Introduction

Biology utilizes phase separation to solve the complexity of different coexisting biological processes by creating different compartments i.e. intracellular organelles1,2. Organelles can either be membrane-bound allowing limited permeability or can be membraneless. These organelles can exhibit selective partitioning and concentrate biomolecules within their dense phase. Although imagining the coexistence of different membrane-bound organelles is straightforward, but how different membraneless organelles exist without mixing its components in the cytosol is deceptive. In other words, these active membraneless entities demix from cytosol, disfavor thermodynamic dissolution, providing spatial stability and thus sustain away from equilibrium3,4. The formation and growth of these membraneless organelles, which play vital roles in extant biochemistry, are controlled by a network of catalytic biochemical reactions5,6. Intracellular membraneless organelles are accompanied by several specific ribonucleoprotein functions and enzymatic activities which are the result of evolutionary selection and adaptation7. Arguments are made for their possible role on early Earth acting as a reaction hub and as the minimalistic form of compartment1,2.

The importance of non-equilibrium behavior of membraneless organelles can be seen in different biological phenomena like mitosis, oocyte to embryo transformation, etc8,9,10. These biochemical processes harness energy from molecules of high chemical potential to sustain the organelles in the activated state, subsequently converting them into low chemical potential molecules11,12,13,14,15. Although several reports demonstrate coacervate lifetime modulation by utilizing different stimuli such as pH, light, temperature, and so forth16,17,18,19,20,21,22,23,24, there has been a recent interest in designing active coacervates utilizing (bio)chemical reactions where coacervation is achieved with the help of polyelectrolytes25,26,27,28,29,30,31,32. However, small molecules based liquid-liquid phase separation (LLPS) that can operate non-equilibrium, facilitated by their native catalytic potential, have not yet been reported33,34,35,36,37,38,39,40. It would be exciting in the context of chemical evolution if low molecular weight coacervates could utilize their native catalytic potential to achieve spatial control as a function of time. Herein, we report a two-component phase separation model triggered by a dynamic covalent bond that results in the formation of coacervate droplets. Interestingly, the coacervate phase harnesses intrinsic (covalent) catalysis and provides negative feedback to achieve active coacervation. These coacervate droplets behave as a chiral microenvironment that was not accessed by the building blocks individually. Further, the catalytic non-equilibrium coacervate droplets show selectivity for one enantiomer over the other.

Results

Chiral coacervate from non-equilibrium demixing

To access the phase separation, we started with a diphenylalanine core tetrapeptide (NH2-HFFP-CONH2, 1, Fig. 1a) with free amine and histidine at the N-terminal to bind a thermodynamically activated substrate A (4-((benzyloxy)carbonyl)−1-(4-formylbenzyl)pyridin-1-ium, a cationic aldehyde, Fig. 1a) via dynamic Schiff base. The pyridinium ion in A was strategically selected to enhance the water solubility of the ester and to engage in supramolecular interactions that could promote coacervation. Further, histidine residues are known to be catalytically proficient towards hydrolysis of ester bonds, particularly in the self-assembled state41,42,43,44. The C-terminal of the peptide was installed with proline, which can resist amyloid fibril formation and introduce disorderness by providing less degree of freedom in ϕ-ψ space45. Upon mixing 1 (40 mM) with A (10 mM) in HEPES buffer (pH 8, 80 mM), the solution turned turbid within ca. 2 min (vial image, Fig. 1a and Supplementary Fig. 1, individual components were soluble). Brightfield optical microscopy of the turbid solution was performed to probe the morphology of the structures. Micron-sized droplets were observed, which showed coalescence that underpinned their liquid-like nature (Fig. 2a). As coacervates are known to sequester a variety of molecules, confocal laser scanning microscopy (CLSM) was performed using different dyes16,17,18,30,31,32,33,34,35. High fluorescence intensities of various dyes inside the droplets demonstrated that the coacervate droplets can indeed host both hydrophobic and hydrophilic molecules (Supplementary Fig. 2). Further, fluorescence recovery after photobleaching (FRAP) experiment was performed to verify the liquid-like nature of the droplets (Fig. 2b). Indeed, the fluorescence recovery within a minute demonstrated the liquid-like nature of the droplets on phase separation (Fig. 2c)30,31,32,33,34,35,46,47,48. The water content of the 1 + A coacervate droplets was determined from thermogravimetric analysis carried out on the coacervate pellet obtained from centrifugation (see Supplementary Information). A gradual loss of 82% mass over a temperature range of 30–110 °C was observed, which represents the water present in the coacervate pellet (Supplementary Fig. 3)49. High performance liquid chromatography (HPLC) and high resolution mass spectrometry (HRMS) of the mixture of 1 and A showed Schiff base formation A1 (Figs. 1a and 2d, and Supplementary Fig. 4, quantification of A1 was not possible considering its reversibility in aqueous medium). To probe the role of imine, we synthesized a control substrate B (1-benzyl-4-((benzyloxy)carbonyl)pyridin-1-ium, where the aldehyde group was absent, Fig. 1b). Turbidity did not appear for the mixture of B and 1 (Fig. 2e and Supplementary Fig. 1), which suggested the role of imine-based building block to access phase separation50,51,52.

Fig. 1: Non-equilibrium demixing and dissolution of coacervate.
figure 1

a Scheme showing the formation of active coacervate via a dynamic covalent bond and its dissolution with time. b Chemical structure of peptide 1–3, and ester without aldehyde B.

Full size image
Fig. 2: Formation and characterization of chiral coacervate.
figure 2

a Brightfield microscopy images showing the fusion, b FRAP images (rhodamine B dye, 5 μM) (scale bar, 5 μm), and c corresponding intensity time trace of fluorescence recovery on photobleaching of 1 + A coacervate droplet (shaded gray region shows the standard deviation of triplicate experiments). d Representative HPLC chromatogram and HRMS of A1. e Vial image of the solution containing a mixture of 1 + B. f Phase diagram showing two (coacervate) and single phase formation (red cross, black, and hollow blue circles represent the measured points). g CD plot of 1, A, coacervate of 1 + A mixture, and 1 + B mixture. h Chemical structure of 2. i Inverted vial image of gel formed by 2 (5% DMSO/buffer), and corresponding j TEM micrograph (scale bar, 500 nm). [1], [2] = 40 mM, [A], [B] = 10 mM.

Full size image

The phase diagram showed that for 40 mM of 1, at least 8 mM of A was required for phase separation, whereas for 10 mM of 1, a minimum of 40 mM of A was needed (Fig. 2f) in HEPES buffer (pH 8, 80 mM). Next, to determine the interactions responsible for phase separation, the effect of various additives, i.e., urea, 1,6-hexanediol, and NaCl, were explored to disrupt hydrogen bonding, hydrophobic, and electrostatic interactions, respectively53,54,55. Notably, turbidity was not observed when the concentration of urea and 1,6-hexanediol exceeded 1 M, whereas NaCl did not affect the appearance of turbidity, which indicates that hydrogen bonding and hydrophobic effect were driving the phase separation (Supplementary Figs. 5–7).

To further characterize these coacervate droplets, circular dichroism (CD) spectroscopy was performed. Interestingly, a new intense CD band was observed when 1 and A were mixed (negative CD peak at 261 nm, Fig. 2g). This peak was absent in the case of the individual components. This suggested the emergence of a chiral environment in the presence of coacervate droplets56,57. Notably, this new peak was not observed in the CD spectrum for the mixture of B and 1 either (Fig. 2g). Next, the coacervate formation was monitored at different pHs to check its pH compatibility. It was found that coacervation was suppressed below pH 6 (Supplementary Fig. 8). Interestingly, no new peak was observed in CD spectroscopy at pH ca. 4, i.e., under the conditions where coacervate formation was suppressed (Supplementary Fig. 9). These results underpinned that the emergence of the chiral microenvironment was observed exclusively in the demixed state. Further, the role of proline in prohibiting the formation of solid-like aggregate was probed by mutating 1 at the C-terminal by alanine to give 2 (NH2-HFFA-CONH2, Fig. 2h). Peptide 2 (40 mM) was found to be insoluble under the same buffer conditions while in 5% DMSO/Buffer (80 mM HEPES, pH 8), 2 (40 mM) formed a self-supporting gel within 5 min (Fig. 2i). Transmission electron microscopy (TEM) images of gel sample revealed the formation of fibrous morphology (Fig. 2j). These results depicted the role of proline in providing a kinetically feasible pathway for phase separation by evading the generation of amyloid-like fibrils45.

Dissolution of chiral coacervate via intrinsic catalysis

Membraneless organelles of extant biochemistry are known to be regulated by a network of biochemical reactions and show active behavior. These condensates are maintained out of equilibrium by utilizing coupled biochemical transformations36,37,58,59,60. We asked whether coacervates formed by mixing 1 and A could exhibit such active behavior under non-equilibrium conditions. To monitor the changes in coacervate droplets’ population with time, CLSM and brightfield optical microscopy were performed. Indeed, time-lapse CLSM and brightfield microscopy images and movies showed the dissolution of coacervate over time (Fig. 3a, b, Supplementary Fig. 10, Supplementary Movies 1, 2). The violin plot showed a significant decline in the population of the coacervate droplets after 60 min (Fig. 3c and Supplementary Fig. 10). This observation is in contrast to what is expected from LLPS where coacervates undergo size enlargement via coalescence. Interestingly, CLSM images further revealed the generation of vacuoles inside the coacervate after ca. 60 min.; vacuolization was not seen earlier in coacervates from low molecular weight components (Fig. 3d–f, Supplementary Movies 1, 2, image at 5 h, Supplementary Fig. 11)61,62. We speculated that the vacuolization observed in 1 + A system might be due to the catalytic degradation of A in the coacervate phase assisted by imidazole present in the building block. Notably, when the hydrolyzed products A’, A” (Fig. 1a) were mixed together with peptide 1 under the same conditions, no imine formation and no phase separation was observed (Supplementary Fig. 12). Hence, when the activated ester moiety began to deplete due to hydrolysis, the interior of the coacervate likely started accumulating the hydrolyzed products (A’ and A”), which do not have the ability to phase separate, subsequently resulting in the formation of a dilute phase inside the coacervate. This dilute phase, generated due to the intrinsic hydrolytic activity of the coacervate or the peptide itself, appeared as vacuoles. Additionally, the hydrolyzed products (A’ and A”), which are not involved in coacervation, build osmotic pressure inside the droplet and drive water transport into it, leading to vacuole formation21,28. In combination, the results suggested that this is one of the first examples of vacuole formation, which was led via intrinsic catalytic roles of the coacervate phase. Interestingly, multiple vacuole formation and subsequent fusion within a single coacervate droplet could also be observed in CLSM and brightfield microscopy images and movies (Fig. 3g, h, Supplementary Movies 3, 4). To show whether the ester hydrolysis is indeed responsible for coacervate dissolution, time-dependent HPLC was performed. HPLC studies showed the generation of A’ and consumption of A with time (Fig. 3i, j, and Supplementary Figs. 13–15, rate of hydrolysis was found to be 141.64 ± 4.04 µM min−1). Next, we demonstrated the recyclability of the system by refueling the 1 + A system again with batchwise addition of A (10 mM) after 240 min. Upon addition of A, the coacervate droplets reformed and subsequently dissolved as a result of catalytic hydrolysis (Supplementary Fig. 16). This cycle of regeneration and dissolution was repeated for up to three cycles, which supports the fuel-driven non-equilibrium nature of the system. To probe the catalytic role of histidine residue in 1 for the temporal control over coacervation and active vacuole formation, a control peptide 3 (NH2-AFFP-CONH2, Fig. 1b) was chosen where histidine at N-terminal was replaced with alanine. The mixture of 3 (40 mM) and A (10 mM) became turbid within 2–3 min under the same buffer conditions (Fig. 4b, Supplementary Figs. 17, 18, imine confirmed from HRMS and HPLC, 3 alone was soluble). CLSM and brightfield microscopy analysis of the turbid system showed the formation of coacervate droplets with coalescence and dye sequestration abilities (Fig. 4a, Supplementary Fig. 19, FRAP plot in Fig. 4d, e). Importantly, time-dependent CLSM images for 3 + A revealed a significantly lesser depletion of droplets, suggesting the importance of the catalytic potential of 1 (Fig. 4c, Supplementary Figs. 20, 21, turbidity was not observed for the mixture of 3 and A’). Notably, the violin plot for coacervate droplets (3 and A, Fig. 4f and Supplementary Fig. 20) showed negligible change in size and population with time, which was in contrast to what was observed for 1 + A (Fig. 3c, size enlargement was not observed due to background hydrolysis of the ester, vide infra). Intriguingly, vacuole formation was not observed for 3 + A coacervate, which again suggested the critical role of the catalytic (histidine) unit in building blocks (vide infra). Also, fluorescence recovery time after photobleaching was found to be lower for 3 + A than 1 + A coacervate (DFRAP were (7.2 ± 0.4) ×10−3 μm2 s−1 and (1.3 ± 0.1) ×10−2 μm2 s−1 for 1 + A and 3 + A coacervate respectively, Supplementary Fig. 22), which suggested that the molecular diffusion within 3 + A coacervate was higher than 1 + A (ca. 1.8 times). The difference in the molecular diffusion between the two coacervates could be due to the difference in the molecular structure of peptides (histidine can participate in H-bonding and cation-pi interactions through its imidazole moiety, while alanine lacks this capability)63. CD spectrum of the 3 + A turbid solution showed a new broad peak around 275 nm (in contrast to 1 + A coacervate, a positive signal was observed) which was not present in either of the components and the mixture of 3 with the control B (Supplementary Fig. 23, no turbidity was observed for the mixture of B and 3, Supplementary Figs. 17a, 24), which suggested the generation of a chiral microenvironment in the coacervate phase. To investigate that indeed catalytic abilities played important roles, HPLC was performed, where the rate of hydrolysis for 3 + A system was found to be ~3.3 fold lower than the 1 + A system (43.29 ± 4.33 µM min−1) and was close to what was observed for background hydrolysis of A (Fig. 4g, Supplementary Figs. 25, 26). In combination, the higher hydrolytic rate in the case of 1 + A underpinned the role of histidine catalytic unit, which provided negative feedback towards coacervation.

Fig. 3: Dissolution of chiral coacervate via intrinsic catalysis.
figure 3

Time-dependent a brightfield microscopy and corresponding b CLSM images with fluorescein dye (10 μM) of 1 + A coacervate droplets. c Size distribution plot of 1 + A coacervate droplets (violin plot was prepared from the CLSM images at different time using ImageJ software, the number at each violin and green square represents number of droplets and median respectively, increment in droplet size observed after 60 min could be attributed to the formation of vacuoles and coalescence). d 3D reconstruction of the vacuole formed at 120 min. e CLSM image and f intensity profile across the vacuole formed at 120 min (dotted white line, scale bar 5 μm). g Brightfield and corresponding h CLSM images showing fusion of two vacuoles with time within 1 + A coacervate droplet. i Representative time-dependent HPLC chromatogram of 1 + A coacervate system. j Plot to show the change in A and A’ concentration with time. (Error bars represent the standard deviation of triplicate experiments). [1] = 40 mM, [A] = 10 mM, HEPES buffer (pH 8, 80 mM).

Full size image
Fig. 4: Role of histidine in augmented catalysis.
figure 4

a Brightfield microscopy images of 3 + A coacervate droplets formation and its coalescence with time. b Chemical structure of A3. c CLSM images of 3 + A coacervate at different times with rhodamine B dye (5 μM). d FRAP images of 3 + A coacervate droplet (fluorescein dye, 10 μM), and corresponding e intensity time trace of fluorescence recovery on photobleaching (shaded gray region shows the standard deviation of triplicate experiments). f Size distribution plot of 3 + A coacervate droplets (violin plot was prepared from the CLSM images at different time using ImageJ software, the number at each violin and red square represents number of droplets and median respectively). g Plot comparing the hydrolytic cleavage rate of A (Error bars represent the standard deviation of triplicate experiments). [1], [3] = 40 mM, [A] = 10 mM, HEPES buffer (pH 8, 80 mM).

Full size image

Kinetic resolution towards hydrolysis from chiral coacervate

Intracellular membraneless organelles like stress granules, ribosomes, etc., formed by the phase separation of RNA and RNA-binding proteins, are responsible for carrying out specific chemical transformations64. Here, we asked whether the chiral microenvironment of the 1 + A coacervate phase (Fig. 2g) could discriminate one of the enantiomers from a racemic mixture and facilitate kinetic resolution in reactivity/consumption65. To explore this, we chose (±) 4-formylphenyl 2-phenylpropionate (CE-CHO, Fig. 5a), a racemic ester substrate having an aldehyde group to get better proximity utilizing dynamic covalent linkage and facilitating hydrolytic activity. Brightfield microscopy showed that the coacervate droplet formation was not affected when CE-CHO (2 mM) was mixed in 1 + A system (Supplementary Fig. 27). The rate of hydrolysis of CE-CHO was found to be 138.7 ± 9.9 μM min−1 in 1 + A coacervate system (Fig. 5b, Supplementary Figs. 28, 29), which is about 3 times higher than the activity in case of only peptide 1 ca. 44.81 ± 2.91 μM min−1 (Fig. 5b, Supplementary Fig. 29, 30). This emphasized the effect on reaction rate in a crowded environment, as the interior of the coacervate droplet is dense with the building blocks (A1), which generates a crowded environment of catalytic moieties that was intrinsically generated by coacervate droplets31,32,33,34,47. The crowded catalytic environment and high partitioning of CE-CHO (vide infra) in the coacervate phase result in a higher hydrolytic rate in the phase separated state. Interestingly, HPLC analysis of the coacervate mixture showed the preferential hydrolysis of R-enantiomer (ee = 12.6 ± 0.4%) in 20 min. (Fig. 5b, c)66. Next, to experimentally verify whether the reaction is occurring within the coacervate phase, we probed the partitioning behavior of the substrate (CE-CHO) via centrifugation. The concentration of CE-CHO within the coacervate was found to be 199.9 ± 2.4 µM, which is ~4 times higher than the supernatant (KD = 4.12 ± 0.96, see Supplementary Information), suggesting that hydrolysis of CE-CHO occurs primarily in coacervate droplets. To probe further, a control experiment was performed with the chiral peptide 1 in the absence of A. Despite the capability to form imines and being the only chiral component, the soluble peptide showed a negligible preference for one enantiomer (2.5 ± 0.2% within 20 min, Fig. 5b inset, and Supplementary Fig. 31). To investigate further the role of coacervate chiral microenvironment, the kinetic resolution was monitored in the supernatant obtained on centrifuging 1 + A coacervate that resulted in 2.18 ± 0.15% of ee (Supplementary Fig. 32). Next, kinetic resolution was monitored with the addition of 1 M 1,6-hexanediol in 1 + A system, which led to dissolution of the coacervate and resulted in a decrease in the selectivity (2.9 ± 0.8% ee, Supplementary Fig. 33). These results underpinned the role of coacervate chiral microenvironment towards the observed enantiomeric excess.

Fig. 5: Kinetic resolution towards ester hydrolysis.
figure 5

a Scheme showing preferential selection of one enantiomer over another by kinetic resolution of ester hydrolysis from 1 + A coacervate droplets. b Bar plot showing rate of hydrolysis of CE-CHO from 1 + A coacervate and peptide 1, inset shows the % ee obtained from 1 + A coacervate and peptide 1 within 20 min. c Bar diagram showing change in % ee with time in 1 + A coacervate system, inset shows representative HPLC chromatogram at 20 min for (±) CE-CHO hydrolysis. d Bar diagram showing the % ee obtained for CB-CHO and CE-CL from 1 + A coacervate. e Bar plot showing the rate of hydrolysis of CB-CHO and CE-CL in 1 + A coacervate system. (All the error bars represent the standard deviation of duplicate experiments). [1] = 40 mM, [A] = 10 mM, [CE-CHO] = 2 mM, 1% MeOH/HEPES buffer (pH 8, 80 mM).

Full size image

We were intrigued to explore the effect of catalytic unit in peptide toward enhancing kinetic resolution. To this end, we monitored the kinetic resolution of CE-CHO in 3 + A coacervate (peptide 3 lacks catalytic histidine unit). Notably, the enantiomeric excess (% ee) observed from the 3 + A coacervate was only 2.45 ± 0.13% (Supplementary Fig. 34), suggesting that considerable % ee could only be realized in the presence of the catalytic residue. However, in case of peptide 2, which featured catalytic unit but formed gel (in 5% DMSO/buffer) instead of coacervate showed 6.07 ± 0.21% ee (Supplementary Fig. 35). When CE-CHO was replaced with (±) 4-chlorophenyl 2-phenylpropanoate (CE-CL, Fig. 5d), which has -Cl group instead of -CHO, the % ee was found to be 14.1 ± 0.4% (Fig. 5d and Supplementary Fig. 36). The initial rate of hydrolysis in this case was 69.2 ± 9.8 μM min−1 (Fig. 5e, Supplementary Figs. 37, 38). To further demonstrate the ability of the 1 + A chiral microenvironment to facilitate kinetic resolution, we monitored another racemic ester, (±)−1-phenylpropyl 4-formylbenzoate (CB-CHO, Fig. 5d). The enantiomeric excess observed with this substrate was found to be 12.54 ± 9.8% (Fig. 5d and Supplementary Fig. 39), where the rate of hydrolysis was 114.7 ± 19.6 μM min−1 (Fig. 5e, Supplementary Figs. 40, 41). Together, the results showed the ability of 1 + A chiral microdroplets, which, despite its highly diffusive environment exhibited higher catalytic performance along with modest yet notable difference in the preference for one enantiomeric guest.

Discussion

In conclusion, we have demonstrated non-equilibrium coacervate formation accessed via a dynamic covalent bond between a tetrapeptide and a positively charged thermodynamically activated ester. We showed the importance of proline residue by mutating the N-terminal of the peptide in prohibiting the formation of fibrils and providing a kinetically feasible path for coacervation. The catalytic activity of the coacervate triggered vacuole formation within the coacervates and subsequently drove their dissolution. Despite the liquid-like fluidity of the coacervates, the dense phase featured chiral microenvironment as suggested from the strong CD signal. Interestingly, this chiral microenvironment that emerged from the coacervate phase was utilized to achieve kinetic resolution of a racemic substrate67,68. The capability of coacervate with chiral microenvironment to facilitate such chemical transformation highlights a potential pathway for the enrichment of enantiopure chemical inventory, a fundamental question in chemical evolution. These dynamic, non-equilibrium coacervate with functional capabilities demonstrates how compartmentalization, catalysis, and symmetry breaking might have converged in early chemical evolution.

Methods

Preparation of coacervates and gels

Coacervates were prepared by dissolving A in the required volume of 40 mM of 1 or 3 (80 mM, pH 8 HEPES buffer stock). Peptide 1 was diluted by the addition of HEPES buffer (pH 8, 80 mM) for the phase diagram. The pH of coacervate samples was measured initially and after 4 h, showing no significant change (ca. ±0.03). The appearance of turbidity indicated the formation of coacervates. In the case of gel preparation, the required amount of 2 was dissolved in 5% DMSO/buffer in a vial through sonication, and gel formation was monitored by inverting the vials.

Turbidity measurement

The freshly prepared solution was put into a 96-well plate. The turbidity was measured in the BioTek SYNERGY H1 microplate reader at room temperature (25 ± 2 °C). OD values at 700 nm wavelength were considered for turbidity measurements. A well containing an equivalent volume of HEPES buffer (80 mM, pH 8) solution served as the blank.

Phase diagram

To determine the phase diagram, the concentrations of 1 and A were varied from 5 mM to 40 mM in HEPES buffer (pH 8, 80 mM). The appearance of milky (or light) turbidity indicates the phase separation (two phase). The red cross, black, and hollow blue circle show the points of observation with no, light, and milky turbidity representation, respectively, and the solid black line represents the apparent critical concentrations for the phase separation. The phase diagram for the pH compatibility study was obtained by measuring turbidity after 5–6 min of preparing coacervate samples (40 mM of 1 and 10 mM of A) in different pH buffers (Acetate, MES, HEPES, and TRIS buffers were used for pH 3–5.5, 6, 7–7.7, 8.6, respectively). For probing the role of various additives in disrupting specific interactions involved in coacervate formation, the samples were prepared by adding the required amount of 1,6-hexanediol, urea, or NaCl, along with 1 (40 mM) in HEPES buffer (pH 8, 80 mM), to A (10 mM), turbidity of the mixture was measured after 5–6 min of the addition. The stock solutions of the additives and 1 were prepared with HEPES buffer (pH 8, 80 mM).

Brightfield and confocal laser scanning microscopy

Coacervate samples were prepared by the pre-mentioned method. The samples were deposited on the 35 mm confocal dish with a well size of 10 mm. Images were captured in Olympus Laser Scanning Confocal System Model FV3000 using 100× objective lens (oil immersion, part of the Atomic Force Microscope with Rheological Measurement and Confocal Imaging Unit Facility, supported by Swarnajayanti fellowship (SB/SJF/2020-21/08)). For confocal, fluorescein dye (10 µM), rhodamine B (5 µM), and thioflavin T (10 µM) were used with coacervate samples. The excitation wavelength of 405, 488, and 561 nm was used for thioflavin T, fluorescein, and rhodamine B, respectively.

Fluorescence Recovery after Photobleaching (FRAP)

FRAP measurement was performed by Olympus Laser Scanning Confocal System Model FV3000 using 100×/1.45 numerical aperture objective lens (oil immersion). Photobleaching was performed with 488 (or 561) nm laser within a circular region of interest (ROI) having a diameter of 1.9 µm. The fluorescence intensity of ROI was monitored with ImageJ. Measurements were conducted on three different droplets from the same sample. Data was normalized through the double normalized method using Eq. (1). Data was fitted to a mono-exponential function using Eq. (2) in OriginLab to give fluorescence recovery time constant (\({\tau }_{f}\)). The half-life for the fluorescence recovery time was calculated using Eq. (3). Then the apparent diffusion coefficient (\({D}_{{FRAP}}\)) was obtained by the formula of 2D diffusion Eq. (4)69.

$$I\left(t\right)=\frac{\left[P\left(t\right)-B\left(t\right)\right]\cdot \left[R\left(0\right)-B(0)\right]}{\left[R\left(t\right)-B\left(t\right)\right]\cdot [P\left(0\right)-B\left(0\right)]}$$
(1)
$$I\left(t\right)=a+b\cdot \exp \left(-\frac{t}{{\tau }_{f}}\right)$$
(2)
$${t}_{1/2}={\tau }_{f}\,{{\mathrm{ln}}}\left(2\right)$$
(3)
$${D}_{{FRAP}}=\frac{0.88{r}^{2}}{4{t}_{1/2}}$$
(4)

Where \(I\left(t\right)\) represent the normalized fluorescence intensity, \(P\left(t\right)\) is the fluorescence intensity of the photobleached area, \(R\left(t\right)\) is the fluorescence intensity of the droplet other than the bleached area for reference, and \(B\left(t\right)\) is the intensity of the arbitrarily chosen background outside of the droplet at different times t. P(0), B(0), and R(0) represents the corresponding intensities at time t = 0 (before photobleaching). \({\tau }_{f}\) is the fluorescence recovery time constant, \({t}_{1/2}\) is the half-life for the fluorescence recovery time, \({D}_{{FRAP}}\) is the apparent diffusion coefficient, and r is the radius of the bleached area.

Transmission Electron Microscopy (TEM)

Samples were prepared by drop casting the 80 times diluted gel sample in Milli Q water (at ca. 12 h) on the TEM grid and allowed to adsorb for 60 s. Excess volume was wicked off by blotting paper. Thereafter, 1 wt% uranyl acetate solution was added as a staining agent and waited for 2–3 min. Then, the excess solution was removed with blotting paper. The drop-casted samples were dried and kept in vacuum until the images were recorded in JEOL JEM-2100 Plus.

NMR studies

All the 1H NMR studies were carried out in Bruker (400 MHz)/JEOL (400 MHz) at 25 °C in respective solvents.

HRMS measurement

All the HRMS studies were carried out in Waters Xevo G2-XS QTof system.

For the detection of Imine, the coacervate samples (1 + A, 3 + A) were prepared by following the above-mentioned procedure ([1]/ [3] = 40 mM, [A] = 10 mM in 80 mM HEPES buffer) and diluted one-fifth in MeOH before injection in HRMS after ~5 min of coacervate sample preparation. The imine peak was detected in HPLC-MS QDa and further probed by HRMS.

Circular dichroism spectroscopy

JASCO J-1500 circular dichroism spectrometer fitted with a Peltier temperature controller to maintain the temperature at 25 °C was used to record the CD spectra. Coacervate samples were prepared using the aforementioned method and were transferred to a quartz cell with a path length of 0.1 mm for CD measurement. For the CD spectra of samples at different pH, acetate buffer (pH 4, 80 mM) and HEPES buffer (pH 8, 80 mM) were used. To record each spectrum, scanning wavelengths from 600 nm to 190 nm were used at a 100 nm/min scanning rate. At least three successive wavelength scans were taken to average for each sample.

Thermogravimetric analysis (TGA)

The water content of the coacervate droplet was analyzed by TGA experiment. The coacervate sample was prepared by following the aforementioned procedure (40 mM of 1 and 10 mM of A were used). The coacervate pellet was collected by centrifuging the turbid solution, followed by removing the supernatant from the tube. The coacervate pellet was immediately loaded onto a ceramic crucible. TGA was carried out on a Perkin-Elmer TGA 4000 Simultaneous Thermal Analyzer. The sample was heated at 1 °C/min over a range of 30–300 °C. Nitrogen gas flow was maintained at 20 mL/min. The data over 30–110 °C were presented. Above 150 °C the residual materials underwent thermal decomposition.

High-performance liquid chromatography-mass spectrometry (HPLC-MS)

HPLC was performed in Waters HPLC system (1525 binary pump equipped with 2998 PDA detector) connected with Waters QDA detector. Flow rate of 0.7 mL/min was maintained with a gradient of acetonitrile/water (both containing 0.1% TFA). Synergi 4 μm Fusion-RP 80 Å analytical HPLC column was used to monitor the changes in concentration of A and A’. The coacervate mixture was diluted 40 times to completely hydrolyzed the imine before injection to HPLC and the hydrolysis of A and formation of A’ were monitored by extracting the chromatogram at λ = 250 nm. For the detection of Imine, the coacervate sample (1 + A or 3 + A) was diluted one-fifth with 50% ACN/water. Flow rate of 0.7 mL/min was maintained with a gradient of acetonitrile/water. Synergi 4 μm Fusion-RP 80 Å analytical HPLC column was used and chromatogram was extracted at 260 nm and 290 nm for the detection of A1 and A3, respectively.

XBridge C18 5.0 µm 4.6 ×250 mm analytical HPLC column was used to monitor the changes in concentration of CE-CHO/CB-CHO/CE-CL with time. The sample was prepared by adding the required amount of CE-CHO/CB-CHO/CE-CL (200 mM stock in methanol) just before the appearance of turbidity from the HEPES buffer (80 mM, pH 8) mixture of 1 (40 mM) and A (10 mM). The hydrolysis of CE-CHO/CB-CHO/CE-CL was monitored by extracting the chromatogram at λ = 252 nm (or 240 nm). The enantiomeric excess for the hydrolysis of CE-CHO/CB-CHO/CE-CL was determined by using HPLC performed with a Chiralpak AD-RH (Daicel) reverse phase 5.0 µm 4.6 ×150 mm column. The chromatogram for CE-CHO/CB-CHO/CE-CL was extracted at λ = 252 nm (or 240 nm). All the samples were diluted with 50% ACN/water (0.1% TFA) immediately before injection into HPLC.

Partition study

The partition study of CE-CHO was performed by preparing the 1 + A coacervate sample (40 mM of 1 and 10 mM of A) by adding CE-CHO (2 mM). The sample was centrifuged for ca. 5 min immediately after the turbidity appeared, and the supernatant (volume 98.2 μL) and coacervate pellet (volume ~1.8 μL) were analyzed by HPLC. The partition coefficient (KD) value was determined by using the formula.

$${K}_{D}=\frac{{C}_{P}}{{C}_{S}}$$

Where, CP and CS are the concentration of CE-CHO in coacervate pellet and supernatant, respectively.