Cells can react to their environment by changing the activity of enzymes in response to specific chemical signals. Artificial catalysts capable of being activated by chemical signals are rare, but of interest for creating autonomously responsive materials. We present an organocatalyst that is activated by a chemical signal, enabling temporal control over reaction rates and the formation of materials. Using self-immolative chemistry, we design a deactivated aniline organocatalyst that is activated by the chemical signal hydrogen peroxide and catalyses hydrazone formation. Upon activation of the catalyst, the rate of hydrazone formation increases 10-fold almost instantly. The responsive organocatalyst enables temporal control over the formation of gels featuring hydrazone bonds. The generic design should enable the use of a large range of triggers and organocatalysts, and appears a promising method for the introduction of signal response in materials, constituting a first step towards achieving communication between artificial chemical systems.
Control over enzymatic activity is at the basis of cellular communication and the regulation of a wide range of biological processes. Enzymes are often activated in a covalent manner by (de-)phosphorylation, for example in cell signalling. Allosteric enzymes, such as haemoglobin, are hindered or stimulated by non-covalent interactions with a regulatory molecule1, 2. Such rudimentary forms of communication and regulation are almost entirely absent in artificial materials, but could lead to the development of soft materials capable of autonomously responding to changes in their environment. Although enzymes are typically very efficient catalysts, the type of reactions they catalyse and their operating conditions are limited. An artificial catalyst that is activated by a chemical signal opens possibilities for autonomous spatial and temporal control over systems at a molecular and supramolecular level. Only a few examples of synthetic catalysts with addressable activity exist, predominantly controlled using light as a signal3,4,5,6,7. Catalysts whose activity is controlled using a chemical signal are nearly absent from the literature8,9,10,11 and often suffer from complex design and synthesis as well as a very specific operating mechanism.
Here, we present proof of principle of an organocatalyst that is activated by a chemical signal. This concept provides a generic design to enable autonomous response to biologically and mechanically relevant signals from the environment. The signal responsive catalyst is designed as a self-immolative molecule12,13,14,15, which fragments upon reaction with a specific chemical signal to release an active catalyst. We synthesise a protected aniline (pro-aniline 1) that liberates the organocatalyst aniline 2 upon reaction with the chemical signal H2O2. Aniline is a nucleophilic catalyst for hydrazone formation and exchange16,17,18, a reaction frequently used in soft19 and dynamic covalent materials20. H2O2 is released by the enzymatic oxidation of many different disease related biomarkers such as glucose, lactose, sarcosine, uric acid, choline and acetylcholine, making it a highly relevant biological signal21,22,23. Furthermore, strained polymers have been shown to generate H2O2 in the presence of water, making H2O2 an interesting mechanically generated chemical signal24. Activating pro-aniline 1 with H2O2 as a chemical signal leads to an almost instant 10-fold increase in rate of hydrazone formation. In addition, using the pro-catalyst in gels featuring hydrazone bonds enables control over material formation, creating materials that can respond to a specific chemical signal.
Activation of the catalyst
The pro-catalyst pro-aniline 1 is activated by the chemical signal H2O2 and catalyses hydrazone formation (Fig. 1a, b). Pro-aniline 1 was synthesised in only two steps in good yields (Fig. 1c). We confirmed the release of aniline 2 from pro-aniline 1 (72 mM) upon reaction with H2O2 (18 equivalents) using GC/MS (gas chromatography/mass spectrometry), showing that 1 fragments completely to release 2 (Supplementary Fig. 1). We analysed the kinetics of aniline 2 release using UV/vis spectroscopy (Supplementary Fig. 2a, b). The timescale for the liberation of 2 from 1 (0.5 mM) depends on the amount of H2O2 added, taking 5 min using 50 equivalents of H2O2, and 15 min using 10 equivalents. In the absence of the H2O2 signal, pro-aniline 1 is stable in solution for over 15 h (Supplementary Fig. 2c, d).
We investigated the catalytic activity of pro-aniline 1 upon activation, by comparing the rates of hydrazone formation in a model reaction (Fig. 2a)25, 26. Aldehyde 4 (0.5 mM) reacts with hydrazide 3 (0.1 mM) to form hydrazone 5 in a buffered medium (100 mM phosphate buffer pH 7.4) with 20% dimethylformamide (DMF) (Fig. 2b). This reaction is catalysed by aniline 2 (0.5 mM), giving a 19-fold increase in reaction rate with respect to the uncatalyzed reaction. Unactivated pro-aniline 1 (0.5 mM) does not influence the reaction rate of hydrazone formation. Addition of H2O2 (2.5 mM) to pro-aniline 1 (0.5 mM) gives a relative reaction rate of 10, indicating efficient activation of the organocatalyst (Fig. 2b and Table 1). H2O2 alone (2.5 mM) does not increase the reaction rate of hydrazone formation (Supplementary Fig. 3a, b). Furthermore, the pH was monitored during the reaction with pro-aniline 1 and H2O2: the pH remained stable at a value of 8. As our solvent system alone (100 mM phosphate buffer with 20% DMF) gives a pH of 8, we conclude that the reactions were sufficiently buffered. After activation of pro-aniline 1, the reaction rate is lower than when using native aniline 2. In an attempt to explain this apparent loss of catalytic activity, we investigated the influence of H2O2 and of boric acid on the activity of aniline 2, but did not observe a lower rate (Supplementary Fig. 3c, d). Although we confirmed complete conversion of pro-aniline 1 after addition of more than 1 equivalent of H2O2 and we were able to detect aniline 2 after an overnight hydrazone reaction in the presence of a substoichiometric amount of pro-aniline 1 + H2O2 (Supplementary Fig. 4), it might be the case that a small amount of aniline 2 is degraded over time. Importantly, using our signal responsive catalyst, we should be able to elicit a change in reaction rate at any given time during the process. To show this, pro-aniline 1 (0.5 mM) and reactants 3 (0.1 mM) and 4 (0.5 mM) were dissolved in buffer and mixed. Upon addition of the signal H2O2 after 1 h, we observed an immediate 9-fold increase in the reaction rate (Fig. 2c). The rapid and significant increase of reaction rate when exposed to a chemical signal shows that the activation of a pro-catalyst enables an instant and autonomous response to a chemical change in the environment.
Control over gel formation by activation of the catalyst
As a first application of the signal responsive catalyst, we chose to couple the formation of a hydrazone polymer gel material27 to a chemical signal. We synthesised an alternating polyethylene glycol/benzaldehyde copolymer using mesylated polyethylene glycol (molecular weight 5.4–6.6 kg mol−1) and 3,4-dihydroxybenzaldehyde. A polydisperse polymer 6 (M w ~ 1×105 g mol−1) featuring benzaldehyde groups was obtained (Fig. 3a). Hydrazone formation between polymer aldehyde 6 and trishydrazide 7 is catalysed by aniline 2 and leads, under the right conditions, to the formation of transparent polymer gels (Fig. 3a, c). Using the signal induced activation of pro-aniline 1 we are able to control the rate of gel formation, the moment of gel formation (temporal control), and the mechanical properties of the obtained gels. An inverted tube test was used to investigate the influence of catalyst activation on gel formation. The gel tests presented here were performed using 136 mg ml−1 polymer aldehyde 6, 10 mM hydrazide 7 and 10 mM catalyst in 20% DMF in aqueous buffer (100 mM phosphate buffer pH 7.4) in 4 ml vials. The hydrazide 7 and aldehyde 6 mixtures with either aniline 2 or pro-aniline 1 activated with H2O2 (10 equivalents) form gels within 1 h (Fig. 3b). In contrast, for the mixtures with unactivated pro-aniline 1 or without catalyst, gelation takes 2 h.
We performed oscillatory rheology to quantify the rate of gel formation under influence of the catalyst and to investigate the mechanical properties of the formed materials (Fig. 3b). The gel prepared with unactivated pro-aniline 1 is comparable in stiffness and formation time to the uncatalyzed gel, whereas the gel prepared with pro-aniline 1 and H2O2 (10 equivalents) is comparable to the gel formed using aniline 2 as a catalyst. After 6 h of gelation time the catalysed gels show elastic moduli (G′) that are 1.5 times higher than the G′ values we measured for the uncatalyzed gels, which indicates that the gel stiffness is controlled by catalysis19, 28. The difference in timescale of gel formation is especially apparent in rheology, as the cross-over for G′ and G″ (the gel point) for the catalysed gels is observed after 30 min, whereas this cross-over takes place after almost 2 h for uncatalyzed gels (Supplementary Fig. 5a). Thus, activation of the catalyst by the chemical signal influences the gel formation rate as well as the gel stiffness.
With the signal responsive catalyst, we can now attempt to control the moment of material formation using a chemical signal. In the rheometer, we added a chemical signal to a solution of 6 and 7 containing pro-aniline 1, 20 min after mixing all components (Fig. 3c). We observed a significant increase in the rate of gel formation shortly after addition of the chemical signal. A control experiment lacking the added signal showed a delayed and smaller increase of the elastic modulus. Importantly, these observations show that our system allows for temporal control over reaction rates and material formation.
Control over supramolecular gel formation by activation of the catalyst
To investigate the scope of our signal responsive catalyst, we also used pro-aniline 1 to control the formation of a supramolecular trishydrazone hydrogel described previously by our group19, 29, featuring trishydrazide 7 (16 mM) and aldehyde 8 (96 mM) (Fig. 4a). Using either aniline 2 (10 mM), or pro-aniline 1 (10 mM) with H2O2 (10 equivalents), the supramolecular gels forms overnight. In contrast, without catalyst or with unactivated pro-aniline 1, no gels are obtained (Fig. 4b). The activated pro-aniline can thus be used to control the formation rate and properties of polymer gels as well as supramolecular gels featuring hydrazone bonds.
In summary, we report a deactivated organocatalyst that is activated by a chemical signal. The pro-aniline 1 was synthesised in only two steps in good yields. Activating the pro-aniline 1 organocatalyst with H2O2 as a chemical signal increases the rate of hydrazone formation 10-fold almost instantly. We obtained control over gel formation for a polymer gel and for a supramolecular gel, both featuring hydrazone bonds. For the polymer gel, we controlled the moment of gelation: addition of the chemical signal to the unactivated pro-aniline 1 catalyst at an arbitrary moment after mixing all gelation components leads to a significant increase in the rate of gel formation. This shows that it is possible to control the moment of material formation in response to a chemical signal. As self-immolative trigger groups allow response to a wide range of signals12, we have aimed to develop a generic method for the design of activatable catalysts. Furthermore, since self-immolative molecules can be used to incorporate more than one molecule of interest, the design can be used for signal amplification12. We are currently working on several organocatalysts that are responsive to a range of specific chemical signals. With the current results, we have developed a promising method for the introduction of signal response in molecular materials, constituting a first step towards achieving communication between artificial chemical systems.
Instrumentation and characterisation
Nuclear magnetic resonance (NMR) spectra were recorded on an Agilent-400 MR DD2 (400 MHz for 1H and 100.5 MHz for 13C) at 298 K using residual protonated solvent signals as internal standard (1H: δ(CHCl3) = 7.26 p.p.m., δ(CH3OH) = 3.31 p.p.m. and 13C: δ(CHCl3) = 77.16 p.p.m., δ(CH3OH) = 49.00 p.p.m.). Thin layer chromatography (TLC) was performed on Merck Silica Gel 60 F254 TLC plates with a fluorescent indicator with a 254 nm excitation wavelength and compounds were visualised under UV light of 254 nm wavelength. GC/MS was performed with a Shimadzu QP-2010S GC/MS. Liquid chromatography–mass spectrometry (LC/MS) was performed on a Shimadzu Liquid Chromatograph Mass Spectrometer 2010, LC-8A pump with a diode array detector SPD-M20. The column used was the Xbridge Shield RP 18.5 μm (4.6 × 150 mm). UV/Vis spectroscopic measurements were performed on an Analytik Jena Specord 250 spectrophotometer; quartz cuvettes with a path length of 1 cm were used. Oscillatory experiments were performed using an AR G2 rheometer from TA Instruments in a strain-controlled mode; the rheometer was equipped with a steel plate-and-plate geometry of diameter 40 mm and a water trap. The temperature of the plates was controlled at 25 ± 0.2 °C. Measurements were performed at a frequency of 1 Hz while applying 1% strain. Gel permeation chromatography (GPC) was performed using a Shimadzu Prominence GPC system equipped with 2× PL aquagel-OH MIXED H columns (Agilent, 8 µm, 300 × 7.5 mm) and refractive index detector.
For the synthesis of cis,cis-cyclohexane-1,3,5-tricarbohydrazide 7 and 3,4-bis(2-(2-methoxyethoxy)ethoxy)benzaldehyde 8 we refer to procedures described in the literature30. Detailed procedures for the synthesis of pro-aniline 1, hydrazone 5 and PEG-aldehyde copolymer 6 are described in the Supplementary Methods, NMR and GPC spectra are shown in Supplementary Figs 9–17.
Investigation of the release of aniline 2 from pro-aniline 1
The self-immolative reaction of pro-aniline 1 with H2O2 was investigated with TLC, GC/MS and UV/vis spectroscopy. TLC: pro-aniline 1 (4.8 mg, 13.6 µmol) was dissolved in acetonitrile (0.5 ml) and a solution of H2O2 in deionized water (18 equivalents, 0.245 mM, 0.5 ml) was added. TLCs (eluent = 1:1 petroleum ether: ethyl acetate) were taken at t = 0, 1, 5, 10, 15 and 20 min. The reaction mixture and aniline 2 were run on the TLC plate for comparison. After 10 min all starting compound had disappeared, replaced by aniline 2 and side products. GC/MS: pro-aniline 1 (5 mg, 18 µmol) was dissolved in 0.25 ml ethyl acetate. H2O2 (18 equivalents, 25.4 µl in 0.25 ml deionized water) was added and after the mixture was allowed to stir for 1 h at room temperature, the reaction was quenched with a saturated sodium thiosulfate solution. The organic layer of the reaction mixture was analysed by GC/MS: after the reaction the product peak had disappeared and aniline 2 (m/z 93) was detected in the reaction mixture (Supplementary Fig. 1). UV/vis: the self-immolative reaction of pro-aniline 1 followed in UV/vis was performed in a 20% DMF in phosphate buffer (100 mM, pH 7.4) solution, using 0.5 mM concentration of pro-aniline 1 and 18 equivalents of H2O2 (Supplementary Fig. 2a, b)). The self-immolative reaction is complete when the increase of absorption reaches a plateau. The stability measurement of pro-aniline 1 overnight was performed using the same conditions of pro-aniline 1 (0.5 mM) in a mixture of 20% DMF in buffer. Without H2O2 pro-aniline 1 is stable for over 15 h (that is, does not release aniline 2, Supplementary Fig. 2c, d).
Investigation of the catalytic activity of pro-aniline 1 after activation
A calibration line was measured for hydrazone product 5 in 20% DMF in phosphate buffer (100 mM, pH 7.4). The extinction coefficient of hydrazone 5 at 350 nm under these conditions is 1.3 ± 0.087×104 M−1 cm−1 (Supplementary Fig. 6). The hydrazone reaction (Fig. 2a) was performed in 20% DMF in a 100 mM phosphate buffer pH 7.4, containing 0.5 mM of 4-nitrobenzaldehyde 4, 0.5 mM of catalyst, 0.1 mM of 4-hydroxybutyric acid hydrazide 3 and 2.5 mM of H2O2. The quartz cuvettes contained a total reaction volume of 2 ml. The stock solutions of the reagents were added as follows: aldehyde 4 solution (10 mM in DMF), phosphate buffer, DMF, catalyst solution (10 mM in DMF), the hydrazide 3 solution (2 mM in buffer). The cuvettes were closed using Teflon caps and thoroughly mixed by converting the cuvette upside down four times. The spectra of the starting reaction mixtures were measured (reference measurement, 250–450 nm, 10 nm s−1). The product peak at 350 nm was followed using a 6-sample holder (standard absorption measurement, slow time scan, measuring wavelength 350 nm, scan every 60 s, Fig. 2b, c). At the end of the measurement, single scans were measured again using the same settings as for the starting reaction mixtures (Supplementary Fig. 7). The pH was monitored for the reaction with pro-aniline 1 (0.5 mM) and H2O2 (2.5 mM), every 10 min for the first 2 h and once after 18 h. At all times a pH of 8.0 was measured. As the phosphate buffer (100 mM, pH 7.4) with 20% DMF alone also gives a pH of 8.0, this indicates that the pH did not change during the reaction. We ensured that no side reactions occurred (Supplementary Table 1). Addition of aniline 2 to the uncatalyzed reaction after 1 h of reaction time gave a similar response in reaction rate as addition of H2O2 to the reaction with pro-aniline 1 (Supplementary Fig. 3e). The pseudo-first-order rate constants were determined by converting the absorbance measured during the reactions to concentration using the extinction coefficient of the product and by fitting the natural logarithm of the concentration (M) over time (s). The graph was fitted using linear regression in Origin Pro 2015 to yield the pseudo-first-order reaction rate constant. For the uncatalyzed reaction and the reaction with pro-aniline 1, the first 10 h of the reaction were taken to determine the rates. For the reaction in the presence of activated pro-aniline 1 and for the reaction catalysed by aniline 2, the first 15 min were used to determine the rate (Supplementary Fig. 8 and Supplementary Table 2). To ensure that aniline 2 was not used up during the reaction, we performed a 65 h reaction of hydrazide 3 with aldehyde 4 in the presence of 0.5 equivalents of pro-aniline 1 with H2O2 and analysed the reaction mixture afterwards with GC/MS. We were able to detect aniline 2 (m/z 93) in the reaction mixture. Conditions: hydrazide 3 (20 mM), aldehyde 4 (20 mM), pro-aniline 1 (10 mM), H2O2 (11 mM) in 20% DMF in 100 mM phosphate buffer pH 7.4. The reaction mixture was extracted after 65 h of reaction time with dichloromethane. The organic layer was evaporated and re-dissolved in ethyl acetate for GC/MS analysis. Aniline 2 was detected in the reaction mixture, MS (GC/MS) m/z: 93 [M], 66 [C5H6]+• (expected m/z = 93.06), retention time: 11 min.
Rheology of the polymer gel
The storage and loss moduli G′ and G″ were followed in time during the formation of the gel, using a rheometer (Supplementary Fig. 5a). Total volume of the gels is 0.6 ml. Composition of the gels: 140 mg ml-1 polymer aldehyde 6, 10 mM aniline 2 or 10 mM pro-aniline 1, 10 mM hydrazide 7, 20% DMF in 100 mM phosphate buffer pH 6.0. First, the polymer aldehyde 6 (84 mg) was weighed out in the shell vial, then we added buffer, the catalyst solution (120 µl of 50 mM stock solution in DMF) or 120 µl DMF, the H2O2 solution (120 µl of 500 mM stock solution in buffer, 10 equivalents) and last, the hydrazide 7 solution (150 µl of 40 mM stock solution in buffer). The mixture was vortexed and poured directly on the rheometer plate. The rheometer plate was rotated slowly when it was lowered to ensure equal spreading of the sample. The G′ and G″ were measured while the gel formed on the plate. When no significant change in the moduli was observed anymore, the measurement was stopped. The frequency dependence of the gels was measured (Supplementary Fig. 5b): none of the gels showed frequency dependency in the frequency range that was used (0.01–100 Hz). A strain sweep was measured additionally; a strain higher than 100% could usually not be applied, as the gel gave too much resistance to the rheometer (Supplementary Fig. 5c). At increasing strain, the oscillatory stress increases linearly (Supplementary Fig. 5d). The rheology experiments where the chemical signal was added during the measurement, a gel volume of 1 ml was used. After mixing the stock solutions of polymer aldehyde 6, hydrazide 7 and pro-aniline 1, we poured the mixture on the rheometer plate and allowed gelation while measuring rheology. After 20 min the measurement was stopped, the top rheometer plate was raised and H2O2 (50 µl, 2 M, 10 equivalents) was added (Fig. 3c). We also performed an experiment where we let the uncatalyzed gel form and added aniline 2 (50 µl, 200 mM) after 60 min (Supplementary Fig. 5e).
Inverted tube test supramolecular gel
Aldehyde 8, hydrazide 7 and catalyst solutions were mixed in 4 ml vials and left overnight. The next day, the vials were turned upside down. When the content of the vial would stay at the bottom of the vial for at least 5 min and was able to sustain its own weight, we assumed a gel had formed. Gelation conditions: 16 mM cis,cis-cyclohexane-1,3,5-tricarbohydrazide 7, 96 mM 3,4-bis(2-(2-methoxyethoxy)ethoxy)benzaldehyde 8, 10 mM catalyst, 100 mM H2O2, solvent 20% methanol in phosphate buffer (100 mM, pH 6.0). Sample size = 1 ml. The gelation took place overnight at room temperature (Fig. 4a, b).
Data relevant to the findings of this study are available from the corresponding author on request.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This work was supported by the Netherlands Organisation for Scientific Research (NWO VIDI grant to R.E.). We thank Maarten C.J.K. Gorseling for GC/MS measurements.