Supramolecular morphological transition driven by nonequilibrium break-and-build self-organisation

Abstract

Controlling the nonequilibrium dynamics in synthetic systems is an important challenge in supramolecular chemistry. Biological morphogenesis exploits nonequilibrium reaction–diffusion processes to organise differentiated cell networks and functions in space. In contrast, artificial assemblies are mostly formed under thermodynamic control, and they thus lack autonomous spatial complexity. Here, we present a supramolecular morphological transition that enables spatial control over the differentiation and decomposition of synthetic self-assembled fibres. Upon hybridisation with a surfactant, self-assembled fibres made from low-molecular-weight peptides undergo nonequilibrium break-and-build dynamics through cycles of fibre decomposition and reformation. Coupling these dynamics with surfactant diffusion generates repeated propagating waves that produce macroscopic nonlinear concentric patterns of chemically distinct fibres. These resultant patterns can be modulated by applying transient thermal pulses or by introducing multiple surfactants. This approach provides a strategy for the hierarchical organisation of supramolecular assemblies, opening up opportunities for constructing complex, functional soft materials with controlled spatial architectures.

Controlling the nonequilibrium dynamics in synthetic systems is an important challenge in supramolecular chemistry but assemblies are often formed under thermodynamic control and lack autonomous spatial complexity. Here, the authors present a supramolecular morphological transition that enables spatial control over the self-assembly of fibres.

Introduction

Molecular self-assembly is an indispensable strategy for constructing a wide variety of supramolecular architectures from the bottom up^1,2. In recent decades, chemists have developed numerous sophisticated synthetic supramolecular systems^3–10. However, synthetic self-assembly still heavily relies on the formation of equilibrium structures under thermodynamic control, and it has not yet fully exploited nonequilibrium processes. Consequently, these abiotic approaches still fall far short of the precision and complexity realised in living organisms^11–13.

In artificial systems, the coupling of reactions and molecular diffusion is considered the key to the bottom-up fabrication of hierarchical self-organised structures spanning length scales from nano- to macroscopic^14–16. Molecular diffusion generates spatiotemporal concentration gradients, which can be fixed through chemical processes such as crystallisation or supramolecular assembly, leading to static spatial patterns. Such nonequilibrium-driven self-organisation has been extensively investigated in inorganic materials, including chemical crystalline gardens, biomorphs, and materials undergoing periodic precipitation¹⁷. Periodic precipitation, known as the Liesegang phenomenon, produces macroscopic, regularly spaced precipitate structures depending on the delicate balance between the nonlinear precipitation kinetics and reactant diffusion (Fig. 1a). More recently, reaction–diffusion systems have been applied to synthetic organic materials, including supramolecular polymers and fibres comprising self-assembling low-molecular-weight (LMW) building blocks^18–34. For example, nonlinear spatial patterns similar to Liesegang ring have been achieved by supramolecular fibres^21,30. However, these approaches have so far been limited to the spatial control of a single type of supramolecular assembly. Bottom-up strategies, which enable the positional formation of multiple supramolecular assemblies into macroscopic, spatially segregated architectures, remain largely undeveloped.

Fig. 1. Concept of a supramolecular morphological transition.

a Schematic illustrations of the Liesegang phenomenon, synthetic periodic precipitation patterns. b Supramolecular morphological transition: formation of a macroscopic concentric structure composed of distinct types of supramolecular fibres, which are generated through nonequilibrium break-and-build dynamics. c Chemical structures of fluorescent probes for (left) Ox fibres and (right) DTAC.

Here, we demonstrate a supramolecular morphological transition, which involves the formation of spatial patterns under nonequilibrium conditions and allows for the hierarchical organisation of multiple types of supramolecular (co)assembled fibres across length scales ranging from micrometres to centimetres (Fig. 1b). In this study, cationic surfactants were allowed to diffuse into preformed (progenitor) anionic supramolecular fibres, which comprised self-assembling LMW peptides. This induces the decomposition of fibres and the formation of coassembled fibres. Coupling this break-and-build cycle with surfactant diffusion enables positional formation of multiple types of supramolecular fibres, resulting in hierarchically organised, spatially segregated structures. Furthermore, the spatial patterns can be modulated by applying transient thermal pulses or by inducing collisions between distinct surfactant diffusion fronts. Thus, this study provides a framework for supramolecular structural engineering that can guide the organisation of functional complex materials.

Results

Discovery of nonequilibrium break-and-build dynamics

We recently reported the emergence of dynamic instability by hybridising a self-assembling peptide with surfactant micelles (sodium dodecyl sulfate (SDS))³⁵. In the system, the peptide nanofibres showed abrupt switching between growth and shrinkage phases, and the growth and shrinkage autonomously repeated until reaching thermodynamic equilibrium. This research initially aimed to investigate nonequilibrium behaviours upon hybridising the peptide with other types of surfactant molecules. The self-assembling LMW peptide (Ox) comprises a hydrophobic diphenylalanine derivative with a benzyloxime group and a carboxylate group at the N- and C-terminus, respectively (Fig. 1b)³⁶. The anionic Ox peptide self-assembles into a one-dimensional nanofibrous structure mainly through hydrogen bonding and π–π interaction. We discovered nonequilibrium break-and-build dynamics upon addition of a cationic surfactant, dodecyltrimethylammonium chloride (DTAC, critical micelle concentration (CMC) = 9.8 mM; Fig. 1b and Supplementary Fig. 1).

To monitor the nonequilibrium dynamics in real time, time-lapse confocal laser scanning microscopy (CLSM) imaging was performed. We used fluorescent probes, NP-Alexa647³⁷ and NBD-TA (NBD: nitrobenzoxadiazole), to observe the dynamics of the Ox fibres and DTAC, respectively (Fig. 1c, see Supplementary Fig. 2 for the fluorescence spectrum of NBD-TA). These probes contain Alexa Fluor 647 and NBD dyes modified with hydrophobic self-assembling moieties similar to the Ox peptide and DTAC. We prepared an Ox hydrogel by mixing a peptide precursor with a benzaldehyde group (Ald) and O-benzylhydroxylamine in the presence of NP-Alexa647 (Fig. 2a). At the top of the Ox hydrogel, a mixed solution of DTAC and NBD-TA was carefully added, and the bottom of the hydrogel was monitored by CLSM. The Alexa Fluor 647 channel showed that many short and thin nanofibres were well entangled before the addition of DTAC/NBD-TA (Fig. 2b). Eight minutes after adding DTAC/NBD-TA, the Ox fibres had disappeared. Thin fibres stained with both NP-Alexa647 and NBD-TA then formed, suggesting that these fibres consisted of both Ox peptide and DTAC (hereafter called thin coassembled fibres). Subsequently, 60 min after adding DTAC/NBD-TA, these thin coassembled fibres had decomposed into spherical aggregates. Finally, 36 h after adding DTAC/NBD-TA, the self-assembled fibrous structures had formed again. These fibres were coassembled fibres that were longer and thicker than the thin coassembled fibres and were observable even in the bright field (BF) image of wide-field (WF) microscopy (hereafter called thick coassembled fibres). These time-lapse CLSM images showed that hybridisation of anionic Ox fibres and cationic DTAC induces multistep nonequilibrium break-and-build dynamics: (1) decomposition of the initial Ox fibres, (2) formation of thin coassembled fibres, (3) decomposition of the thin coassembled fibres and concurrent formation of spherical aggregates, and (4) formation of thick coassembled fibres.

Fig. 2. Nonequilibrium break-and-build dynamics induced by DTAC addition.

a Schematic illustration of an experimental setup for real-time CLSM observation. b, c Time-lapse CLSM images after addition of DTAC to Ox fibres. The images acquired by NP-Alexa647 (magenta) and NBD-TA (green) channels are overlaid. The rightmost images show BF images in the WF mode. d ¹H NMR spectrum of thick coassembled fibres isolated by centrifugation (500 MHz, 5:2 DMSO-d₆/D₂O, room temperature). Signals of h and l were overlapped with water peaks. IS internal standard. *: signals for MES. ¹H NMR spectra of thin coassembled fibres are shown in Supplementary Fig. 8. e Phase diagram of supramolecular fibres plotted as concentrations of Ox and DTAC at the thermodynamic equilibrium. Different points indicate the dominant self-assembled state under given conditions. Red: Ox fibres, yellow: thin coassembled fibres, grey: thick coassembled fibres, green: fibre decomposition. f A mechanism of break-and-build cycles from Ox fibres to thin and thick coassembled fibres, followed by fibre decomposition. Red, yellow, grey, and green circles correspond to the respective regions in the phase diagram shown in (e). Condition: [Ald] = 3.6 mM, [O-benzylhydroxylamine] = 3.6 mM, [DTAC] = 5.0 (for b, d), 36 mM (for c), 50 mM MES, pH 7.0, 30 °C.

The concentration dependence of Ox and DTAC revealed that their concentrations determine the final stage at which the multistep dynamic process stops under equilibrium conditions (Fig. 2c and Supplementary Figs. 3–6; see Supplementary Fig. 7 for Ox formation at different concentrations). At low concentrations of both Ox and DTAC (1.8 mM and 1 mM, respectively), the initial Ox fibres had decomposed after 8 min, and thin coassembled fibres appeared at 20 min (Supplementary Fig. 4a). No significant changes were then observed. In the case of high DTAC concentration (36 mM), every step proceeded faster than for low DTAC concentration (5.0 mM, Fig. 2b). The thick coassembled fibres also degraded under this condition, resulting in no fibres being detected in either the fluorescence image or BF image after 60 min (Fig. 2c). This observation suggests that an excess amount of DTAC micelles decomposes the thick coassembled fibres, probably through electrostatic and hydrophobic interaction with the inner hydrophobic environment. At low and high concentrations of DTAC and Ox (1.0 mM and 7.2 mM, respectively), the morphology of the Ox fibres remained intact, but NBD-TA stained the Ox fibres (Supplementary Fig. 6a). This suggests that NBD-TA (and DTAC) interacted with the Ox fibres through electrostatic interaction, but the molar amount of DTAC was insufficient to decompose the Ox fibres. ¹H NMR measurement after fibre isolation by centrifugation revealed that the thin and thick coassembled fibres were indeed composed of both Ox peptide and DTAC (Fig. 2d and Supplementary Fig. 8). Therefore, the concentrations of Ox and DTAC and the ratio of Ox to DTAC determine the nonequilibrium dynamics, as shown in a phase diagram (Fig. 2e).

Considering these experimental results, we propose a mechanism for the nonequilibrium break-and-build dynamics in Fig. 2f. After DTAC addition at the top of the Ox hydrogel, the concentration of DTAC at the bottom (i.e., the focal plane of CLSM imaging) gradually increases with time. At low DTAC concentration, the DTAC monomers can interact with the initial Ox fibres through electrostatic interactions and hydrophobic effects, leading to the decomposition of the Ox fibres, followed by the formation of the coassembled fibres. As the DTAC concentration increases, the coassembled fibres decompose into spherical aggregates, from which new fibres are generated again. The DTAC concentration then exceeds its CMC, and the DTAC micelles degrade the coassembled fibres, probably because of the incorporation of Ox peptides in the hydrophobic interior. Because each step of the dynamic process could be observed even at high DTAC concentration, the kinetics of fibre decomposition and formation may be comparable with the diffusion rate of DTAC. This behaviour differs from the dynamic instability induced by anionic SDS micelles, which is characterised by stochastic fluctuations in fibre growth and shrinkage³⁵. The dynamics observed here fundamentally manifests as a multistep structural transition, albeit with the potential for transient, localised dynamic instability-like behaviours.

Spatial pattern formation in two dimensions

With the unique break-and-build dynamics in hand, we next sought to apply the break-and-build dynamics to pattern formation in two dimensions. The experimental setup was inspired by the classical Liesegang ring formation. To suppress thermal convection, a composite hydrogel comprising agarose and Ox was used. It has been confirmed that the agarose network has a negligible effect on the formation of peptide fibres^32,38. We prepared a thin cylindrical composite hydrogel of Ox peptide, agarose, and NP-Alexa647 with a small central hole (gel diameter: 14 mm, hole diameter: 2.5 mm, gel thickness: 1.4 mm) (Fig. 3a). A mixed solution of DTAC and NBD-TA was added into this hole. DTAC and NBD-TA subsequently diffused outwards and interacted with the Ox fibres. After incubation at 30 °C for 18 h, the initially transparent and homogeneous hydrogel transformed into a heterogeneous structure comprising four concentric ring patterns with varying opacity (Fig. 3b, c). Line plot analysis of the patterned hydrogel showed sharp transitions between the 1st, 2nd, 3rd, and 4th rings (numbered from centre outwards) (Fig. 3d). The resultant concentric patterns showed negligible changes thereafter, implying that they can be regarded as thermodynamically stable or metastable states. The emergence of such abrupt, discontinuous spatial transitions suggests an underlying nonlinear self-organisation process, analogous to Liesegang ring formation.

Fig. 3. Supramolecular morphological transition.

a Schematic illustration of an experimental setup for supramolecular morphological transition. b, c Photos of composite hydrogels of Ox peptide and agarose b before and c 18 h after addition of DTAC. d Line plot analysis of brightness at a yellow rectangle region shown in (c). Red arrowheads represent interfaces between each ring. a.u. arbitrary units. e BF image of the patterned hydrogels in the WF mode. The brightness of the BF images is inverted for clarity. f CLSM image of the patterned hydrogel. The images acquired by (magenta) NP-Alexa647 and (green) NBD-TA channels are overlaid. g Magnified BF images of the 1st, 2nd, 3rd, and 4th rings at rectangle regions shown in (e). The brightness of the BF images is inverted for clarity. h CLSM images of the 1st, 2nd, 3rd, and 4th rings. i Violin and box plots of fibre widths. The white dots represent the median, and the boxes indicate the interquartile range. Statistical significance was evaluated by one-way ANOVA followed by Dunnett’s post hoc test (***p < 0.001; ns, not significant, n = 20). j, k Concentration of j Ox peptide and k DTAC at the (black) 1st, (dark grey) inner, (light grey) outer 2nd, (yellow) 3rd, and (red) 4th rings determined by ¹H NMR measurements. The data represent the mean ± SD (n = 4). l Scatter plot of the observed fibre morphology as a function of the concentrations of Ox and DTAC, overlaid with the phase diagram shown in Fig. 2e. The data points confirm that the experimentally observed morphologies are consistent with the thermodynamically stable states identified in the phase diagram. Condition: [Ald] = 3.9 mM, [O-benzylhydroxylamine] = 3.9 mM, [agarose] = 0.25 wt%, [DTAC] = 3.4 mM (if homogenously distributed), volume of the hydrogel and DTAC solution: 220 µL and 2.5 µL (300 mM), respectively, 50 mM MES, pH 7.0, 30 °C.

We then performed multiscale structural analysis of the patterned hydrogel, ranging from microscale (µm) to macroscale (mm/cm), by WF microscopy and CLSM. The macroscopic BF and CLSM images clearly showed that the fibre morphology, density, and spatial distribution varied across the concentric rings (Fig. 3e, f). BF imaging revealed that the 1st and 2nd rings mainly consisted of densely packed, well-entangled thick fibres, which appeared to be randomly oriented (Fig. 3g). The 3rd ring consisted of thick fibres that were slightly less dense than those in the inner rings, but they aligned along the direction of the DTAC diffusion. In sharp contrast, such thick fibres were rarely observed in the 4th ring. These observations corresponded well with the gel opacity. Fluorescently-stained coassembled fibres were mainly located at the 3rd and 4th rings (Fig. 3f). Higher-magnification CLSM images showed that many of the short and thin fibres were well entangled (Fig. 3h). Fluorescently-stained coassembled fibres indeed formed even in the 1st and 2nd rings, whose widths were larger than those of the 3rd and 4th rings (0.43 ± 0.06, 0.44 ± 0.06, 0.32 ± 0.04, and 0.36 ± 0.02 µm for 1st, 2nd, 3rd, and 4th rings, respectively) (Fig. 3i).

To further characterise the fibres, we determined the concentrations of Ox and DTAC by ¹H NMR spectroscopy. The patterned hydrogel was cut into individual rings, with the 2nd ring further divided into the inner and outer regions, and each gel piece was dissolved in dimethyl sulfoxide-d₆ for analysis (Supplementary Figs. 9, 10). The concentrations of Ox peptide and DTAC varied across the rings, as shown in Fig. 3j and k. Compared with the initial Ox concentration of 3.9 mM, the Ox concentration was significantly elevated in the 1st and inner 2nd rings (7.0 ± 0.8 and 4.7 ± 0.4 mM, respectively), similar in the outer 2nd and 3rd rings (3.8 ± 0.4 and 3.83 ± 0.17 mM, respectively), and slightly lower in the 4th ring (3.4 ± 0.4 mM). The ¹H NMR spectra revealed a monotonic decrease in the DTAC concentration from the 1st ring to the 4th ring (11.4 ± 1.6 mM to 1.9 ± 0.3 mM). We then plotted the concentration data for Ox and DTAC in the phase diagram shown in Fig. 2e (Fig. 3l). The concentrations of Ox and DTAC in the 1st and 2nd rings fell within the region corresponding to the formation of the thick coassembled fibres. Those in the 3rd ring lay at the boundary between the thick and thin coassembled fibre regions, whilst the concentrations in the 4th ring fell within the region of the thin coassembled fibres. This characterisation is consistent with the microscopic observations. Overall, it is reasonable to conclude that the concentric ring patterns emerged from spatially distinct fibre formation regimes, presumably resulting from multistep nonequilibrium dynamics.

To assess the contribution of multistep dynamics to the observed gel pattern formation, we added the anionic surfactant SDS, instead of cationic DTAC, to the composite hydrogel. As previously reported, hybridisation of Ox fibres with SDS induces distinct nonequilibrium dynamics, autonomous repetition of fibre growth and shrinkage. However, in this case, the hydrogel remained homogeneous even 18 h after SDS addition (Supplementary Fig. 11). These findings strongly suggest that the nonequilibrium break-and-build dynamics triggered by cationic DTAC is crucial for the emergence of the gel pattern.

Repeated propagation waves generate macroscopic concentric patterns

Visualisation of the formation process of the concentric ring pattern revealed the generation of repeated propagation waves during multiple ring formation. Real-time microscopic observations using the BF, Alexa647, and NBD channels revealed that the concentric rings appeared immediately after the addition of DTAC and propagated outwards along the direction of DTAC diffusion (Fig. 4a and Supplementary Movie 1). For example, 30 min after the addition of DTAC (green line in Fig. 4b–d), four concentric rings with distinct fluorescence intensities were clearly observed in both the Alexa647 and NBD channels. The innermost ring, approximately 250 µm wide, exhibited the highest intensities across all three channels, and it could be assigned to the 1st ring. The 2nd ring, located just outside the 1st ring, showed gradual decreases in the Alexa647 and NBD fluorescence intensities towards its outer edge, whilst the BF intensity remained relatively constant (approximately 600 µm wide), corresponding well with the 2nd ring identified in Fig. 3. Beyond this region, sharp increases in the fluorescence intensities in the Alexa647 and NBD channels were observed, accompanied by a gradual decrease in the BF intensity. This region, approximately 300 µm wide, was assigned to the 3rd and 4th rings. Interestingly, in the Alexa647 channel, a dark region was detected outside the 3rd and 4th rings, where the fluorescence intensity dropped below that of the initial Ox hydrogel (green open triangle in Fig. 4c). This may reflect local decomposition of the Ox fibres. As shown in the line plot analyses (0.5–2 h in Fig. 4b–d) and the time course tracking of the ring position (Fig. 4e), all the rings gradually propagated outwards over time. By 4 h, the propagation of the 1st, 2nd, and 3rd rings had almost stopped, whilst the outer edge of the 4th ring and the adjacent dark region continued to slowly expand, probably reflecting the ongoing growth of the 4th ring. After 12 h, a sharp boundary between the 2nd and 3rd rings became clearly defined, characterised by a decrease in the BF intensity and a concurrent increase in the fluorescence in the Alexa647 and NBD channels. Notably, the fluorescence intensity of NBD-TA in the central hole gradually decreased over 12 h, indicating that NBD-TA/DTAC steadily diffused outwards from the hole with time.

Fig. 4. Repeated propagation waves produce macroscopic nonlinear concentric patterns.

a Time-lapse wide-field microscopic images after addition of DTAC to the composite hydrogel. (Left) BF images and (right) overlay fluorescent images acquired by (magenta) NP-Alexa647 and (green) NBD-TA channels. b–d Time-dependent line plot analysis of BF and fluorescence intensities along a yellow rectangle indicated in (a). Plots correspond to b BF, c NP-Alexa647, and d NBD-TA images, showing the time evolution of signal intensities along the selected region. Filled triangles indicate the maxima of the 3rd ring, and open triangles represent the minima of dark regions attributed to fibre disassembly. *: Signal intensity mismatch due to image tiling. a.u. arbitrary units. e Time course of the position of the concentric interfaces. f A mechanism of the supramolecular morphological transition. Condition: [Ald] = 3.9 mM, [O-benzylhydroxylamine] = 3.9 mM, [agarose] = 0.25 wt%, [DTAC] = 0.91 mM (if homogenously distributed), volume of the hydrogel and DTAC solution: 818 µL and 2.5 µL (300 mM), respectively, 50 mM MES, pH 7.0, 30 °C.

Using the spatiotemporal imaging and structural analysis described above, we propose a mechanism for the emergence of the two-dimensional concentric patterns in Fig. 4f. Upon addition of DTAC into the central hole, the DTAC molecules/micelles begin to diffuse outwards, generating a concentration gradient across the composite hydrogel. At low DTAC concentrations near the diffusion front, the progenitor Ox fibres begin to disassemble upon interaction with DTAC, followed by the formation of differentiated thin and thick coassembled fibres. These events are likely responsible for the emergence of the dark zone and the outer rings (3rd and 4th). As the DTAC concentration further increases, the thin coassembled fibres become unstable and decompose, whilst the thick coassembled fibres further grow, corresponding to the innermost rings (1st and 2nd). With continued DTAC diffusion, these distinct fibre populations propagate outwards along the concentration gradient. Once the system approaches equilibrium, the propagation of the ring structures slows and eventually stops. Overall, the spatial patterning observed in this system can be ascribed to the interplay between the DTAC diffusion-driven concentration gradient and the nonequilibrium break-and-build dynamics comprising supramolecular fibre formation and decomposition. Notably, this mechanism is fundamentally different from that of the classical Liesegang phenomenon, in which periodic band formation is governed by nucleation and precipitation, rather than multistep supramolecular assembly and disassembly.

Modulation of the self-organised ring patterns via transient low-temperature pulses

Given that the formation of the current ring patterns is driven by the diffusion-controlled propagation wave, we sought to modulate the pattern by external factors, such as the temperature pulse and the interference of the two distinct propagation waves. Brief low-temperature pulses (4 °C for 1 min) were applied 30 min, 1 h, and 2 h after the addition of DTAC, corresponding to when the leading edge (the dark region) of the propagating wave reached the positions of the 2nd, 3rd and 4th rings at 18 h, respectively (Fig. 5a). The macroscopic and BF images revealed the emergence of new dark rings located near the centres of the 2nd and 3rd rings (Fig. 5b–d and Supplementary Fig. 12). Conversely, the fluorescence signals in the Alexa647 and NBD channels remained largely unchanged, indicating that temperature perturbation selectively affects the formation of thick coassembled fibres rather than thin coassembled fibres. Given the outward propagation of the ring patterns (Fig. 4e), the positions of the new dark rings likely correspond to the location of the outer edge of the 4th ring at the time of each pulse. Although three temperature pulses were applied, only two new dark rings were observed. This is probably because, by the time of the third pulse (at 2 h), the outer edge of the 4th ring had already extended beyond the position of the 3rd ring. Because thick coassembled fibres do not form in the 4th ring region, the third pulse had no observable effect. Overall, brief low-temperature pulses can preferentially modulate the spatial patterning of the thick coassembled fibres during the self-organising process.

Fig. 5. Modulation of the supramolecular morphological transition by low-temperature pulses.

a Experimental protocol for low-temperature pulses. b Photograph of the patterned composite hydrogel. Red triangles indicate low temperature-induced new dark rings. c Wide-field microscopic images acquired by (top) BF, (middle) NBD-TA, and (bottom) Alexa647 channels. Red triangles indicate low temperature-induced new dark rings. d Line plot analysis of BF and fluorescence intensities along a yellow rectangle indicated in (c). Plots correspond to (top) BF, (middle) NBD-TA, and (bottom) Alexa647 images. The comparison of line plot analysis is shown in Supplementary Fig. 12 to clarify the impact of the low temperature pulse. a.u. arbitrary units. Condition: [Ald] = 3.9 mM, [O-benzylhydroxylamine] = 3.9 mM, [agarose] = 0.25 wt%, [DTAC] = 3.4 mM (if homogenously distributed), volume of the hydrogel and DTAC solution: 220 µL and 2.5 µL (300 mM), respectively, 50 mM MES, pH 7.0, 30 °C.

Complex pattern formation by collision of propagating waves induced by distinct surfactants

We finally investigated the pattern formation when two distinct propagating waves were generated by adding different types of cationic surfactants, DTAC and trimethyltetradecylammonium chloride (TTAC), into separate holes (Fig. 6a, CMC of TTAC = 0.98 mM, see Supplementary Fig. 1). First, we investigated the nonequilibrium dynamics and spatial pattern formation induced by TTAC. Microscopic images revealed that TTAC induced similar break-and-build dynamics to DTAC and that concentric patterns consisting of multiple rings formed upon TTAC addition (Supplementary Figs. 13–15). However, the resultant pattern differed from that observed for DTAC. In the case of TTAC, three distinct rings were observed: the innermost and 2nd rings mainly consisted of thick coassembled fibres, whereas such fibres were not detected in the 3rd ring, as judged from the BF image. Therefore, the 1st, 2nd, and 3rd rings that formed upon TTAC addition are considered to correspond to the 1st, 2nd, and 4th rings observed upon DTAC addition. Next, we simultaneously added DTAC and TTAC into two holes spaced 6.2 mm apart (Fig. 6a). After incubation for 18 h, a highly-opaque region formed in the middle area between the two holes (Fig. 6b–f). A time-lapse microscopic movie revealed that this new opaque ring region emerged when the 3rd ring of DTAC and the 2nd ring of TTAC collided (Fig. 6g and Supplementary Movie 2). Quantitative analysis indicated that the initial propagation velocities were 110 µm/min for DTAC and 71 µm/min for TTAC, demonstrating that DTAC exhibited a higher propagation velocity than TTAC (Supplementary Fig. 16). This difference caused the collision to occur closer to the TTAC injection point (with distances of 4611 and 3621 µm from the respective injection points). Furthermore, we observed that the collision patterns strongly depended on the distance between the two holes. When the holes were spaced 2.6 mm apart, the mixed opaque regions were produced at both the top and bottom of the region between the holes (Supplementary Fig. 17). In contrast, no collision was observed at a distance of 8.2 mm (Supplementary Fig. 18). We also confirmed that this opaque ring did not form when adding only DTAC to the two holes; instead, the rings simply fused (Supplementary Fig. 19). These results demonstrate that the pattern formation is strongly affected by the type of cationic surfactant, probably the hydrophobicity and self-assembly properties (e.g., CMC) of the surfactant, and the resulting pattern can be further modulated by the interactions between propagating waves generated by surfactant diffusion.

Fig. 6. Morphology control by diffusing two distinct types of surfactants.

a Experimental setup: addition of two distinct types of cationic surfactants, DTAC and TTAC, into different holes. b Photographs of the patterned composite hydrogel 18 h after adding DTAC and TTAC. c Line plot analysis of brightness along a yellow rectangle shown in (b). Green and red arrows and letters represent the patterns induced by DTAC and TTAC, respectively. The yellow arrow and letters indicate a newly-formed band by mixing DTAC and TTAC. a.u. arbitrary units. d BF image of the patterned composite hydrogel. e, f Magnified BF images at square and rectangle regions shown in (d). g Time-lapse wide-field microscopic images after addition of DTAC and TTAC to the composite hydrogel. (Left) BF images and (right) fluorescent images acquired by the NP-Alexa647 channel. Condition: [Ald] = 3.9 mM, [O-benzylhydroxylamine] = 3.9 mM, [agarose] = 0.25 wt%, [DTAC] = [TTAC] = 0.91 mM (if homogenously distributed), volume of the hydrogel and DTAC solution: 818 µL and 2.5 µL (300 mM for both DTAC and TTAC), respectively, 50 mM MES, pH 7.0, 30 °C.

Discussion

We have demonstrated the macroscopic spatial arrangement of distinct types of supramolecular fibres within a hydrogel through a bottom-up approach. By coupling nonequilibrium cycles of fibre decomposition and reformation with the diffusion of surfactants, self-propagating waves repeatedly emerged to generate nonlinear concentric patterns. This system provides a strategy for spatially patterning artificial soft materials where progenitor peptide fibres differentiate into coassembled fibres in response to a surfactant gradient, and the resulting patterns can be modulated by using multiple surfactants. The central challenge in current supramolecular chemistry is to orchestrate the organisation of multiple types of molecular assemblies within a single system^{29,36,37,39–53}. While the macroscopic spatial control of a single type of supramolecular assembly has been extensively investigated^20,42,54,55, the orchestrated patterning of multicomponent systems rarely extends beyond tens of micrometres⁵⁶. The spatial organisation of multiple distinct types of supramolecules at the macroscale from the bottom up remains a challenge, as demonstrated here (Supplementary Fig. 20). Our method bridges the gap between nanoscale design and macroscale function, providing a versatile platform for the development of complex, functional biomaterials with controlled spatial architectures.

Methods

General

Unless stated otherwise, all commercial reagents were used as received. Thin-layer chromatography (TLC) was performed on silica gel 60F₂₅₄ (Merck). ¹H and ¹³C NMR spectra were obtained on a JEOL JNM-ECZ500 and a JEOL JNM-ECZ400 with residual non-deuterated solvents (CD₃OD: 3.31 ppm, DMSO-d₆/D₂O: 2.5 ppm (DMSO-d₅) for ¹H, CD₃OD: 49.00 ppm for ¹³C) as the internal reference. ESI mass spectra were recorded using an Exactive (Thermo Fisher Scientific). Reversed-phase HPLC (RP-HPLC) was carried out on a Hitachi Chromaster system equipped with a diode array and YMC-Pack Triart C18 column. All runs used linear gradients of acetonitrile (CH₃CN) containing 0.1% trifluoroacetic acid (TFA) and 0.1% aqueous TFA. The UV-vis absorption measurement was conducted with Shimadzu UV-2600 with a temperature-controlled cell holder. Fluorescent measurement was performed on a PerkinElmer LS55 fluorescent/phosphorescent spectrophotometer. The images of confocal laser scanning microscopy (CLSM) were acquired by an LSM 900 equipped with an Airyscan 2 unit (Carl Zeiss). Plan-Apochromat objectives (×63, 1.40 numerical aperture, oil immersion, Carl Zeiss), Plan-Apochromat objectives (×20, 0.8 numerical aperture, Carl Zeiss), and EC Plan-Neofluar objectives (×5, 0.16 numerical aperture, Carl Zeiss) were used for CLSM and wide-field (WF) imaging. Definite Focus 2 (Carl Zeiss) was employed to compensate for the sample drift. The sample temperature was controlled at 30 °C with a stage-top incubator (STXG-WSBX-SET, Tokai Hit). The analysis of the microscopic images and photographs was performed by Zen 3.4 (blue edition, Carl Zeiss) and Fiji. Photographs were taken by iPhone 12 mini (Apple).

Real-time CLSM observation of nonequilibrium dynamics

A suspension of Ald (1.2-fold higher than the desired final concentration) and NP-Alexa647 (4.0 µM) in 50 mM MES (pH 7.0, containing 1 vol% DMSO) was heated with a heating gun (PJ-206A1, Ishizaki) until fully dissolved. After cooling at room temperature (sample temperature: ca. 50 °C), the resultant mixture (100 µL) was transferred to a multiwell glass-bottom dish (Matsunami Glass Ind. Ltd., D141400) and maintained at 45 °C. O-Benzylhydroxylamine (43 mM, 10 µL in 100 mM MES, pH 7.0) was added and thoroughly mixed at 45 °C. The mixture was incubated at 30 °C for 18 h with water placed around the wells to prevent drying. Subsequently, a solution of DTAC and NBD-TA (10 µL in 50 mM MES, pH 7.0; 12-fold higher than the desired final concentration) was carefully added onto the top of the thin cylindrical hydrogel, and CLSM imaging was immediately performed using a ×63 objective lens. After time-lapse imaging, WF images were acquired with a ×20 objective lens. Detailed assay conditions were described in the figure captions.

Concentric patterning in a composite hydrogel

A suspension of Ald (4.3 mM) and NP-Alexa647 (4.0 µM) in 50 mM MES buffer (pH 7.0, containing 1 vol% DMSO) was heated with a heating gun (PJ-206A1, Ishizaki) until completely dissolved. After cooling to room temperature, the solution was mixed with agarose LE (0.25 wt%) and was reheated with a heating gun until fully dissolved. The resulting solution (200 µL) was transferred to a glass-bottom dish (Matsunami Glass Ind. Ltd., D11530H, diameter: 14 mm, height: 1.4 mm) and maintained at 45 °C. A solution of O-benzylhydroxylamine (43 mM, 20 µL in 50 mM MES, pH 7.0) was added and thoroughly mixed at 45 °C. A plastic cylinder (diameter: 2.5 mm) was then placed at the centre of the dish. The mixture was incubated at room temperature until gelation, followed by further incubation at 30 °C for more than 18 h. The plastic cylinder was removed, and any solution remaining in the hole and on the gel surface was carefully removed. A solution of DTAC and NBD-TA (300 mM, 1.2 mM, 2.5 µL in 50 mM MES pH 7.0) was added into the hole, and the hydrogel was incubated at 30 °C for 18 h. The patterned hydrogel was subsequently analysed by CLSM and WF microscopy.

Real-time imaging of the formation process of the concentric patterns

The hydrogel was prepared according to the protocol similar to that described in “Concentric patterning in a composite hydrogel”, except that a larger glass-bottom dish (Matsunami Glass Ind., Ltd., D910400, diameter: 27 mm, height: 1.4 mm) was used. Briefly, a mixture of Ald, NP-Alexa647, agarose-LE (744 µL) was transferred into a glass-bottom dish and maintained at 45 °C. O-Benzylhydroxylamine (43 mM, 74.4 µL in 100 mM MES, pH 7.0) was added and thoroughly mixed at 45 °C. After incubation at 30 °C for more than 18 h with a plastic cylinder in place, the cylinder was removed, and a solution of DTAC (300 mM) and NBD-TA (1.2 mM) in 50 mM MES (pH 7.0, 2.5 µL) was added into the hole. and time-lapse imaging was immediately started thereafter.

Author contributions

R.K. and I.H. conceived the project. S.T. discovered the nonequilibrium dynamics and spatial patterning. Y.I. conducted all of the experiments, except for the propagating wave collision experiment of DTAC and TTAC at a distance of 2.5 mm, which was performed by N.K. R.K. and I.H. wrote the manuscript with contributions from all co-authors.

Peer review

Peer review information

Nature Communications thanks Rienk Eelkema and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

I.H. discloses support for the research of this work from the Japan Science and Technology Agency (JST) ERATO Grant Number JPMJER1802, a Grant-in-Aid for Specially Promoted Research (JSPS KAKENHI Grant JP23H05405). R.K. discloses support for the research of this work from a Grant-in-Aid for Scientific Research (B) (JSPS KAKENHI Grant JP22H02195, JP26K01632) and JST FOREST (Grant Number JPMJFR2328). I.Y., S.T. and N.K. declare no relevant funding.

Data availability

Additional data to support the findings of this study are provided in the Supplementary Information. All data are available from the corresponding author upon request. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-76336-3.