Gelation-driven component selection in the generation of constitutional dynamic hydrogels based on guanine-quartet formation

Abstract

The guanosine hydrazide 1 yields a stable supramolecular hydrogel based on the formation of a guanine quartet (G-quartet) in presence of metal cations. The effect of various parameters (concentration, nature of metal ion, and temperature) on the properties of this gel has been studied. Proton NMR spectroscopy is shown to allow a molecular characterization of the gelation process. Hydrazide 1 and its assemblies can be reversibly decorated by acylhydrazone formation with various aldehydes, resulting in formation of highly viscous dynamic hydrogels. When a mixture of aldehydes is used, the dynamic system selects the aldehyde that leads to the most stable gel. Mixing hydrazides 1, 9 and aldehydes 6, 8 in 1:1:1:1 ratio generated a constitutional dynamic library containing the four acylhydrazone derivatives A, B, C, and D. The library constitution displayed preferential formation of the acylhydrazone B that yields the strongest gel. Thus, gelation redirects the acylhydrazone distribution in the dynamic library as guanosine hydrazide 1 scavenges preferentially aldehyde 8, under the pressure of gelation because of the collective interactions in the assemblies of G-quartets B, despite the strong preference of the competing hydrazide 9 for 8. Gel formation and component selection are thermoreversible. The process amounts to gelation-driven self-organization with component selection and amplification in constitutional dynamic hydrogels based on G-quartet formation and reversible covalent connections. The observed self-organization and component selection occur by means of a multilevel self-assembly involving three dynamic processes, two of supramolecular and one of reversible covalent nature. They extend constitutional dynamic chemistry to phase-organization and phase-transition events.

Supramolecular entities present the ability to reversibly modify their constitution through exchange and rearrangement of their molecular components because of the lability of the noncovalent interactions that hold them together (1, 2). Similar features may be imported into molecular species if reversible covalent bonds are introduced into their structure, allowing cleavage and formation of interatomic connections with fragment exchange under specific conditions.

Thus, entities, capable of reversible modification of their constitution, define a constitutional dynamic chemistry on both the supramolecular and the molecular levels (3). Because the constitutional changes may be expected to respond to external factors, constitutional dynamic chemistry is the basis for the design and development of adaptive chemical systems. It generates constitutional dynamic libraries (CDLs) whose constituents are in dynamic equilibrium, such that they can exchange their components and express all of the entities that are potentially accessible through recombination by means of reversible covalent bonds and noncovalent interactions. The CDL may then adapt to (internal or) external physical factors or chemical effectors by selection (3, 4) of the appropriate components for the optimal constituent.

Such processes form the basis of the recently developed dynamic combinatorial chemistry (5, 6), in which molecular recognition events have been implemented toward the generation of optimal binding agents toward artificial or biological (7) molecular receptors through a target-driven shift in the distribution of the library constituents toward the best binder(s). Changes in library constituents may also be caused by redistribution of components induced by metal ion binding (8, 9) or environmental factors such as temperature and pH (N. Giuseppone and J.-M.L., unpublished data).

The amplification of a given constituent of a CDL under the pressure of a self-organization process, such as the formation of an organized phase (for example, a gel), would be of special interest. It would represent a process of self-organization by selection (3, 4) by which the formation of a structured phase drives the selection of the components that make up the dynamic constituent producing the most highly organized and most stable assembly.

Gels attract much current interest for their potential as intriguing materials (10, 11) and as substrates for biomedical applications (12, 13). In particular, hydrogels formed from low-molecular-weight compounds that respond to pH are suitable candidates for oral drug delivery as well as biosensor technology, especially when biochemical components are involved. Hydrogels have also been shown to respond to ligand-receptor molecular recognition (14, 15) and redox stimuli (16).

Here, we describe our studies of a system in which the formation of a supramolecular hydrogel drives the selection of the components that form the constituent leading to the most stable gel. It embodies a process of self-organization by selection under the pressure of gelation. It presents triple constitutional dynamics, two at the supramolecular level and a third one of covalent dynamic nature, which involves selection by covalent self-assembly of the component that generates the hydrogel of highest cohesive strength.

The system brings together several features of particular interest, namely (i) self-organization and dynamics at both the supramolecular and molecular levels; (ii) generation of dynamic hydrogels; (iii) dynamic selection of the optimal components; (iv) implementation of biochemical components; and (v) adaptive behavior in response to external factors.

Materials and Methods

Instrumental Techniques. NMR spectra were recorded on a 400-MHz spectrometer (Bruker, Wissembourg, France). The chemical shifts are reported in ppm downfield from tetramethylsilane; coupling constants are given in Hz. Electrospray ionization (ESI)-MS was carried out on a Bruker MicroTOF MS coupled with liquid chromatography. Samples were prepared at a concentration of 200 μM in Milli-Q water or in 0.5 M ammonium acetate buffer. Before injection, a small aliquot of the sample was diluted 20-fold and used for ESI-MS. The following mild conditions were used to detect the supramolecular assembly of guanine quartets (G-quartets): dry heater temperature was set at 120°C, ion polarity was positive, nebulizer pressure was 0.4 bar, capillary voltage was 4,000 V, end plate offset voltage was -400 V, and dry gas flow was 3.0 liters/min. Viscosity measurements were performed on a digital rheometer (model DV-III; Brookfield, Middleboro, MA) fitted with a CPE-40 spindle model of 4 cm in diameter and a 1° angle.

Preparation of Gels and Measurement of Gel Melting Temperature. The guanosine hydrazide 1 was dissolved in 0.5 M acetate buffer (pH 6, 500 μl) to make up a concentration of 15 mM. The container was heated until guanosine hydrazide 1 dissolved completely, and it was then cooled to room temperature. Gelation was observed, and the gel melting temperature (T_gel) was determined visually by the vial-inversion method. The sample vials were immersed in an inverted position in an oil bath, and the temperature was increased slowly. T_gel was taken as the point at which the gel started to flow.

Determination of the Gelled Fraction by Proton NMR Spectroscopy. We added guanosine hydrazide 1 (2.3 mg), ²H₂O (450 μl), 45 μl of KCl stock solution (1 M), and 5 μl of dioxane stock solution (300 mM) to an NMR tube, yielding 15 mM 1/90 mM KCl/3 mM dioxane solution. The NMR tube was gently heated until guanosine hydrazide dissolved, and it was then cooled to room temperature. The ¹H-NMR spectra were recorded when the sample was fully gelled. The percentage of free hydrogelator was determined by integrating the proton signal of H-8 on the guanine group with respect to the internal dioxane reference. The proton signal of H-8 is sharp for free 1 in solution, whereas that for 1 engaged in the gel is broadened beyond detection. Integration of the observable H-8 signal with respect to the internal reference gave the amount of 1 still free in the gelled solution. The gel was destroyed by addition of a drop of concentrated DCl giving a clear solution, and then the proton NMR spectrum was recorded; the integration of H-8 of 1 (100% free in the solution) with respect to the internal dioxane reference gave the amount of 1. The difference between the integral of total free 1 (100% in solution) and the integral of free 1 in the gelled solution yielded the percentage of gelation. The integration of the internal dioxame signal was also checked against an external dioxane reference contained in a capillary.

Dynamic Combinatorial Library Generation. Stock solutions (150 mM) of the aldehydes (6-8) and hydrazides (9 and 10) were prepared by dissolving a given compound in ²H₂O or deuterated buffer solution [0.5 M sodium acetate or potassium acetate, p²H (pD) 6.0]. Guanosine hydrazide 1 (2.3 mg) was dissolved in a 500-μl buffer in an NMR tube to make up a 15 mM solution. Then, we added 50 μl of hydrazide (9 or 10) solutions and 50 μl of aldehyde (6 and 8) solutions from the stock solutions. The NMR tube was gently heated to 50-60°C for 5-6 h to reach equilibrium. It was then cooled to room temperature, and the ¹H-NMR spectrum was recorded once the solution was fully gelled. The CH═N proton signals of the anti and syn isomers of each acylhydrazone (≈75% anti and 25% syn ± 10%, depending on the compound) were integrated, giving the fraction of the library constituents present as free in solution. The fraction of guanosine acylhydrazone in the gel was obtained by difference. Although the acylhydrazone (1+6) does not give a gel, a small amount (≤3%) of it could be trapped in the gel formed by the acylhydrazone (1+8) in the mixtures of (1+9 or 10) with (6+8). The spectra of the individual acylhydrazones (15 mM) showed a weak signal (≤5%) of unreacted aldehyde proton. On heating, the CH═N signals broadened both for the individual compounds and for the mixtures, and the signals due to the anti and syn forms merged progressively.

Synthesis. All commercial reagents were purchased from Aldrich and used without any further purification.

Guanosine-5′-Hydrazide 1. We added hydrazine hydrate (80 mg, 1.6 mmol) to a suspension of guanosine-5′-methylester (17) (100 mg, 0.32 mmol) in methanol (150 ml), and the mixture was refluxed for 12 h. The reaction mixture was concentrated to one-third of its volume, filtered, and dried under a vacuum to give 76 mg of hydrazide 1 (yield, 76%; white solid; mp, 241-243°C): ¹H NMR (400 MHz, DMSO-d₆): 10.8 (br s, 1H), 10.6 (br s, 1H), 7.95 (s, 1H), 6.57 (br s, 2H), 5.78 (d, J = 7.6 Hz, 1H), 5.73 (m, 1H), 5.54 (br s, 2H), 4.52 (m, 1H), 4.45 (m, 1H), 4.33 (d, J = 1.2 Hz, 1H), 4.1 (s, 1H). ¹³C-NMR (100.6 MHz, DMSO-d₆): δ 169.2, 157.1, 153.9, 150.3, 137.7, 117.8, 88.0, 84.6, 73.5, 72.6. ESI-MS m/z 312.1 [M+H]⁺; high-resolution MS (fast atom bombardment MS) found 312.1083, (C₁₀H₁₄N₇O₅ calculated 312.0978). CHN analysis calculated for C₁₀H₁₃N₇O₅ 0.5 H₂O: C, 37.57; H, 4.39; N, 30.67; found: C, 37.01; H, 4.35; N, 30.67.

Results and Discussion

In an ongoing exploration of the dynamic combinational chemistry of derivatives of biological substances [amino acids, peptoids (18), carbohydrates (19, 20), nucleic acid components (D. T. Hickman, N. Sreenivasachary, and J.-M.L., unpublished data)], we synthesized the guanosine-5′-hydrazide derivative 1 by treatment of the corresponding methylester (17) with hydrazine.

Figure 6.

Compounds containing guanine groups are well known to undergo quadruple association into G-quartets through Hoogsten-type hydrogen-bonding forming supramolecular macrocycles that stack into G₄ assemblies in the presence of cations such as Na⁺,K⁺, and NH₄⁺ with formation of hydrogels (Fig. 1) (21, 22). Organo-soluble derivatives of guanine have also been shown to form organized materials such as liquid crystals (23, 24). These results led us first to explore the physicochemical properties of 1.

Fig. 1.

Guanine derivatives self-assemble to G-quartets in the presence of metal ions.

Gelation Properties of Guanosine Hydrazide 1. Guanosine hydrazide 1, although not soluble by itself in pure water, was found to form stable free-standing gels at 15 mM in the presence of Na⁺, K⁺, and NH₄⁺, as well as of the much larger Me₄N⁺ cation at neutral pH (phosphate buffer). The hydrogels were strong enough not to flow on inversion of the container (Fig. 2a) and were found to be stable at room temperature for several days. They are pH-sensitive, forming around neutral and slightly acidic pH (acetate buffer) but not at pH 8.0 or 9.0 (borate buffer). Electrospray MS showed a peak for (G₄+Na) at m/z 1,267.4, confirming the identity of the G-quartet. The transmission electron microscopy observation of the gel prepared from 1 revealed fibers of several micrometers in length (Fig. 2b).

Fig. 2.

Hydrogel formed from 1.(a) Picture showing that the sample does not flow when the vial is inverted at 15 mM, 23°C in 0.5 M sodium acetate buffer at pH 6.0. (b) Transmission electron microscopy images of fibers forming the gel.

The temperatures of gelation, T_gel, were measured at pH 6 in sodium acetate buffer (0.5 M) as a function of gelator concentration (Fig. 3a). The results indicated that hydrazide 1 was able to gelate the buffer solution even at a concentration as low as 10 mM [i.e., ≈0.3% (wt), giving a T_gel value of 33°C]. At 50 mM, T_gel was 65°C, and increasing the concentration up to 100 mM gave product precipitation. The formed hydrogel was thermally reversible but unstable to shear.

Fig. 3.

Temperature of gelation T_gel determined visually as a function of hydrazide 1 concentration in buffer (0.5 M) of the acetates of various cations as follows. (♦), Na⁺;(▪), K⁺;(▴), NH₄⁺;(⋄), Me₄N⁺.(a) T_gel values (in °C for 10 and 50 mM, ion): (33, 65, Na⁺), (59, 87, K⁺), (45, 73, NH₄⁺), and (54, 79, Me₄N⁺). (b) T_gel of an aqueous solution of 1 (15 mM) as a function of K⁺ concentration (in mM, °C): (15, 41), (30, 51), (45, 61), (60, 61), and (90, 61).

Determination of T_gel as a function of the concentration of 1 for different cations showed that K⁺ was the most efficient gelator/G₄ assembler. The other cations showed the sequence of gelation efficiency Me₄N⁺ > NH₄⁺ > Na⁺ (Fig. 3a), in line with the results reported for guanosine gels (21).

The gelation temperature T_gel of 1, determined visually by container inversion, as a function of K⁺ concentration, remained unchanged at 61°C above 45 mM and up to 90 mM salt, indicating that it was independent of ion concentration when gelation was complete (Fig. 3b).

According to previous studies (21), the gelation properties of compound 1 may be attributed to the formation of G-quartets and subsequent stacking into columns induced by binding of metal cations between the G₄ species (Fig. 1). However, the present gels are much more stable than reported for guanosine (21).

Proton NMR Determination of Guanosine Hydrazide 1 Gelation. The fraction of 1 engaged in the hydrogel was determined by ¹H-NMR spectroscopy. Indeed, whereas the signal of H-8 on the guanine group is sharp for free 1, that for 1 engaged in the gel is broadened beyond detection. Integration of the observable H-8 signal with respect to an internal reference signal (3 mM dioxane) gave the fraction of 1 free and in the gel (by difference) as a function of K⁺ concentration (Fig. 4a) (see Materials and Methods). It was found that practically total gelation (≥98%) occurred above ≈45 mM KCl. Thus, NMR spectroscopy offers a general method for studying the efficiency of the gelation processes. Similarly, the variation of the fraction of free 1 as a function of temperature was followed by integration of the H-8 proton signal (Fig. 4b). It displays a sigmoidal shape similar to the melting curve of double-stranded DNA, yielding a transition temperature T_t of 43°C.

Fig. 4.

Investigation of the gelation of 1 (15 mM, ²H₂O) by ¹H-NMR spectroscopy. (a) Fraction of free 1 (▴), 1 engaged in the gel (▵) as function of K⁺ concentration (lines drawn through the experimental data points). (b) Fraction of free 1 in the presence of 90 mM KCl as a function of temperature. Experimental data points and the calculated best fit to a sigmoidal curve are given. (Inset) Derivative curve giving the transition temperature T_t

The difference between the visually determined T_gel (61°C) and T_t (43°C) may be ascribed to the fact that they concern two different events. Different physical methods refer to different microscopic processes (25). T_t refers to the variation at the molecular level of the amounts of free and bound (in the gel) states of 1, possibly involving motions within fibrils without depolymerization. T_gel describes the gel-to-sol transition at the macroscopic level, when the material flows under gravity shear because of loss of cohesion of the assembly. Also, the NMR data indicate full melting of the gel at ≈55-60°C (onset of the plateau of the curve in Fig. 4b). Such data may be of much interest for the understanding of the relationship between microscopic and macroscopic collective events in gels and organized phases in general. They also point to the fact that motion within or exchange in and out of an organized phase may occur before phase transition, a feature of significance for delivery processes.

Formation of Dynamic Gels Derived from Guanosine Hydrazide 1. The G-quartets formed by hydrazide 1 are potentially interesting scaffolds for dynamic decoration. This feature offered the possibility of exploring the effect of side chains attached through reversible condensation with various aldehydes (in particular, of biological type) on the properties of the hydrogel. More specifically for the present purposes, 1 displays a hydrazide functionality that may engage into the formation of reversible acylhydrazones by condensation with various aldehyde (or ketone) compounds, thus generating dynamic libraries of acylhydrazones as documented (18, 20, 26-28). As a result, reversible decoration of 1, its G₄ derivative Q₁ (Fig. 5), and the gel-forming assemblies by diverse groups becomes possible, with the perspective of modulating the properties of the organized phase and inducing the selection of a specific group. Thus, we conducted an extensive investigation involving hydrazide 1 and a series of aldehydes 2-8 (Fig. 5; in all cases, 15-mM solutions of each component in sodium acetate buffer in ²H₂O at pD 6.0 and 25°C).

Fig. 5.

Reversible decoration of guanosine hydrazide 1 and of its G-quartet assembly Q₁ by condensation with various aldehydes (2-8).

The addition of stoichiometric amounts of the aldehydes 2-4 to the gel formed by 1 disrupted the gel and resulted in product precipitation, whereas aldehydes 5 and 6 gave solutions at the same concentration. The reaction of 1 with 1-formyl furan-3-sulfonic acid 7 and with pyridoxal-5-phosphate 8 yielded highly viscous gels. The condensation reactions with 5-8 were followed by ¹H-NMR spectroscopy observing the disappearance of the aldehyde proton signal (see above). The acylhydrazones generated from 1 with 5 or 6 each showed two imine proton signals corresponding to the syn and anti imine isomers.

Rheological measurements on the gels formed by the acylhydrazone derivatives Q₂ obtained from aldehydes 7 and 8 indicated that they had a much higher viscosity (2,400 and 1,900 mPa, respectively, at 0.38 turns s^-1 shear rate) than the gel formed by the hydrazide 1 itself and that they presented both thermal and shear stress reversibility (Fig. 6). Similar shear stress reversibility is observed for supramolecular polymers (ref. 29 and E. Kolomiets and J.-M.L., unpublished data).

Fig. 6.

Viscosity of the gels formed from the acylhydrazone derivatives obtained by reacting guanosine hydrazide 1 with the aldehydes 7 (▴) and 8 (▵).

Gelation-Driven Component Selection. The promising results obtained for the formation of highly viscous hydrogels by the acylhydrazone quartet derivatives Q₂ of guanosine hydrazide 1 prompted us to explore whether the gelation process would drive the selection, within a dynamic library of constituents, of the component generating the constituent that yields the strongest gel. It would represent a process of self-organization by selection, driven by the cohesive strength due to the collective interactions in the resulting assembly. On a related note, size-selective synthesis of oligomers may be induced by folding (30).

To investigate the evolution of the system toward the “best-fit” dynamic hydrogel, the aldehydes 6, 8 and the hydrazides 1, 9 were selected (Fig. 7). The dynamic library was generated at 15 mM concentration for each compound, at pD 6 in sodium acetate buffer (0.5 M). It consists of four acylhydrazones, each presenting two configurational isomers, undergoing interchange continuously by acylhydrazone bond formation and cleavage in aqueous medium (18, 20, 26-28).

Fig. 7.

Generation of a dynamic library of acylhydrazones C, D and of the acylhydrazone G-quartets A and B from hydrazides 1, 9 and aldehydes 6 and 8.

A mixture of the hydrazides 1, 9 and the aldehydes 6, 8 was prepared (see Materials and Methods) in a molar ratio of 1:1:1:1 (Fig. 7 and Table 1, entry 1). After gently heating the mixture at 60°C and then cooling it to room temperature, the ¹H-NMR spectra were recorded when equilibrium was attained (<6 h). The distribution of acylhydrazone components at 25°C was analyzed by ¹H-NMR spectroscopy, by integrating the CH═N imine protons of the free (nongelated) acylhydrazones, which, although broadened, could be clearly identified for each constituent of the mixture (Table 1). A markedly uneven distribution was obtained (Fig. 8). The guanosine hydrazide 1 gave 8% and 39% of the acylhydrazones A and B, resulting from its reaction with aldehyde 6 and pyridoxal monophosphate 8, respectively. Similarly, the serine hydrazide 9 reacted with aldehydes 6 and 8 to give ≈42% of C as well as 11% of D.

Table 1. Equilibrium distribution of acylhydrazones A-F in the dynamic mixtures generated from the hydrazides 1, 9, and 10 and the aldehydes 6 and 8.

	Hydrazides			Aldehydes		Acylhydrazones at the equilibrium, %
Entry	1	9	10	6	8	A	B	C	D	E	F
1	1	1	0	1	1	8	39	42	11	—	—
2	1	0	1	1	1	9	37	—	—	40	12
3	0	1	0	1	1	—	—	15	85	—	—
4	0	2	0	1	1	—	—	50	50	—	—
5	1	1	0	0	1	—	87	—	13	—	—
6	1	0	1	0	1	—	96	—	—	—	4
7	1	1	0	1	0	48	—	52	—	—	—
8	1	0	1	1	0	51	—	—	—	49	—
9	0	1	1	1	1	—	—	25	25	25	25

All compound fractions have been determined by proton NMR spectroscopy. The concentration of all compounds was 15 mM in sodium acetate buffer at pD 6 and 25°C (see Materials and Methods).

Fig. 8.

Distribution of the acylhydrazones A-D in the CDL generated from a mixture of the hydrazides 1, 9 and the aldehydes 6 and 8 (all 15 mM, sodium acetate buffer, pD 6 in ²H₂O) as a function of temperature after reaching equilibrium, as calculated from integration of the CH═N ¹H-NMR signals. At 25°C: A, 8%; B, 39%; C, 42%; D, 11%; at 55°C: A, 15%; B, 35%; C, 35%; D, 15%; at 80°C: A, 22%; B, 28%; C, 28%; D, 22%.

When the ¹H-NMR spectra were measured at 55°C, the distribution of acylhydrazones was found be become less uneven. On further temperature increase up to 80°C, the gel was melted completely and the distribution of acylhydrazones was close to equal (Fig. 8). Cooling the reaction mixture slowly over a period of 60 min back to 25°C restored the initial distribution, indicating that a selection process occurred, by which the mixture evolved to favor the constituent B forming a thermodynamically stable dynamic hydrogel over constituents A, C, and D, which do not give such an organized phase. As indicated by the ¹H-NMR data, two acylhydrazones B (in the gel) and C (free in solution) clearly dominate in the CDL at equilibrium. The latter is expressed as “image” of B, as a consequence of D being depressed by the trapping of pyridoxal monophosphate 8 in B in the gel. Note that the dynamic selection is reversible and depends on the temperature; at high temperature, when the gel has melted, the selection disappears, whereas it operates at 25°C when the medium is gelled. Thus, a self-organization by selection occurs, driven by gelation and inducing the preferential formation of the most stable hydrogel. Its generality is shown on replacement of serine hydrazide 9 by alanine hydrazide 10, which results in almost the same distribution of acylhydrazones (Table 1, entry 2).

Several control experiments were performed (Table 1). Equimolar amounts of hydrazide 9 and aldehydes 6 and 8 (1:1:1 ratio) gave 15% of acylhydrazone C and 85% of acylhydrazone D (Table 1, entry 3), indicating that hydrazide 9 forms preferentially D with aldehyde 8. As expected, hydrazide 9 and aldehydes 6 and 8 in a 2:1:1 ratio generate equal amounts of C and D at equilibrium (Table 1, entry 4). Aldehyde 8 and hydrazides 1, 9 in a 1:1:1 molar ratio resulted in gel formation, giving 87% acylhydrazone B and 13% of D at equilibrium (Table 1, entry 5). Together, these results stress the ability of gelation to redirect the acylhydrazone distribution because hydrazide 1 is able to scavenge 8 from 9 in D despite the strong preference of 9 for 8 (entry 3). Similarly, compounds 1, 10 and 8 in a 1:1:1 ratio yielded acylhydrazones B and F at 96% and 4% respectively, indicating again strong selection when there is gelation (Table 1, entry 6). Reacting 1, 9 with 6 (1:1:1 ratio) or 1, 10 with 6 (1:1:1 ratio) gave almost equal distribution of imines because no gelation occurs to drive a selection (Table 1, entries 7 and 8). Similarly, when 9, 10 were mixed with 6, 8 in a 1:1:1:1 ratio, there was no component selection and the fractions of acylhydrazones C, D, E, and F were found to be equal (Fig. 9 and Table 1, entry 9). These experiments highlight that the guanosine hydrazide 1 cation-templated self-assembly drives component selection enforced by the ability of the supramolecular assembly B to form a stable hydrogel.

Fig. 9.

Dynamic library of acylhydrazones C-F generated from hydrazides 9, 10 and aldehydes 6 and 8.

Conclusions

The pH-sensitive guanosine hydrazide 1 was found to be a powerful hydrogelator, capable of undergoing dynamic decoration through formation of reversible acylhydrazone bonds with various aldehydes. It is a versatile building block for combinations presenting various gelation abilities depending on the appended group. It opens up the possibility to generate biologically relevant dynamic hydrogels by means of the constitutional dynamic approach. New gelators, bearing in particular biologically relevant recognition groups, can be explored systematically for the generation of aqueous gels that may find useful applications in the area of medicinal chemistry and material science. The gelation process results from a multilevel self-assembly based on the cation-templated self-assembly of quartets of guanosine acylhydrazone derivatives. In a broader perspective, within the frameworks of self-organization by selection and constitutional dynamic chemistry (3), the present results demonstrate the ability of a physical parameter, the formation of a cohesive organized phase (here, a hydrogel), to direct a constitutional dynamic system toward the selection of the components that lead to the expression of the “fittest” constituent for the directing process in a function-driven fashion.