Review
Self-assembly of Urea Derivatives into Supramolecular Gels
2025 Volume 73 Issue 6 Pages 497-510
Details
2025 Volume 73 Issue 6 Pages 497-510
In this review, we present our development of low-molecular-weight gelators (LMWGs) with urea moieties. A C3-symmetric tris-urea framework was found to be an excellent structure for LMWG. Tuning the molecular structure has enabled the creation of LMWGs that gel in a wide range of media, ranging from organic solvents to water. The introduction of appropriate functional groups into LMWGs can be reflected in a function of the resulting supramolecular gel. Our developed supramolecular gels were applied as substrates for electrophoresis, and the separation of proteins and DNA was achieved. Considering material applications, we also developed structurally simplified mono-urea derivatives as LMWGs.
Urea is one of the iconic molecules in organic chemistry. In 1828, Wöhler discovered that the organic substance urea could be formed from inorganic substances. This result disproved vitalism that only living organisms could synthesize organic substances from inorganic substances. For life, urea is the final metabolite. Proteins that have completed their roles are broken down and eventually converted into urea. Urea is then excreted from the body as a waste product of life. We have been attracted to the structure and function of urea and have been working on the creation of supramolecules based on urea derivatives. This can be seen as a challenge in creating functional molecules (supramolecules) by utilizing the waste products of life. In this review, we introduce our laboratory’s efforts to create functional supramolecular gels based on urea structures.
The structure of urea is simple yet distinctive. The structural properties of the urea molecule are as follows: 1) high planarity, 2) oxygen atoms acting as hydrogen bond acceptors, and 3) hydrogen atoms acting as hydrogen bond donors. This structure is ideal for forming hydrogen bonds (Fig. 1). A variety of supramolecular architectures have been developed through the rational design and synthesis of urea derivatives.1,2) By designing homolytic intermolecular hydrogen bonds, supramolecular capsules with isolated nano-spaces, supramolecular polymers as pseudo-polymer substances, and supramolecular gels as functional soft materials can be created.
The self-assembly of organic molecules, called low-molecular-weight gelators (LMWGs), in an appropriate medium gives a gel-like object, that is, a supramolecular gel.3–5) Supramolecular gels, which are formed by non-covalent interactions throughout the formation process, show highly flexible and stimulus-responsive properties.6,7) Because the molecular structure of an LMWG can be reflected in a property of the resulting supramolecular gel, functional supramolecular gels can be created based on the molecular design of an LMWG. As a pioneering study in this field, Shinkai and colleagues synthesized an LMWG with a photo-responsive azobenzene group, and the supramolecular gel showed a gel-to-sol phase transition in response to photoirradiation.8) Research on supramolecular gels has been stimulated by this result, and the development of a wide variety of functional supramolecular gels has been realized.9,10) Naturally occurring structures such as steroids, fatty acids, and peptides are commonly used as the basic framework of LMWGs. On the other hand, artificial molecules, such as 1,2-cyclohexanediamine derivatives, have also been found to be LMWGs, and the structural diversity of LMWGs has been dramatically increasing.11,12) In the molecular design of LMWGs, relatively strong intermolecular interactions, such as hydrogen bonding and/or π–π interactions, are often introduced into their structures. Urea is a typical hydrogen-bonded functional group found in LMWGs. We are interested in the high designability of urea, and have been working on the creation of supramolecules using urea derivatives. The following is an overview of our urea-based LMWGs.
At the beginning, our research interest was in supramolecular capsules formed by discrete self-assembly.13) In particular, we focused on the development of hybrid supramolecular capsules that form driven by non-covalent bonds of different strength, namely hydrogen bonds and metal-ligand coordination bonds.14,15) In this pursuit, we developed a hybrid supramolecular capsule with a cavitand framework, a urea derivative as the hydrogen bonding site, and a pyridyl group and divalent palladium as the coordination bonding sites. We also achieved control of the dissociation rate of guest molecules in response to external stimuli.16) As a development of this research, we were interested in developing a new framework to replace the cavitand, and designed and synthesized C3-symmetric tris-urea 1 with 1,3,5-triethylbenzene as its framework. Our research into supramolecular gels began when 1 unexpectedly formed supramolecular gels.
C3-Symmetric tris-urea 1 was synthesized in four steps using 1,3,5-triethylbenzene as the starting material17) (Fig. 2a). A mixture of 1 and acetone gave an insoluble suspension, which became a supramolecular gel upon sonication using an ultrasonic cleaner (Fig. 2b). The minimum gelation concentration (MGC) in the formation of the acetone gel was 1.5 wt%. The IR spectra supported hydrogen bonding between the ureido groups, and scanning electron microscopy (SEM) measurements revealed fibrous aggregates with a high aspect ratio. These results suggest that 1 forms a supramolecular gel via dimensionally controlled self-assembly.
As the ureido groups are known to strongly associate with anions, we expected the supramolecular gel of 1 to change its association mode in response to the addition of anions, resulting in a phase transition of the supramolecular gel. Therefore, when fluoride ions were added as tetrabutylammonium salts, the supramolecular gels underwent a phase transition to a homogeneous solution, as expected (Fig. 2c). The addition of anions other than fluoride, such as chloride, bromide, and acetate, also caused a phase transition of the supramolecular gel of 1 (Fig. 2c). However, the amounts of anions required for the phase transition were unique for each anion, and the NMR titration results showed a correlation between 1 and the association constant of the anions. Re-gelation of the phase transition solution was also examined. When boron trifluoride etherate was added, only the solution that underwent a phase transition due to the addition of fluoride ions selectively underwent re-gelation, whereas the solution containing other anions remained a solution (Fig. 2d). The fluoride ion reacted with trifluoroborane to form tetrafluoroborate,18) which weakly associates with the ureido groups, resulting in re-gelation, while the same reaction did not occur with the other anions. In contrast, when metal salts such as zinc bromide were added, the solutions generated by adding any of the anions underwent non-selective re-gelation19) (Fig. 2e). NMR analysis showed that the anions associated with the ureido groups dissociated with the addition of the metal salt, indicating that the association between the metal salt and the anion was preferential and that non-selective re-gelation proceeded.
Since the structure of tris-urea 1 facilitated the synthesis of various derivatives, more than 20 derivatives were synthesized, and their gelation abilities were evaluated.20) Interestingly, the introduction of long alkyl groups into the outer shell, which is generally considered effective for gelation, was not effective in this system, and the presence of aromatic substituents was effective for gelation. The fact that the gelation ability was maintained regardless of the substituents introduced into the aromatic ring moiety provides a major guideline for the molecular design of various derivatives, as described in the following sections.
An exceptional derivative, tris-urea 2, in which an octylureido group was introduced into the para-substituted framework derived from phloroglycinol as the starting material, served as the LMWG for haloalkanes21) (Fig. 3a). A mixture of 2 and 1,1,2-trichloroethane produced a supramolecular gel when cooled after being heated to dissolve (Fig. 3b). A mixture of 2 and 1,1,2,2-tetrachloroethane produced a supramolecular gel when cooled after being heated to dissolve in the presence of an appropriate amount of copper(II) bromide (Fig. 3c). A mixture of 2 and chloroform was sonicated to form a supramolecular gel (Fig. 3d). A mixture of 2 and dichloromethane was sonicated in the presence of an appropriate amount of bismuth (III) chloride to form a supramolecular gel (Fig. 3e). Thus, it was found that mixtures of 2 and chloroalkanes produced supramolecular gels when treated under suitable conditions.
Because the outer aromatic moiety of tris-urea 1 is tunable, we investigated the creation of functional supramolecular gels by introducing functional groups to this moiety. For example, tris-urea (R,R,R)-3 [or (S,S,S)-3] with a chiral substituent gelated various organic solvents, and its MGC was 2.0 wt% in acetone22) (Fig. 4). However, its MGC was 4.0 wt% in racemic rac-3, a 1 : 1 mixture of (R,R,R)-3 and (S,S,S)-3 (Fig. 4). As shown in this typical example, the chiral substituent introduced into the tris-urea affected its gelation ability.
Tris-urea 4 with pyridyl groups, which function as coordination sites or hydrogen bonding acceptors, was synthesized23) (Fig. 5a). 4 could co-assemble with 1 to form a fibrous aggregate, and the addition of divalent palladium or dicarboxylic acid as a cross-linker between the fibrous aggregates was expected to reduce the MGC. In fact, the addition of Pd(OAc)2 or isophthalic acid in an amount equivalent (equiv.) to 1 mol% 4 for 1 reduced the MGC from 1.5 wt% to 0.75 wt%.
The mixture of 4 and Pd(OAc)2 also functioned as a gelator for dimethyl sulfoxide (DMSO) or a DMSO-water mixture.24) The supramolecular gel showed phase transition properties in response to chelating agents such as ethylenediamine and acted as a catalyst for an organic reaction (Fig. 5b). The supramolecular gel functioned as a catalyst for the Wacker oxidation reaction that converted styrene to acetophenone, and the catalytic activity tended to be retained for extended periods of time compared to the reaction in the bulk solution (Fig. 5c).
We set our next research target as the development of water gelation, that is, supramolecular hydrogels. Because hydrogels are expected to be highly biocompatible, they can be applied in various fields, such as regenerative medicine and drug delivery. Based on our developed C3-symmetric tris-urea framework, we decided to design a low-molecular-weight hydrogelator (LMWHG). For a molecule to function as a hydrogelator, it must have an amphiphilic structure with both hydrophobic and hydrophilic properties. Therefore, we designed the C3-symmetric amphiphilic tris-urea molecule 5, in which a glucose derivative was introduced as a hydrophilic group into the outer shell of the hydrophobic tris-urea structure25) (Fig. 6a). C3-Symmetric amphiphilic tris-urea 5 was considered to self-assemble in water in a one-dimensional shape so that the central hydrophobic part avoids contact with water, forming a supramolecular hydrogel.
C3-Symmetric amphiphilic tris-urea 5 was mixed with water, heated to dissolve, and allowed to stand at room temperature to form a transparent supramolecular hydrogel with an MGC of 1.5 wt% (Fig. 6b). SEM observation of the xerogels revealed homogeneous fibrous aggregates. Transparent supramolecular hydrogels were also formed by simply mixing 5 with water and allowing it to stand at room temperature (Fig. 6c). The MGC in this method was 2.0 wt%. SEM observations of the xerogels prepared from the room-temperature-conditioned supramolecular hydrogels revealed fibrous aggregates that were finer than those generated by the heating method. The supramolecular hydrogel prepared at room temperature remained transparent for a while but turned into a cloudy gel after several weeks. The transparency of the gel is related to its refractive index, and the presence of thick fibrous aggregates and the heterogeneity of the aggregate structure reduce its transparency. Therefore, SEM observation of the xerogel prepared from the cloudy supramolecular hydrogel revealed heterogeneous fibrous aggregates, including thick fibrous aggregates of several hundred nanometers in size. The supramolecular hydrogels prepared by heating showed homogeneous fibrous aggregates with an appropriate bundling progress, whereas the supramolecular hydrogels prepared at room temperature formed supramolecular hydrogels without sufficient bundling progress. Therefore, it is considered that the bundling progressed gradually after the formation of the supramolecular hydrogels, resulting in the formation of heterogeneous fibrous aggregates.
Supramolecular hydrogel 5 showed anion responsiveness. The phase transition from gel to solution was observed by adding sodium salts of various anions. For most monovalent anions, the phase change was completed by adding two equiv., and for divalent anions, by adding one equiv. Because the anions are hydrated in water, they all associate with the ureido groups with the same strength. Therefore, in supramolecular hydrogels, no difference in the anion species was observed during the anion-induced phase transition. Utilizing this non-selective anion response, we visually evaluated the hardness of mineral water (Fig. 6d). A mixture of Crystal Geyser® and 5 in soft water gave a transparent supramolecular hydrogel (Fig. 6d (i)). A mixture of Paradiso® and 5, classified as medium-hard water, formed a cloudy supramolecular hydrogel (Fig. 6d (ii)). In contrast, mixtures of Wattwiller® or Contrex® and 5, which are classified as hard water, gave only a cloudy suspension and no supramolecular hydrogel (Fig. 6d (iii, iv)).
The glucosides introduced as hydrophilic groups of 5 densely accumulated on the surface of the fibrous aggregates that constituted the supramolecular hydrogel. Therefore, we hypothesized that the macroscopic phase transition in supramolecular hydrogels would proceed even with the addition of concanavalin A (ConA), a sugar recognition protein that recognizes glucosides. When 0.009 equiv. of ConA were added to a supramolecular hydrogel prepared from 5 and water, the gel that was clear before the addition became cloudy. The gel-sol phase transition temperature, which was 38 °C before the addition, increased to 85 °C with the addition of ConA. This suggests that the fibrous aggregates were cross-linked by ConA, resulting in cloudiness of the supramolecular hydrogel and an increase in the gel-sol phase transition temperature (Fig. 6e (iv)). When the amount of added ConA was increased to more than 0.016 equiv., the phase transition from gel to suspension occurred (Fig. 6e (i)). We considered that the addition of sugar was strongly associated with ConA. When α-methylmannose (Me-α-Man), which is strongly associated with ConA, was added, the supramolecular hydrogel was regenerated (Fig. 6e (ii, v)). In contrast, the addition of galactose (Gal), which is not associated with ConA, did not regenerate the supramolecular hydrogel (Fig. 6e (iii)).
C3-Symmetric amphiphilic tris-urea 6 was designed as an LMWHG26) (Fig. 7a). A mixture of 6 and water gave only a suspension containing precipitates, and no supramolecular hydrogel was formed. When sodium dodecyl sulfate (SDS), an anionic surfactant, was added to the suspension, supramolecular hydrogels were formed. The amount of SDS affected the gelation, which changed from suspension to a cloudy gel, a clear gel, and a homogeneous solution as the amount of SDS increased (Fig. 7b). This change was followed by SEM observations; no specific shape of the aggregates was observed in the SEM images of the samples without SDS, but nano-sized fibrous aggregates were observed in the samples with SDS. The thickness of the fibrous aggregates decreased as the amount of SDS increased. Thin, homogeneous fibrous aggregates were observed in a homogeneous solution containing large amounts of SDS.
The properties of surfactants that induce the gelation of 6 were investigated. Various alkyl sodium sulfates were used in this study. Sodium sulfates with short alkyl chains, ranging from sodium methyl sulfate to sodium pentyl sulfate, remained in suspension even when added in excess amounts and did not induce the formation of supramolecular hydrogels. In contrast, all the sodium sulfates with long alkyl chains above the heptyl group induced gelation of the suspension consisting of 6 and water. A similar trend was observed when alkylammonium bromide was used as the additive. The molecules that induced the gelation of 6 were all ionic molecules that functioned as surfactants. Gelation with neutral surfactants such as octyl glucoside and tetraethylene glycol monododecyl ether was attempted, but this did not induce gelation of the suspension composed of 6 and water. From these results, we concluded that the ionic surfactants induced gelation of the suspension composed of 6 and water (Fig. 7c). The concentration of the ionic surfactant required to induce gelation was lower than the respective critical micelle concentration. 6 forms supramolecular polymers in water driven by hydrophobic interactions and intermolecular hydrogen bonds. However, 6 did not have sufficient hydrophilic groups to fully cover the surface of the formed supramolecular polymer. Therefore, excessive bundling occurs due to hydrophobic interactions, resulting in a suspension containing precipitates. When an ionic surfactant is added, the long-chain alkyl groups penetrate the supramolecular polymer through hydrophobic interactions. Consequently, the ionic hydrophilic moieties of the ionic surfactant accumulate on the surface of the supramolecular polymer. Electrostatic repulsion between the ionic hydrophilic moieties unbundles the supramolecular polymers and forms a supramolecular hydrogel when moderately sized fibrous aggregates are formed.
In the process of evaluating the gelation ability of the C3-symmetric amphiphilic tris-ureas 5 and 6, we were reminded of the importance of a balance between hydrophilicity and hydrophobicity in the design of molecules that can function as LMWGs for water. 6, which is insufficiently hydrophilic, cannot function as a hydrogelator. However, 5, which is more hydrophilic than 6, functioned as a hydrogelator. Based on these results, we designed and synthesized C3-symmetric amphiphilic tris-urea 7, which has a reduced hydrophobic moiety in the structure of 6 to make it a suitable amphiphilic molecule that can also function as an LMWG for water27) (Fig. 8a).
C3-Symmetric amphiphilic tris-urea 7 formed a supramolecular hydrogel when mixed with water, dissolved by heating, and allowed to cool, with an MGC of 0.25 wt% (1.9 mM) (Fig. 8b). SEM observations of the xerogels revealed homogeneous fibrous aggregates. The formed supramolecular hydrogel was translucent at 0.25 wt%, but became cloudy with increasing concentration (1.0 wt%, Fig. 8c (iii)). The thermal stability of the supramolecular hydrogel also increased with increasing concentration. The supramolecular hydrogel of 7 was stable in water without swelling or shrinkage. The supramolecular hydrogel of 7 exhibited thixotropic properties and typical thermal reversibility. When the supramolecular hydrogel of 7 was stirred using a vortex mixer, it underwent a phase transition from gel to suspension, and the suspension recovered to a gel by standing. This phase transition was repeated more than 20 times. SEM observations of the sample that underwent a phase transition to a sol by stirring revealed fibrous aggregates that were more heterogeneous than those of the xerogel. From this result, it was considered that the supramolecular hydrogel of 7 underwent a phase transition due to the dissociation of the physical cross-links between the fibrous aggregates constituting the gel and the unbundling of the fibrous aggregates upon stirring.
C3-Symmetric amphiphilic tris-urea 7 was able to gel not only water, but also various other aqueous solutions (Fig. 8c). For example, a wide range of acidic and basic buffers (phosphate buffer, borate buffer, Tris buffer, N-(2-hydroxyethyl)piperazine-N′-2-ethanesulfonic acid (HEPES) buffer, etc.) and Dulbecco’s modified Eagle’s medium (DMEM) used for cell culture could be gelated. Saline, pseudo-body fluids, and even untreated seawater could be gelated with 0.25 wt% of 7. The gelation of aqueous solutions with various salt concentrations was examined, and it was found that even saturated saline solution (5.35 M) could be gelated by 0.25 wt% (1.9 mM) of 7. Aqueous solutions within a wide range of pH could be gelated by 7: it formed supramolecular hydrogels in 0.01 M HCl (pH = 2.0, Fig. 8c (ii)) and 0.01 M NaOH solutions (pH = 12.0, Fig. 8c (iv)) as well as in 8 M HCl (Fig. 8c (i)) and 7 M KOH solutions (Fig. 8c (v)). Under acidic conditions, gels were formed despite the hydrolysis of the acetal moiety of 7, and each supramolecular hydrogel was stable at room temperature for at least one year.
Amphiphilic molecules consisting of a hydrophobic C3-symmetric tris-urea core with sugar and ethylene glycol moieties as the hydrophilic groups in the outer shell were evaluated as LMWGs for water. One problem associated with these compounds is that they require multiple synthesis steps. Therefore, we decided to develop a new LMWG for water that could be synthesized more easily. We designed the C3-symmetric amphiphilic tris-urea 8 with carboxy groups as the hydrophilic moiety28) (Fig. 9a). The synthesis of 8 could be performed in short steps from commercially available compounds.
Mixtures of 8 with water or acidic aqueous solutions gave only insoluble suspensions, even after heating. In contrast, a mixture of 8 and basic NaOH solution gave a highly viscous solution. SEM observations of this solution revealed fibrous aggregates. The reason why supramolecular hydrogels were not formed despite the formation of fibrous aggregates was considered to be the electrostatic repulsion between the deprotonated carboxylates. A supramolecular hydrogel was formed by the addition of hydrochloric acid to a viscous solution (Fig. 9b). The MGC of 8 was estimated to be 0.3 wt% (3.0 mM). The supramolecular hydrogel underwent a reversible phase transition in response to pH. The addition of a base, such as NaOH solution, to the supramolecular hydrogel of 8 caused a phase transition from gel to sol, and the addition of an acid caused a phase transition from sol to gel.
Gelation of the viscous solution formed from 8 and NaOH solution proceeded upon the addition of metal salts, such as calcium chloride (Fig. 9b). This gelation is considered to be caused by the cross-linking of calcium ions between the carboxylates. Therefore, the addition of [2,2,2]cryptand, a host molecule that traps calcium ions, to the supramolecular hydrogel caused a gel-to-sol phase transition. A supramolecular hydrogel was also obtained by adding a rare-earth metal salt, terbium triflate [Tb(OTf)3]. The supramolecular hydrogel exhibited green luminescence based on the energy transfer from 8 to terbium ions.29)
The development of an effective adsorbent for cleansing polluted water is necessary for environmental purification. Supramolecular hydrogels constructed by self-assembly are promising candidates for this purpose. It was found that the supramolecular hydrogel of 8 efficiently adsorbed cationic organic dyes.30) The supramolecular hydrogel of 8 adsorbed 1–2 M equivalents of rhodamine 6G to 8. The adsorption of methylene blue was even more efficient with 4.19 M equivalents to 8. The adsorption mechanism of the supramolecular hydrogel was investigated by measuring its luminescence spectra. Two emission peaks with different wavelengths and fluorescence lifetimes were observed for the rhodamine 6G-adsorbed supramolecular hydrogel, the ratio of which varied with the amount of adsorbed dye. These results suggest that adsorption proceeds initially by interaction with 8 on the fiber, and after the adsorption is saturated, it proceeds in the aqueous space of the supramolecular hydrogel.
Considering applied research using supramolecular hydrogels, a conversation with a friend led to the idea of developing them as substrates for electrophoresis. Polyacrylamide gel is a commonly used substrate for the electrophoresis of protein samples; however, researchers are not completely satisfied with polyacrylamide gels. In particular, protein electrophoresis using polyacrylamide gels often suffers from inefficient recovery of the protein sample from the gel after electrophoresis because of the strong interactions between the protein and polyacrylamide. We hypothesized that it would be possible to adjust the affinity between the protein and the supramolecular hydrogel because the supramolecular hydrogel is formed by weak non-covalent interactions. Furthermore, we believe that the high designability of the supramolecular hydrogels will be useful for the development of substrate-specific electrophoresis. For these reasons, we decided to develop an electrophoresis method based on a supramolecular gel, which we named SUGE (supramolecular gel electrophoresis).31)
Initially, we developed an electrophoresis method using a supramolecular hydrogel as an alternative to the most versatile sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) method. Because C3-symmetric amphiphilic tris-urea 6 can form a gel of tris-glycine-SDS solution (a typical solution used in SDS-PAGE), this supramolecular hydrogel was applied to the electrophoresis of denatured proteins.32) After various investigations, the following experimental procedure was established (Fig. 10): I) Filling a glass capillary (inner diameter: 2 mm, length: 8 cm) with both 2.0 wt% of supramolecular hydrogel of 6 and agarose; II) Adsorbing the denatured protein sample on one end of the gel and filling both ends of the glass capillary with agarose; III) Performing the electrophoresis of the capillary filled with the sample using a submarine electrophoresis apparatus; IV) Removing the electrophoresed gel from the glass capillary and dividing it into 8 equal portions (numbered from 1 to 8 in order from the anode side; V) Extracting protein samples and confirming of the mode of separation by typical SDS-PAGE. One notable property of supramolecular hydrogels is that approximately 50% of the protein sample was recovered by the simple addition of a buffer solution to the gel and centrifugation.
Electrophoresis was performed using protein samples ranging from 6.5 to 116 kDa. Electrophoresis of a mixture of ovalbumin (45 kDa) and β-galactosidase (116 kDa) showed that small ovalbumin was detected in lanes 3 and 4, with a strong band in lane 3 (Fig. 11a). Large β-galactosidase was detected in lanes 3–6, with the strongest band observed in lane 4 (Fig. 11a). Similar separation behavior was observed in the electrophoresis of ovalbumin (45 kDa) and bovine serum albumin (66 kDa), with proteins with smaller molecular weights showing larger mobility (Fig. 11b). However, when proteins with molecular weights less than 45 kDa were used, a unique separation behavior was observed, with larger proteins showing greater mobility. In electrophoresis of a mixture of aprotinin (6.5 kDa) and ovalbumin (45 kDa), ovalbumin was detected in lanes 3–5 and aprotinin in lane 7 (Fig. 11c). We attributed these results to the competition between two different separation mechanisms in electrophoresis using a supramolecular gel. One is the molecular sieving effect, also seen in SDS-PAGE, in which string-like denatured proteins pass through a three-dimensionally entangled fibrous network, with smaller molecular weight proteins showing greater mobility. Another is the size exclusion effect seen in gel filtration, where proteins smaller than a certain size are retained in the isolated space formed by the supramolecular gel, resulting in greater mobility for larger proteins. For proteins with molecular weights larger than 45 kDa, the molecular sieve effect appears to be the dominant separation mechanism, whereas for proteins with molecular weights less than 45 kDa, the size exclusion effect appears to be the dominant separation mechanism. It is also clear that the molecular weight of the protein sample at which the dominant separation mechanism switches can be adjusted by the SDS concentration of the supramolecular gel.33)
In the SDS-SUGE study, it was found that protein samples could be efficiently recovered from supramolecular gel matrices after electrophoresis using a simple operation. To utilize this property more effectively, we decided to investigate native protein electrophoresis (native-SUGE) using a supramolecular gel. In this study, we used C3-symmetric amphiphilic tris-urea 7, which can be used to gel a wide range of aqueous solutions, including TG solutions (Tris: 25 mM, glycine: 195 mM) commonly used in native protein electrophoresis (native-PAGE).
The experimental procedure for native-SUGE was similar to that for SDS-SUGE, but a supramolecular gel prepared from 7 without agarose was used for native-SUGE because the supramolecular gel formed from 7 had sufficient strength for electrophoresis.34) Native acidic proteins, D-lactate dehydrogenase (146 kDa, pI = 4.0), β-galactosidase (540 kDa, pI = 4.6), and ovalbumin (45 kDa, pI = 4.7), were used as samples. In native-SUGE of a mixture of D-lactate dehydrogenase and β-galactosidase, D-lactate dehydrogenase was detected in lanes 4 and 5 and β-galactosidase in lanes 5 and 6 (Fig. 12a). The smaller, more acidic D-lactate dehydrogenase showed greater mobility toward the anode than the larger, less acidic β-galactosidase. In the native-SUGE of D-lactate dehydrogenase and ovalbumin, D-lactate dehydrogenase was detected in lanes 3 and 4, and ovalbumin was detected in lanes 5 and 6 (Fig. 12b). The larger and more acidic D-lactate dehydrogenase showed greater mobility toward the anode than the smaller and less acidic ovalbumin. In the native-SUGE of β-galactosidase and ovalbumin, β-galactosidase was detected in lanes 4 and 5, and ovalbumin in lanes 5 and 6 (Fig. 12c). Again, the larger and more acidic β-galactosidase showed greater mobility toward the anode than the smaller and less acidic ovalbumin. These results indicate that in native-SUGE using the supramolecular gel of 7, native acidic protein samples were separated mainly by their isoelectric points, with little influence from their molecular weights.
The activity of D-lactate dehydrogenase after native-SUGE was measured to confirm that the protein samples maintained their native structures after native-SUGE. Native D-lactate dehydrogenase catalyzes the oxidation of lactic acid to pyruvate in the presence of nicotinamide adenine dinucleotide (NAD+) and retains more than 90% of its activity after native-SUGE. Native-SUGE of green fluorescent protein (GFP, 27 kDa, pI = 5.57) and red fluorescent protein (RFP, 27 kDa, pI = 5.65) was performed. Supramolecular hydrogels were irradiated with UV light during and after electrophoresis, and green and red fluorescent bands were observed. The fluorescent bands observed in the supramolecular gel of 7 were consistent with those of GFP and RFP detected after SDS-PAGE analysis.
The surfaces of the nanofibers formed by the self-assembly of 7 are densely coated with glucosides, which were introduced as hydrophilic groups. Glucoside-coated nanofibers can interact with appropriate lectins during electrophoresis, and the lectins involved in this interaction are not as mobile as non-sugar-binding proteins in supramolecular gels. A representative lectin, concanavalin A (ConA, as tetramer = 112 kDa, pI = 4.4–5.5), was subjected to affinity electrophoresis in a supramolecular gel of 7.
The affinity between α-methyl-D-glucopyranoside and ConA (Ka = 1.96 × 103 M−1) indicates a moderate association between 7 and ConA. Electrophoresis of ConA was performed under typical native-SUGE conditions (100 V, 100 min), and most of the ConA remained at the cathode (lane 8) (Fig. 13a). In contrast, denatured ConA was electrophoresed under the same conditions and was observed in anodic lanes 4 and 5 (Fig. 13b). These results indicated that the interaction of native ConA with glucosides on the nanofiber surface inhibited its electrophoretic mobility. The electrophoretic mobility of native ConA in native-SUGE can be improved by adding sugar with a strong affinity for ConA to the electrophoretic buffer. α-Methyl-D-mannopyranoside (MeαMan) was chosen because of its stronger affinity for ConA (Ka = 0.82 × 104 M−1) than α-methyl-D-glucopyranoside. TG-MeαMan solution (Tris: 25 mM; glycine: 192 mM; MeαMan: 51 mM) was used for the electrophoresis of native ConA. SDS-PAGE analysis showed that native ConA was electrophoresed and was detected in lanes 5 and 6 (Fig. 13c).
Polyacrylamide gels have pore sizes as small as 5–100 nm and are suitable for separating small DNA fragments of 500 base pairs (bp) or less. In contrast, agarose gels with pore sizes of 200–500 nm are commonly used to separate much larger DNA fragments (>100 bp). In these types of electrophoresis, large DNA fragments move more slowly than smaller DNA fragments. In addition, DNA fragments larger than 20 kbp are difficult to separate in general electrophoresis because they move at similar speeds. Pulsed-field gel electrophoresis was developed as a technique to separate large DNA fragments.35,36) Pulsed-field gel electrophoresis is useful for cloning large DNA fragments, karyotyping microorganisms, and epidemiological analysis of infectious diseases, but it requires specialized apparatus and long analysis times. The native-SUGE results suggest that the network structure of the supramolecular gel of 7 is coarse and flexible enough to allow large biomacromolecules to pass through. Large DNA fragments were thought to be separated by the supramolecular gel of 7 using a common electrophoresis apparatus.
Electrophoresis of large DNA fragments (DNA-SUGE) was performed using a supramolecular gel of 7 and TBE solution (Tris: 45 mM, boric acid: 45 mM, ethylenediaminetetraacetic acid (EDTA): 1.0 mM).37) DNA-SUGE was performed using almost the same procedure as native-SUGE. In this study, three types of DNA markers were analyzed: λ-Hind III digest (DNA fragments from 2 to 23 kbp), Lambda DNA-Mono Cut Mix (DNA fragments from 10 to 49 kbp), and Marker 7 GT (DNA fragments from 10 to 165 kbp). After electrophoresis, the supramolecular gels were divided into 8 or 20 equal portions. The extracts were analyzed by typical DNA electrophoresis using agarose H gels to analyze the separation manners.
DNA-SUGE was applied to analyze DNA markers of λ-Hind III digests containing 2.0, 2.3, 4.4, 6.6, 9.4, and 23.1 kbp DNA fragments (Fig. 14a). Electrophoresis was performed using a 2.0 wt% supramolecular gel of 7 at 150 V for 90 min. The DNA fragments were separated by their length, with shorter fragments showing greater mobility than longer fragments. The 2.0 kbp DNA fragments were found in the most anodic lane 2, with 2.3, 4.4, and 6.6 kbp DNA fragments in lanes 3, 4, and 5, respectively. The 9.4 and 23.1 kbp large DNA fragments appeared mainly in lane 6, and a narrow band corresponding to a 23.1 kbp DNA fragment was observed in lane 7. It is noteworthy that under these conditions, the six DNA fragments were precisely separated, with complete separation of the 2.0 kbp and 2.3 kbp DNA fragments.
The separation of Lambda DNA-Mono Cut Mixes containing 10.1, 15.0, 17.1, 24.0, 24.5, 30.0, 33.5, 38.4, and 48.5 kbp DNA fragments was examined (Fig. 14b). Electrophoresis conditions suitable for the separation of λ-Hind III digests were not suitable for the separation of the Lambda DNA-Mono Cut Mix. A low concentration of supramolecular gel and low voltage were effective for separating large DNA fragments. DNA-SUGE of the Lambda DNA-Mono Cut Mix was performed using a 1.0 wt% supramolecular gel of 7 at 50 V for 6 h. After electrophoresis, the supramolecular gel of 7 was divided into 20 equal portions and the separation manner was analyzed by electrophoresis using an agarose H gel. Clear separation of DNA fragments of 10.1, 15.0, and 17.1 kbp was achieved and the 10.1 kbp DNA fragments was mainly detected at lane 4; the 15.0 kbp DNA fragment was mainly detected in lane 7; the 17.1 kbp DNA fragment was detected mainly in lane 8. 24.0–48.5 kbp DNA fragments were observed in lanes 10–13.
The separation of Marker 7 GT containing 10.1, 17.7, 21.1, 23.5, 41.8, 50.3, and 165.7 kbp DNA fragments was examined (Fig. 14c). DNA-SUGE with 2.0 wt% supramolecular gel of 7 showed proper separation at 40 V and 9 h. The 10.1 and 17.7 kbp DNA fragments were detected only in lanes 9 and 11, respectively. DNA fragments of 21.1 and 23.5 kbp were detected in lane 12. The 41.8 kbp DNA fragments were detected in lanes 13–15 and 50.3 kbp DNA fragments in lanes 14 and 15. The largest DNA fragment, 165.7 kbp, was detected only in lane 15. It is worth mentioning that 50.3 and 165.7 kbp DNA fragments showed different mobility in DNA-SUGE.
A stable supply of LMWGs as materials is an extremely important issue for applied research on supramolecular gels. In fact, our SUGE studies were abandoned due to the detailed conditions and scaled-up studies being hindered by the synthesis of LMWG, which requires more than 10 steps. Therefore, we decided to restart our research with a view toward the application and development of LMWGs synthesizable in large quantities. Urea groups are widely known as effective functional groups for LMWGs. Thus, we systematically synthesized N-alkyl-N′-aryl-urea derivatives that can be synthesized in a single step from commercially available compounds. Their gelation experiments showed that many of the urea derivatives formed supramolecular gels with multiple organic solvents.38) Among them, N-alkyl-N′-2-benzylphenyl-urea had a particularly high gelation ability. We synthesized derivatives with alkyl groups ranging from 2 to 18 and evaluated their gelation abilities in a wide range of organic solvents.39) As a result, we were able to identify the appropriate alkyl chain length depending on the polarity of the solvent, which is an important guideline for the design of gelator molecules. In general, the introduction of long alkyl groups with carbon chains of 12 or more is effective for gelation of highly polar organic solvents, whereas shorter alkyl groups are more effective for the gelation of organic solvents with low polarity.
An amphiphilic molecule is a common structure of LMWGs for water. We decided to develop amphiphilic urea derivatives as LMWGs for water. For the hydrophilic groups, saccharides are desirable from the viewpoints of function and structural diversity. During the synthesis of sugar derivatives, the number of steps tends to increase because of the protection and deprotection of the hydroxyl groups. However, we thought that the synthesis could be achieved in short steps by selecting a reaction using an unprotected sugar. Aminoglycosylation reaction using an unprotected sugar and an amine is considered an ideal reaction to meet these requirements. Indeed, the aminoglycosylation reaction allows the synthesis of amphiphilic urea 9 in only two steps40) (Fig. 15a). The synthesized 9 had a high gelation ability in water, and the MGC was estimated to be 0.3 wt%. The hydrophilic lactose moiety of 9 would be hydrolyzed into galactose and glucose by lactase. The supramolecular hydrogel of 9 changed from a gel to a suspension over several hours in the presence of lactase (Fig. 15b). Analysis of the NMR spectra of the suspension revealed that this phase transition was caused by the hydrolysis of the lactose moiety of 9.
As lactase is localized in the small intestine in the human body, oral administration of the supramolecular hydrogel is expected to cause hydrolysis and decomposition of the supramolecular hydrogel in the small intestine. We thought that drug delivery to the small intestine could be realized using a drug-adsorbed supramolecular hydrogel. We first used dyes as pseudo-drugs and evaluated the stability of the supramolecular hydrogels and the retention of the dyes. As a result, cationic dye molecules such as rhodamine 6G were adsorbed without impairing the stability of the supramolecular hydrogel of mono-urea derivative 10, and the adsorption was maintained in water41) (Figs. 15c, d). Furthermore, when lactase was added to the mixture of the rhodamine 6G-adsorbed supramolecular hydrogel and water, the supramolecular hydrogel gradually decomposed, and the adsorbed rhodamine 6G was released accordingly (Fig. 15e). To extend the generality of this concept, we also developed amphiphilic ureas with maltose as the hydrophilic group and found that the supramolecular hydrogels underwent a phase transition in response to maltase.42)
We developed various urea-based LMWGs and evaluated their functions. Our research began with the discovery of C3-symmetric tris-urea as an LMWG for organic solvents. The C3-symmetric tris-urea framework is an excellent structure for LMWGs, and in many cases, its gelation ability is retained even after structural modifications such as the introduction of functional groups. Encouraged by the high gelation ability of this framework, we developed C3-symmetric amphiphilic tris-ureas as LMWGs for aqueous solutions. Some of the supramolecular hydrogels formed from C3-symmetric amphiphilic tris-ureas were found to be applicable as substrates for the electrophoresis of biomacromolecules such as proteins and DNA. To develop LMWGs suitable for material applications, we developed urea and amphiphilic urea derivatives that can be synthesized in short steps. Amphiphilic urea, which forms enzyme-responsive supramolecular hydrogels, has potential applications in DDS. Surprisingly, urea derivatives rarely failed to function as LMWGs, and in many studies, they functioned as competent molecules in our research project. We are convinced that urea, which is a waste product of life, is a very attractive structure for our research and that many more highly functional molecules can be created.
This work was supported by Grant-in-Aid for the Scientific Research (No. 19750111, 22750125, 23107514, 24310089, 25107713, 15H03826, 17H06374, 21K06485 for M.Y.; 23K14327, 24H01729 for S.K.) the Japan Society for the Promotion of Science (JSPS) or the Ministry of Education, Culture, Sports, Science and Technology (MEXT). We would like to thank many co-workers for their kind cooperation in promoting research. These results are the efforts of the graduate students in our laboratory, and we are deeply grateful for their wholehearted efforts. We would like to express our sincere gratitude to our many friends who have made our life of research so glamorous and enjoyable. “A heartfelt thank you to my (M.Y.) family, Seiko, Mao, and Daiki, for always providing a pleasant research environment.”
The authors declare no conflict of interest.
This review is written in commemoration of Professor Yamanaka’s receipt of the 2023 Pharmaceutical Society of Japan Award for Divisional Scientific Promotion. Unfortunately, he passed away on March 12, 2024, at the age of 50 years. This manuscript is based on Prof. Yamanaka’s unfinished draft, which has been carefully refined and completed by Dr. Kimura. We express our deep respect for Prof. Yamanaka in honor of his lifelong achievements and unwavering dedication to research until the very end.
