydrogel 3.

Compound kobs

[�10�6 s�1]t1=2

[h] k’obs

[�10�6 s�1]t’1=2

[h] pH

2 (sol)[a] 7.0�0.19 27.5 – – 6.62 (sol)[a] 14�0.16 13.7 – – 7.23 (sol)[a] 7.2�0.66 26.7 67�0.5 2.9 6.63 (sol)[a] 15�0.48 12.8 130�13 1.5 7.23 (gel)[b] – – 4.3�0.27 45 6.63 (gel)[b] – – 6.3�0.27 31 7.2

[a] For both compounds in solution, kobs and t1=2 are the rate constant andhalf life for hydrolysis of the succinimide moiety and k’obs and t’1=2

are therate constant and half life for the hydrolysis of the maleimide moiety.[b] For hydrogel 3, k’obs and t’1=2

are the rate constant and half life for thehydrolysis of both imide moieties.

www.chemeurj.org � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eur. J. 2011, 17, 9753 – 97619754

and similar for both 2 and 3 (t1/2 = 12 h for 2 and 13 h for 3at pH 7.2). The hydrolysis of the maleimide moiety of 3, onthe other hand, is nearly one order-of-magnitude faster rela-tive to hydrolysis of the succinimide group; this is consistentwith the literature (t1/2 of the imide moiety of 3 is 1.5 h,Table 1 and Figure 4S in the Supporting Information).[41, 42]

In the gel state the hydrolysis of the imide groups is mark-edly slower, which is important for the stability and poten-tial application of the hydrogelators as explained in the fol-lowing sections.

Water-gelating properties of 2 and 3 : The gelation ability of2 and 3 was determined by using the stable-to-inversionmethod (Table 2 and Figure 1).[11] Neither 2 or 3 are readilysoluble in pure water and alkaline solutions of sodium phos-

phate were used to dissolve the compounds. Acidification ofthese solutions induces the formation of a gel suspensionthat becomes transparent by heating at 50 8C and yields ahydrogel upon cooling at 25 8C (Figure 1).

The CGC is pH dependent for both compounds. Changesin CGCs are observed in the pH range 2.0–7.4, and cantherefore be attributed to the presence of the ionizable car-boxylic groups (Table 2). Whereas the CGC values decrease

at acidic pH values for both 2 and 3, the CGCs of 3 aremuch lower than the CGCs of 2 at any of the pH valuestested. The pKa for the carboxylic groups of structurally re-lated compounds, such as 2 and 3, is assumed to be similarin the solution state and around 3.0.[43] Whereas the extentof deprotonation in the gel state may not correlate exactlywith the pKa in solution,[44–46] at high pH values the differ-ence in CGCs can be explained in terms of higher solubilityof the monomers (total charge �2 for 2, and �1 for 3 in so-lution). However, the differences in CGCs are maintainedalso for low pH values for which it is reasonable to assumefull protonation for both gelators. This result shows that thepresence of the additional NAC moiety makes 2 more solu-ble than 3, regardless of the protonation state of the carbox-ylic groups.

TEM revealed the formation of fibrillar aggregates forboth hydrogelators (Figure 2). The fibers are entangled andundergo splitting and overlapping in some areas of the mi-crographs. This arrangement of the fibers is consistent with

a possible mechanism for the formation of cavities and sub-sequent entrapment of the water molecules that leads to ge-lation and that is a typical feature of the assembly ofLMWGs.[1,2,4–7] Whereas the morphologies of the aggregateswere similar, the distribution of fiber widths appears to bedifferent for the two gelators. In aggregates of 2, 67 % of thetotal fibers show 8–11 nm widths and are formed by lateralassembly of thin fibers of 4–6 nm width (Figure 2 c). In the

Table 2. CGC values for gelators 2 and 3 (in mm and %wt in brackets)at different pHs.

Gelator[a] pH 7.4 pH 6.0 pH 4.5 pH 2.0

2 100 (5.5) 80 (4.8) 30 (1.8) 12.5 (0.92)3 15 (0.86) 10 (0.57) 2.5 (0.14) 1.5 (0.10)

[a] The hydrogels were obtained at the pH listed in the table by acidifica-tion of solutions of 2 and 3 in phosphate buffer (80 mm, pH 6.2–8.4) andwere observed approximately 10 min (hydrogel 2) and 20 min (hydrogel3) after formation of the solution phase.

Figure 1. Pictures of self-supporting hydrogel 2 (left) and hydrogel 3(right) in an inverted vial. Both hydrogelators where dissolved up to theCGCs in phosphate buffer (80 mm, pH 8.4) to give a final pH of 4.3 (forhydrogelator 2) or pH 6.0 (for hydrogelator 3). The magnetic bar of2 mm entrapped within the hydrogels qualitatively shows the rigidity ofthe gels.

Figure 2. TEM images of a) hydrogel 2, b) hydrogel 3, and c) histogramsof the fiber-width frequency distribution of both hydrogels (grey bars forgel 2 and black bars for gel 3). In (a) and (b) the widths of some of thefibers are marked by black bars, and black arrows point to regions ofsplitting of the fibers. In (a), the black box contours an area of fibers thatare laterally assembled; this area is further magnified at the bottom ofthe micrograph. The scale bar is 100 nm.

Chem. Eur. J. 2011, 17, 9753 – 9761 � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemeurj.org 9755

FULL PAPERLow-Molecular-Weight Hydrogelators

aggregate of 3, there are no fibers with widths above 8 nmand narrow fibrils of 4–5 nm widths represent nearly 50 %of the total fibers (Figure 2 c). Because both gelators wereanalyzed at concentrations corresponding to half the CGCs,the observed difference in width distribution is consistentwith 3 being a smaller molecule that yields narrower assem-blies than 2 does.

Multistimulus responsiveness of 2 and 3 : The gel–sol transi-tion of 2 and 3 can be triggered by changes in temperature,pH and by the presence of a reducing agent.

Reversibility in response to heat was observed for bothhydrogelators by the stable-to-inversion method and it wasstudied by changes in the absorbance of the gels upon heat-ing and cooling to accurately determine the gelation temper-ature (Figure 3 and Figure 5S in the Supporting Informa-

tion). The hysteresis observed for hydrogel 2 indicates thatgelation is thermally reversible and kinetically controlled.For 3, the hydrogel obtained after the cooling cycle showedan absorption at 330 nm that was more intense than beforethe melting cycle, suggesting thermal degradation (Fig-ure 3 b). The UV/Vis spectrum of the solution obtained bydilution of the hydrogel after the melting/cooling cycleshowed a an increase of the absorption in the 280–320 nmregion relative to the spectrum of the gel before the melt-ing; this is consistent with hydrolysis of the imide moietiesof 3 as observed in the study of the kinetics of hydrolysis of2 and 3 in solution (see above and Scheme 1S in the Sup-

porting Information).[44] The presence of these additionalcarboxylic groups increases the solubility of the hydrolyzedmixture and it is consistent with the observed increase inthe CGC of 3 after the heating/cooling cycle.[48]

In the melting transition of the two hydrogels and in thecooling transition of hydrogel 2, sloping baselines were ob-served. For hydrogel 3, these baselines can be attributed tothe irreversible hydrolytical degradation of the gelator, butfor hydrogelator 2 such degradation was not observed. Thesloping baselines were therefore attributed to structuralchanges that occur within the residual fibrillar assemblies ofthe melted hydrogel. Similar baselines are also observed inthe unfolding of proteins induced by denaturants or temper-ature and they are believed to result from residual structuresin the unfolded state or noncooperative effects, such as theaggregation of hydrophobic surfaces exposed to the sol-vent.[49]

Both hydrogelators showed pH-responsive behavior at-tributed to the ionization of the carboxylic moieties; whenthe pH values of the solutions were changed from alkalineto acidic, the transition to hydrogels was triggered and thesubsequent addition of a base to the hydrogel induced com-plete dissolution, allowing the reversible switching from hy-drogel to solution to be controlled by changes in pH. ThepH reversibility was clearly observed by the stable-to-inver-sion method for both hydrogelators.

A pH-dependent hydrogel-to-solution transition, althougha nonreversible one, could also be triggered by the hydroly-sis of the imide groups of hydrogelator 3. Samples of hydro-gel 3 (15 mm) at pH values of 6.0, 7.0, and 8.0 undergo tran-sition to solution in approximately seven days, three days,and 12 h respectively, as observed by the stable-to-inversionmethod. The kinetics of the hydrolysis of 3 in the hydrogelstate at pH 6.6 and 7.2 was quantified by following thechanges in the UV/Vis spectrum of 3, diluting aliquots ofthe hydrogel below the CGC at different time intervals (Fig-ure 6S in the Supporting Information and Table 1). Similarto what is observed in solution, the hydrolysis in the hydro-gel is pH dependent, but more than an order-of-magnitudeslower. Less than 20 % of the hydrolyzed 3 is observed after10 h of preparation of the gel, at the pH values monitored,in contrast to quantitative hydrolysis observed in solutionsof 3 (Figure 6S in the Supporting Information). This resultshows that the imide moieties are protected from hydrolysisin the hydrogel state; this is consistent with these groupsbeing in a more hydrophobic environment relative to the so-lution state. Such protective modulation of the hydrolysis re-action is also observed for succinimide groups within thefolded state of proteins and is attributed to different expo-sure of these groups to the aqueous environments in addi-tion to other structural factors.[50] The stability to hydrolysisobserved for hydrogel 2 relative to 3 is explained in terms ofpH dependence of the hydrolysis reaction; at the acidic pHthat induces gelation of 2, the hydrolysis reaction is ex-tremely slow (Table 1). Although the gel–sol transition trig-gered by hydrolysis is irreversible, the timescale makes it apotentially useful trigger for drug-release applications as

Figure 3. Thermoreversible gel–sol transitions of a) hydrogel 2 and b) 3monitored by optical density changes at 366 and 370 nm, respectively.Empty symbols represent the melting cycle and filled symbols the coolingcycles. In (a) the two hysteresis cycles corresponds to hydrogelator 2(30 mm) in phosphate buffer (80 mm, pH 4.5) before (black symbols) andafter consecutive addition of DTT and H2O2 (grey symbols). In (b) hy-drogelator 3 (15 mm) in phosphate buffer (80 mm, pH 6.5) was used tocollect the data.

www.chemeurj.org � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eur. J. 2011, 17, 9753 – 97619756

L. Milanesi, S. Tomas et al.

shown by experiments of release of bioactive substancesfrom hydrogelator 3 (see below).

The gel–sol transition of 2 can be reversibly tuned byusing thiol/disulfide exchange as the redox reaction(Scheme 2 a). A gel–sol transition was observed by the vial-

inversion method 30 min after the addition of the reducingagent dithiothreitol (DTT) to hydrogelator 2 (Figure 4 a).The transition to solution is attributed to the reductivecleavage of hydrogelator 2, which yields the nongelatingthiol 2 a (Scheme 2 a), as demonstrated by MS analysis of analiquot of the gelator at different time intervals after addi-tion of DTT (Figure 4 b,c). Addition of an excess of the bio-logical oxidant H2O2 over a DTT-treated sample resulted inthe reverse sol-to-gel transition, attributed to reformation ofhydrogelator 2 by oxidation of thiol 2 a (Scheme 2 a, Fig-ure 4 c).[51] The vial-inversion experiment showed that thesol–gel transition had run to completion in 30 min. An ali-

quot of the hydrogel was analyzed by LC-MS, confirmingthat 2 had been almost completely regenerated after thistime (Figure 4 b,c). Other higher-oxidation-state sulfur com-pounds, which can form by reaction of thiol 2 a with H2O2

and potentially lead to degradation of 2, were not detected(Figure 4 c). Furthermore, the hydrogel obtained after oxida-tion has a melting and cooling cycle identical to that of thehydrogel before the redox cycle, showing that the process isreversible (Figure 3 a). Reformation of the hydrogel also fol-lowed atmospheric oxidation of thiol 2 a. However, in thiscase the sol-to-gel transition took up to 12 h, as showed bythe vial-inversion experiment.

Similarly to 2, hydrogelator 3 contains a disulfide bondthat can undergo reductive cleavage. However, because 3 isa nonsymmetric disulfide, it is not possible to regeneratepure 3 after oxidation of the reduced mixture because thisreaction would lead to the formation of a mixture of disul-fides and thus to a decrease of the CGC of 3.[46] Moreoverthe thiols formed by reduction of 3 can react with the malei-mide moieties leading to the formation of further side prod-ucts (Scheme 2 b).[35] However, whereas the reductive cleav-age of 3 is irreversible it initiates a gel–sol transition that

Scheme 2. a) Conversion of 2 into 2a by reductive cleavage initiated byaddition of excess DTT and reoxidation of 2a to 2 initiated by additionof excess H2O2 or by atmospheric oxygen. b) Reaction of 3 with anexcess of the reducing agents: DTT or GSH.

Figure 4. a) Five photos of inverted hydrogel 3 taken after 10 min of addi-tion of DTT (first test tube on the left) and subsequently at intervals ofapproximately 1 min. b) UV-absorption HPLC traces (detector at250 nm) from the HPLC-MS experiments of an aliquot of 2 in the hydro-gel state (lower trace). The same sample turned into a sol 30 min afteraddition of DTT (middle trace). Addition of H2O2 over the DTT treatedsolution resulted in the hydrogel (30 min after H2O2 addition, uppertrace). The shaded areas labeled with roman numerals yield the corre-sponding MS spectra depicted in (c). c) MS of the corresponding shadedareas of the HPLC chromatogram. Only the masses of the peaks corre-sponding to the major products, 2 and 2a, are shown for clarity. Satellitepeaks labeled with symbols corresponds to fragments of the major prod-ucts 2 and 2a. The masses and structures of these fragments are shown inScheme 2S in the Supporting Information.

Chem. Eur. J. 2011, 17, 9753 – 9761 � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemeurj.org 9757

FULL PAPERLow-Molecular-Weight Hydrogelators

can be used as an effective trigger for the release of drugsentrapped in the hydrogelator (see below).

Release of bioactive substances from 3 : Gelators that can beformed under biocompatible conditions and are responsiveto biologically relevant stimuli are good candidates for thedevelopment of drug-delivery systems.[1,12,22–25,32] We usedhydrogelator 3 with entrapped vitamin B12 as an in vitromodel to investigate the potential of 3 as a controlled-re-lease drug-delivery system. Compound 3 was chosen insteadof 2 because it is the best hydrogelator with a CGC tentimes lower than 2 at a pH range that is biologically relevant(Table 2). Vitamin B12 was chosen as the drug because of thepharmacological importance and spectroscopic propertiesthat allow to follow the release of the drug in a spectralregion (550 nm) in which many gelators, including 3, are op-tically transparent.[24,32] Three different stimuli were used tocontrol the release of the vitamin from the hydrogel:changes in pH, the spontaneous hydrolysis of the gelator,and the reductive cleavage of the disulfide bond of 3 initiat-ed by the addition of glutathione (GSH), a thiol that regu-lates the redox balance in cells (Figure 5).[51]

The release of vitamin B12 was followed by a continuousspectrophotometric assay of a solution that was in contactwith the hydrogel and that was buffered at physiological pH(7.4) or at pH 6.0, (see Experimental Section and Figure 7Sin the Supporting Information, for details of the experimen-tal set up). We found that irrespective of the external stimu-

lus, the release of vitamin B12 is accompanied by the dissolu-tion of the hydrogel by dilution in the surrounding buffer,consistent with a strong affinity of the relatively hydropho-bic drug for hydrogel 3. Release triggered by pH changeswas achieved by preparing the hydrogel with entrapped vita-min B12 at lysosomal pH (5.3) and by using a release solu-tion buffer at pH 7.4.[53] The release of vitamin B12 beginsupon contact with the high-pH buffer. The rate of both vita-min B12 release and gel dissolution is approximately constantfor the first two hours. During this time, around 50 % of thevitamin is released (Figure 5 a). After this time the rate ofvitamin release decreases rapidly and levels off to a valuemore than 30-fold slower; this rate is maintained until totaldissolution of the gel, approximately 60 h after the start ofthe experiment (Figure 5 b). The fast initial release phase isconsistent with a structural change of the gel induced by theexhaustive deprotonation of 3 upon change of the pH. Thesecond, slower release phase is on the other hand attributedto the dissolution of the gel, once the gel reaches the pH ofthe external buffer. This interpretation is supported by theobservation that the rate of release observed in the secondphase is similar to the rate of release observed in absence ofan external stimulus (see below and Figure 5).

Release triggered by spontaneous hydrolysis was achievedby preparing the hydrogel with entrapped vitamin B12 andthe surrounding buffer at the same pH values (7.2 and 6.0,Figure 5). At pH 7.2, the release is biphasic; in the initialphase it is similar to the release observed at pH 6.0(Figure 5) and was therefore attributed to the process of dis-solution of the hydrogel. After this phase during last ap-proximately 40 h, during which nearly 40 % of the vitamin isreleased, the rate suddenly increases four-fold and in thissecond phase the remaining 60 % of the vitamin is released.The onset of the second, faster phase is attributed to thepresence of a critical amount of the more-soluble, hydro-lyzed hydrogelator that increases the rate of gel dissolutionrelative to the first phase of release. At pH 6.0 on the otherhand, the rate of release is approximately constant and simi-lar to that of the first phase of release, observed at pH 7.2,and does not change for the period of time monitored(60 h), consistent with the slower hydrolysis of the hydrogelat lower pH values.

Release of vitamin B12 triggered by the reductive cleavageof the disulfide bond of 3 was achieved by addition of GSHto the buffer surrounding the hydrogel, with the hydrogeland the buffer both at pH 7.2. All the vitamin is released be-tween 3–4 h after addition of GSH in a single phase with arate increase of 30 fold relative to release induced by thedissolution of the hydrogel in the surrounding buffer. Thisincrease in the rate of vitamin release is attributed to the re-ductive cleavage of the disulfide bond of 3 that leads to amixture of thiols that are more hydrophilic and thereforemore readily soluble in the surrounding buffer relative tohydrogel 3 (Scheme 2 b).

Figure 5. a) Initial stages and b) overall time course of vitamin B12 release(5 % mol in the gel) from hydrogel 3 (15 mm) triggered by three differentstimuli : addition of a GSH solution (1 mm, pH 7.4, and hydrogel–vitaminat pH 7.4, &), a pH change (hydrogel–vitamin at pH 5.3 and surroundingbuffer at pH 7.4, ^), and spontaneous hydrolysis of the hydrogelator atpH 7.4 (*, both hydrogel–vitamin and surrounding solution at pH 7.4).The pH 6.0 curve (~), for which no hydrolysis-induced release is ob-served during the monitored period, shows release due purely to the dis-solution of the hydrogel in the surrounding buffer. Note, that the rate ofrelease between approximately 10 and 40 h is very similar for all thestimuli (except reduction, &), and is attributed to the spontaneous disso-lution of the hydrogel.

www.chemeurj.org � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eur. J. 2011, 17, 9753 – 97619758

L. Milanesi, S. Tomas et al.

Conclusion

Our limited ability to predict the assembly properties inwater of novel compounds makes it difficult to design multi-responsive LMWHs. As a consequence, the design of manyLMWHs is typically based on the functionalisation of struc-tural motives that are known to assemble in gel-forming fi-brils and are usually responsive to a single stimulus. The ge-lators described herein—discovered rather than designed—suggest a general strategy to achieve multiresponsivenesswith minimal disruption of the hydrophobic–hydrophilic bal-ance of the molecule to introduce hydrophobic (i.e. , disul-fide bonds) and hydrophilic (i.e. , carboxylic moieties) stimu-li-responsive moieties in the appropriate parts of the mole-cule. The ability of LMWH 2 to gelate water can be con-trolled by at least three stimuli: changes in pH, temperature,and the presence of specific redox agents (Figure 6), show-

ing that 2 is an excellent blueprint for the development ofsmart materials. LMWH 3, on the other hand, shows reversi-ble gel-to-sol transitions upon pH and temperature changes,but also a nonreversible transition due to hydrolytic degra-dation of the imide moieties (Figure 6). Gel 3 has also a re-markable low CGC at physiologically relevant pH values.Therefore 3 is an excellent model of drug-release vehicles,with a release rate that can be modulated by the presence ofthree different stimuli. It is worth noting that the develop-ment of 2 and 3 into derivatives that are more efficient fordrug release or for other soft-nanotechnology applicationsshould be relatively straightforward thanks to the modularstructures that are easily amenable to molecular design.

There are obvious points of this structure that can be modi-fied in the search for novel gelators: other thiols or ligandswith amines can be added to the maleimide moieties anddifferent aromatic disulfides can be appended to the malei-mide group. Previous work has shown that these aromaticdisulfides undergo reversible photocleavage and we antici-pate that this could provide a further tool to expand the re-sponsiveness and hence applicability of these gelators. Cur-rent work in our lab is focusing on these developments.

Experimental Section

Materials and methods : Chemicals and solvents were obtained from com-mercial sources and used without further purification. 1H and 13C NMRspectra were recorded on a Bruker AMX400. All chemical shifts arequoted in ppm and spectra were referenced to the residual solvent signal.All buffers were filtered through a 0.2 mm cellulose membrane (Amicon)and the pH values given are not corrected for the deuterated solutions.Positive electrospray (ES+)-MS were recorded on a Fison VG platformwith a quadrupole detection system. UV/Vis absorbance spectra were re-corded with a Thermoelectron Corporation UV1 Thermospectronic UV/Vis spectrometer and with a Varian Cary 3 spectrophotometer equippedwith a Varian Cary temperature controller for the gel melting measure-ments. For ES+-MS analysis of hydrogel 2, after subsequent addition ofDTT and H2O2, an Agilent 1100 series LC-MS platform equipped with aPhenomenex-Gemini C-18 (3 mm) prepacked column (20 � 4.0 mm) wasused. RP-C18 preparative chromatography was performed with a Phe-nomenex-Jupiter Proteo (90 �, 10 mm) prepacked column (150 �21.2 mm) connected to a Gilson 300 series HPLC setup equipped with aquaternary gradient pump (flow rate: 10 mL min�1), and a UV/Vis detec-tor set at l=220 nm.

Synthesis of 2 and 3 : The procedure for the synthesis of 2 and 3 wasmodified from a previously reported procedure[34] as follows: a solutionof NAC (0.58 g, 3.60 mmol) was added slowly and under vigorous stirringto a solution of 1 (1 g, 2.45 mmol) in MeCN. The reaction mixture wasleft overnight, stirring in a nitrogen atmosphere. The solvent was then re-moved under reduced pressure and the resulting hydrogel was dissolvedin water/dioxane 50:50 v/v. The solvent was then removed by using afreeze drier to give a white residue (1.5 g). This solid was dissolved inDMF (10 mL) and was purified by RP-HPLC eluting with MeCN andH2O, both containing 0.1% trifluoroacetic acid (TFA), with the followinggradient: 0–15 min 50% MeCN, 15–20 min 50!20% MeCN, 20–30 min20!50% MeCN. Fractions corresponding to 2 (tR =15 min) and 3 (tR

=20 min) were collected and concentrated under reduced pressure. Re-sidual water was removed with a freeze drier to give pure 2 (0.9 g, 50 %yield) and 3 (0.42 g, 30 % yield). For the characterization of 2 and 3 seethe Supporting Information.

Hydrolysis measurements : Stock solutions of 2 (2.5 mm) or 3 (2.5 mm) inphosphate buffer (50 mm, pH 6.6, 7.2, and 7.7) were diluted in the samebuffers down to 120 mm (2) and 41 mm (3). The diluted solutions weremaintained at 25 8C and UV/Vis spectra were recorded in the 250–450 nm region by using the automated acquisition of the scanning kinet-ics program of the spectrophotometer. Rate constants were derived byfitting the spectral data to the relevant kinetic models by using SPEC-FIT 3.0.[54] For hydrolysis of 2, an irreversible one-step kinetic model ofthe type A!B was used, whereas for hydrolysis of 3 an irreversible two-step consecutive model of the type A!B!C was used. In these modelsonly the following species are considered: unmodified (A), single hydro-lyzed (B), and hydrolyzed (C) species and it is assumed that hydrolysis ofthe succinimide rings of 2 and imide rings of 3 occurs independentlygiven the distance between the two rings (see the Supporting Informa-tion, Figure 4S for further details).[10] Hydrolysis measurements of hydro-gel 3 were carried out as follows: two samples of hydrogel 3 were pre-pared in phosphate buffer (80 mm , pH 6.6 and pH 7.2) and then aliquotsof the hydrogels were withdrawn and transferred in a vial containing the

Figure 6. Stimuli responsiveness of gel 2 and 3 characterized herein.

Chem. Eur. J. 2011, 17, 9753 – 9761 � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemeurj.org 9759

FULL PAPERLow-Molecular-Weight Hydrogelators

same buffers and diluted down to 41 mm. These solutions containing hy-drogels were sonicated for five minutes and aliquots of the supernatantwere withdrawn for UV/Vis measurements at different time intervals (1–47 h). The absorbance of these solutions was compared with the UV/Visspectra recorded to monitor hydrolysis of 3 below the CGC by usingSPECFIT 3.0.[54] This allowed determination of the percentage of hydro-lyzed 3 in gel 3 from the known percentages of hydrolyzed species deter-mined with SPECFIT for a solution of 3 (e.g., below the CGC, see Fig-ure 6S in the Supporting Information).

Gelation : In a typical experiment, 2 (0.9–5 % w/v) and 3 (0.1–09 % w/v)were dissolved in phosphate buffer (80 mm), at pH 8.4 (for 2) and pH 7.4or 6.2 (for 3) in a vial equipped with a magnetic stirrer. To this solution,small aliquots (1–5 mL to 1 mL sample) of HCl (5–10 m) were added untilformation of a precipitate. This precipitate was dissolved by heating (ap-proximately 50 8C). The warm solution was left to cool at RT. The vialwas then inverted to test for gelation. If the system was not self support-ing, more HCl was added and the above procedure repeated. The finalpH of the hydrogels was then measured.

Electron microscopy : Hydrogel 2 (CGC 30 mm, pH 4.5) and 3 (CGC15 mm, pH 7.4) prepared in phosphate buffer (80 mm) were melted byheating at 50 8C and aliquots of the warm solutions were diluted in dis-tilled water to 15 mm for 2 and 7.5 mm for 3. These solutions (85 mL)were loaded on freshly discharge carbon-coated grids, blotted with filterpaper and negatively stained with 2 % (w/v) uranyl acetate. Low-doseimages were recorded on a 200 keV F20 microscope (FEI) at magnifica-tions of 29000 � , 50000 � , and 80 000 � on Kodak SO-163 film with a de-focus range of 1–3 mm. Micrographs were digitized on a Zeiss SCAI scan-ner at a pixel size of 14 mm corresponding to 4.82, 2.80, and 1.75 � pixelsize on the specimen for magnifications of 29 000 � , 50000 � , and 80 000 �, respectively. The width of approximately 100 fibrils for each hydrogela-tor was determined by using the measure-distance and angles tool ofGIMP 2.6.[55] Only segments of fibrils that appeared straight and flatwere selected for the measurements.

Gel-melting-temperature measurements : Hydrogel 2 (30 mm) and 3(15 mm) prepared in phosphate buffer (80 mm, pH 4.5 and 6.5, respective-ly) were melted to approximately 50 8C and aliquots (200 mL) were trans-ferred to a quartz cuvette with a 1 mm path length and allowed to cool atRT. For gel 2, the sample was first heated to 60 8C and then allowed tocool to 15 8C. The cooled sample was then reheated to 60 8C. The ratewas 2 8C min�1 for both cycles and the absorbance of the sample at366 nm was recorded. For gel 3, the sample was first subjected to a melt-ing cycle from 15 to 60 8C and then to a cooling cycle from 60 to 15 8C.The rate was 3 8C min�1 for both cycles and the absorbance of the sampleat 370 nm was recorded.

Reversible tuning of gelation by pH changes : Hydrogel 2 (1 mL, 30 mm,pH 4.5) was melted to 50 8C under stirring and to the solution, aliquots(1–5 mL) of NaOH (1–5 m) were added up to complete dissolution of thegel phase (pH 5.2, NaOH 6 mm). To the cold solution, aliquots of HCl(1 m) were added under stirring until formation of the gel phase (pH 4.5,HCl 5 mm), as tested by the vial-inversion method.[11] The above proce-dure was repeated three times and the resulting gel was subjected to acycle of melting and cooling as described in the section gel-melting-tem-perature measurements. A similar procedure was used to test pH-reversi-ble gelation of 3 : gel 3 (1 mL, 2.5 mm in 80 mm phosphate buffer pH 4.5)was melted and to a solution NaOH (3 mL, 1 m) was added while stirring.The cold solution was then added aliquots of HCl (1 m) up to formationof the gel phase as test by the vial-inversion method. This procedure wasrepeated three times and a hydrogel was formed for each cycle of the pHswitch. This indicates that hydrolysis of hydrogelator 3 did not take placeduring the course of the pH switch, probably due to the slow kinetic ofthis reaction relative to the timescale of the pH-switch experiment.

Reversible tuning of gelation of 2 by redox reactions : Gelator 2 (30 mm)in phosphate buffer (80 mm, pH 4.5) was melted by heating it to 50 8C ina water bath. An aliquot (10 mL) was diluted in water (500 mL) and wasimmediately subjected to LC-MS analysis (see Figure 4c). To the remain-ing sample, a solution of DTT (5 m) in H2O was added to yield 36 mm

DTT in the sample. An aliquot (130 mL) was transferred to a vial andthen inverted to take pictures at intervals of 1 min. After 30 min after

DTT addition, the hydrogel had turned into a solution. This solution(10 mL) was diluted in water (500 mL) and immediately subjected to LC-MS analysis (see Figure 4c). To the remaining sample, a solution of H2O2

(8.5 m) in H2O was added to yield 42 mm H2O2 in the sample. After30 min the sol had become a gel again. The gel (10 mL) was diluted in ofwater (500 mL) and immediately subjected to LC-MS analysis (see Fig-ure 4c). The remaining gel sample was subjected to a cooling and meltingcycle as described above (see Figure 3a). The timescale and reproducibil-ity of the redox switch were assessed by repeating the above procedurethree times.

Release experiments : In a typical experiment, melted hydrogel 3 (55 mL,30 mm in phosphate buffer 160 mm at the required pH) were mixed withvitamin B12 (55 mL, 0.76 mm) in water (obtained by dilution of a 7 mm

stock solution) to yield a solution containing 15 mm 3 and 0.38 mm vita-min B12 in phosphate buffer (80 mm) at the required pH. This solution(30 mL) was quickly transferred into a volume-displacement plastic pip-ette tip that had had the thin end cut off so that the diameter of the aper-ture of the end of the pipette tip was identical to the diameter of its cy-lindrical section. After the solution became a gel upon cooling at RT, thepipette tip containing the vitamin-loaded hydrogel was place on top of aquartz cuvette in contact with the phosphate buffer (1 mL of 80 mm) atthe appropriate pH and UV/Vis spectra of the surrounding buffer wererecorded up to 80–100 % release of the vitamin (see Figure 7S in the Sup-porting Information). The pH of the stock hydrogel loaded with vita-min B12 and supernatant buffers were identical, except for measurementsof release triggered by a pH change. For these experiments, hydrogel 3was prepared in phosphate buffer (80 mm, pH 5.3) and the same solutionbuffered at pH 7.4 was used as the supernatant. The pH of the hydrogeland the supernatant for measurements of release triggered by hydrolysiswere 6.0 and 7.4, respectively. For measurements of release triggered bya redox reaction, hydrogel 3 was prepared in phosphate buffer (80 mm,

pH 7.4) and the same buffer added glutathione (of 1 mm) was used as thesupernatant. During all the experiments, the pink color of the remaininggel inside the tip did not change, qualitatively showing that the release ofthe vitamin required the dissolution of the hydrogel.

Acknowledgements

We thank Robert Hanson and Simon Thorpe (The University of Shef-field) for technical assistance in RP-HPLC, Dr. Philip Lowden (BirkbeckUniversity of London) for the use of mass spectrophotomer, Prof. SteveCaddick (University College London) for the use of NMR facilities, Prof.Helen Saibil (Birkbeck University of London) for the use of EM facili-ties, the BBRSC, and the Wellcome Trust for funding

[1] M. de Loos, B. L. Feringa, J. H. van Esch, Eur. J. Org. Chem. 2005,3615 – 3631.

[2] L. A. Estroff, A. D. Hamilton, Chem. Rev. 2004, 104, 1201 – 1217.[3] D. J. Abdallah, R. G. Weiss, Adv. Mater. 2000, 12, 1237 –1247.[4] L. A. Estroff, L. Leiserowitz, L. Addadi, S. Weiner, A. D. Hamilton,

Adv. Mater. 2003, 15, 38 –42.[5] R. J. Mart, R. D. Osborne, M. M. Stevens, R. V. Ulijn, Soft Matter

2006, 2, 822 –835.[6] W. H. Binder, O. W. Smrzka, Angew. Chem. 2006, 118, 7482 –7487;

Angew. Chem. Int. Ed. 2006, 45, 7324 –7328.[7] K. S. Moon, H. J. Kim, E. Lee, M. Lee, Angew. Chem. 2007, 119,

6931 – 6934; Angew. Chem. Int. Ed. 2007, 46, 6807 –6810.[8] E. F. Banwell, E. S. Abelardo, D. J. Adams, M. A. Birchall, A. Corri-

gan, A. M. Donald, M. Kirkland, L. C. Serpell, M. F. Butler, D. N.Woolfson, Nat. Mater. 2009, 8, 596 –600.

[9] M. Suzuki, K. Hanabusa, Chem. Soc. Rev. 2009, 38, 967 – 975.[10] K. J. C. van Bommel, C. van der Pol, I. Muizebelt, A. Friggeri, A.

Heeres, A. Meetsma, B. L. Feringa, J. van Esch, Angew. Chem. 2004,116, 1695 –1699; Angew. Chem. Int. Ed. 2004, 43, 1663 – 1667.

[11] F. M. Menger, K. L. Caran, J. Am. Chem. Soc. 2000, 122, 11679 –11691.

www.chemeurj.org � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eur. J. 2011, 17, 9753 – 97619760

L. Milanesi, S. Tomas et al.

[12] J. C. Tiller, Angew. Chem. 2003, 115, 3180 – 3183; Angew. Chem. Int.Ed. 2003, 42, 3072 –3075.

[13] A. R. Hirst, B. Escuder, J. F. Miravet, D. K. Smith, Angew. Chem.2008, 120, 8122 –8139; Angew. Chem. Int. Ed. 2008, 47, 8002 –8018.

[14] Z. Yang, G. Liang, B. Xu, Acc. Chem. Res. 2008, 41, 315 –326.[15] D. Kamarun, X. W. Zheng, L. Milanesi, C. A. Hunter, S. Krause,

Electrochim. Acta 2009, 54, 4985 –4990.[16] N. M. Sangeetha, U. Maitra, Chem. Soc. Rev. 2005, 34, 821 – 836.[17] J. J. D. de Jong, L. N. Lucas, R. M. Kellogg, J. H. van Esch, B. L. Fer-

inga, Science 2004, 304, 278 –281.[18] X. B. Zhao, F. Pan, H. Xu, M. Yaseen, H. H. Shan, C. A. E. Hauser,

S. G. Zhang, J. R. Lu, Chem. Soc. Rev. 2010, 39, 3480 –3498.[19] G. John, P. K. Vemula, Soft Matter 2006, 2, 909 –914.[20] R. G. Ellis-Behnke, Y. X. Liang, S. W. You, D. K. C. Tay, S. G.

Zhang, K. F. So, G. E. Schneider, Proc. Natl. Acad. Sci. USA 2006,103, 5054 –5059.

[21] G. A. Silva, C. Czeisler, K. L. Niece, E. Beniash, D. A. Harrington,J. A. Kessler, S. I. Stupp, Science 2004, 303, 1352 –1355.

[22] Z. Yang, G. Liang, Z. Guo, Z. Guo, B. Xu, Angew. Chem. 2007, 119,8364 – 8367; Angew. Chem. Int. Ed. 2007, 46, 8216 –8219.

[23] F. Plourde, A. Motulsky, A. C. Couffin-Hoarau, D. Hoarau, H. Ong,J. C. Leroux, J. Controlled Release 2005, 108, 433 – 441.

[24] S. Ray, A. K. Das, A. Banerjee, Chem. Mater. 2007, 19, 1633 – 1639.[25] P. K. Vemula, G. A. Cruikshank, J. M. Karp, G. John, Biomaterials

2009, 30, 383 –393.[26] F. Rodr�guez-Llansola, J. F. Miravet, B. Escuder, Chem. Commun.

2009, 7303 – 7305.[27] S. Kiyonaka, K. Sada, I. Yoshimura, S. Shinkai, N. Kato, I. Hamachi,

Nat. Mater. 2004, 3, 58 –64.[28] G. L. Liang, H. J. Ren, J. H. Rao, Nat. Chem. 2010, 2, 54 –60.[29] D. Lçwik, E. H. P. Leunissen, M. van den Heuvel, M. B. Hansen,

J. C. M. van Hest, Chem. Soc. Rev. 2010, 39, 3394 –3412.[30] L. Frkanec, M. Jokic, J. Makarevic, K. Wolsperger, M. Zinic, J. Am.

Chem. Soc. 2002, 124, 9716 –9717.[31] A. Pal, S. Shrivastava, J. Dey, Chem. Commun. 2009, 6997 –6999.[32] H. Komatsu, S. Matsumoto, S. Tamaru, K. Kaneko, M. Ikeda, I. Ha-

machi, J. Am. Chem. Soc. 2009, 131, 5580 –5585.[33] P. Terech, R. G. Weiss, Chem. Rev. 1997, 97, 3133 –3159.[34] L. Milanesi, G. D. Reid, G. S. Beddard, C. A. Hunter, J. P. Waltho,

Chem. Eur. J. 2004, 10, 1705 – 1710.[35] L. Milanesi, C. Jelinska, C. A. Hunter, A. M. Hounslow, R. A. Stani-

forth, J. P. Waltho, Biochemistry 2008, 47, 13620 – 13634. J. D.

[36] C. J. Bowerman, B. L. Nilsson, J. Am. Chem. Soc. 2010, 132, 9526 –9527.

[37] J. D. Hartgerink, E. Beniash, S. I. Stupp, Proc. Natl. Acad. Sci. USA2002, 99, 5133 –5138.

[38] D. K. Smith, Chem. Soc. Rev. 2009, 38, 684 –694.[39] M. Jokic, V. Caplar, T. Portada, J. Makarevic, N. S. Vujicic, M. Zinic,

Tetrahedron Lett. 2009, 50, 509 – 513.[40] Y. Watanabe, T. Miyasou, M. Hayashi, Org. Lett. 2004, 6, 1547 –

1550.[41] J. Schlesinger, C. Fischer, I. Koezle, S. Vonhoff, S. Klussmann, R.

Bergmann, H. J. Pietzsch, J. Steinbach, Bioconjugate Chem. 2009,20, 1340 –1348.

[42] J. Kalia, R. T. Raines, Bioorg. Med. Chem. Lett. 2007, 17, 6286 –6289.

[43] J. T. Doi, M. H. Goodrow, W. K. Musker, J. Org. Chem. 1986, 51,1026 – 1029.

[44] F. M. Menger, L. Shi, S. A. A. Rizvi, J. Colloid Interface Sci. 2010,344, 241 –246.

[45] A. H. Fuhrhop, T. Y. Wang, Chem. Rev. 2004, 104, 2901 – 2937.[46] O. Tr�ger, S. Sowade, C. Bottcher, J. H. Fuhrhop, J. Am. Chem. Soc.

1997, 119, 9120.[47] P. Knight, Biochem. J. 1979, 179, 191 –197.[48] A. R. Hirst, I. A. Coates, T. R. Boucheteau, J. F. Miravet, B. Escud-

er, V. Castelletto, I. W. Hamley, D. K. Smith, J. Am. Chem. Soc.2008, 130, 9113 –9121.

[49] M. A. C. Reed, C. Jelinska, K. Syson, M. J. Cliff, A. Splevins, T. Ali-zadeh, A. M. Hounslow, R. A. Staniforth, A. R. Clarke, C. J.Craven, J. P. Waltho, J. Mol. Biol. 2006, 357, 365 –372.

[50] L. Athmer, J. Kindrachuk, F. Georges, S. Napper, J. Biol. Chem.2002, 277, 30502 –30507.

[51] M. D. Rees, E. C. Kennett, J. M. Whitelock, M. J. Davies, Free Radi-cal Biol. Med. 2008, 44, 1973 – 2001.

[52] L. Milanesi, C. A. Hunter, S. E. Sedelnikova, J. P. Waltho, Chem.Eur. J. 2006, 12, 1081 –1087.

[53] R. G. W. Anderson, L. Orci, J. Cell Biol. 1988, 106, 539 –543.[54] H. Gampp, M. Maeder, C. J. Meyer, A. D. Zuberbuhler, Talanta

1986, 33, 943 –951.[55] http://www.gimp.org.

Received: February 28, 2011Published online: July 26, 2011

Chem. Eur. J. 2011, 17, 9753 – 9761 � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemeurj.org 9761

FULL PAPERLow-Molecular-Weight Hydrogelators