Temporal Control of Gelation and Polymerization Fronts Driven by an Autocatalytic Enzyme Reaction

Abstract Chemical systems that remain kinetically dormant until activated have numerous applications in materials science. Herein we present a method for the control of gelation that exploits an inbuilt switch: the increase in pH after an induction period in the urease‐catalyzed hydrolysis of urea was used to trigger the base‐catalyzed Michael addition of a water‐soluble trithiol to a polyethylene glycol diacrylate. The time to gelation (minutes to hours) was either preset through the initial concentrations or the reaction was initiated locally by a base, thus resulting in polymerization fronts that converted the mixture from a liquid into a gel (ca. 0.1 mm min−1). The rate of hydrolytic degradation of the hydrogel depended on the initial concentrations, thus resulting in a gel lifetime of hours to months. In this way, temporal programming of gelation was possible under mild conditions by using the output of an autocatalytic enzyme reaction to drive both the polymerization and subsequent degradation of a hydrogel.

There is much interest in the design of functional and adaptive polymer systems by the use of reaction networks under kinetic control. [1] Recent strategies for the spatial and temporal control of gelation were inspired by the dissipative structures that form far from equilibrium in natural systems, such as actin filaments.H ence,g el lifetime has been controlled by tuning the timescale of competing self-assembly and disassembly processes by using enzyme catalysts,t he in situ formation of gelators,orthe injection of promotors for self-assembly and/or deactivators for self-destruction. [2] In these systems,g elation began immediately after the addition of the catalysts/fuel. Many materials-chemistry applications,such as adhesives, coatings,s ealants,a nd injectable biomedical formulations, require an initial slow reaction followed by rapid curing.F or injectable biomedical formulations,s ubsequent degradation of the gel for drug release is also desirable.A ni nduction period before the rapid reaction can be introduced either through the consumption of an inhibitor that prevents the accumulation of products,f or example,t ime-lapse polymerization was possible in ab ase-catalyzed thiol-Michael addition reaction by the use of acid inhibitors, [3] or as the result of an initial slow evolution of achemical species or heat. In freeradical polymerization, the exothermicity of the reaction can lead to an increase in the rate as the reaction progresses (thermal feedback);h owever,i tm ay result in thermal runaway. [4] Other rate-acceleration processes are also difficult to control, such as the lag phase that occurs in supramolecular polymerization as ar esult of slow initial nucleation steps [5] and the Tr ommsdorff-Norrish gel effect with adecrease in the rate of termination as gelation proceeds. [6] One advantage to the presence of thermal feedback in polymerization is the ability to create cure-on-demand systems in which the formulation does not react until the external application of localized heating and then propagates as ac onstant-velocity (cm min À1 )p olymerization front. [7] Adhesives,f or example,c an be readily applied as aliquid or paste and then rapidly cured even in inaccessible locations by the propagation of the front from the initiation site. [8] Frontal polymerization has also been used to create intricate endoskeletons in flexible materials. [9] However,t hermal frontal polymerization requires relatively thick layers and/or surfaces that are poor thermal conductors lest heat loss quench the propagation. It also involves large temperature changes (> 100 8 8C).
Other techniques for initiating polymerization, such as irradiation with UV light, have also been used for generating fronts.P hotofrontal polymerization typically requires ac onstant input of light for propagation. [10] Isothermal fronts can be initiated from ap olymer seed to produce gradient-index optical materials,b ut these fronts are limited to free-radical systems that exhibit the gel effect and whose polymers are soluble in their monomers. [11] We have developed amethod for time-lapse gelation and polymerization fronts for cure-on-demand applications under mild conditions.R ather than exploiting intrinsic rate acceleration in the polymerization process,weused the product of an aqueous-phase autocatalytic reaction to drive the formation of athiol-acrylate hydrogel. Thetime to gelation can be controlled through the initial concentrations.F urthermore, gel lifetime can be tuned, as the gel is susceptible to hydrolytic degradation, the rate of which also depends on the initial composition.
Autocatalytic reactions are frequently exploited for the design of complex dynamic behavior in systems chemistry and synthetic biology. [12] There are many autocatalytic systems that show induction periods and propagating fronts,i ncluding inorganic systems, [13] self-replicating organic reactions, [14] and biological and biopolymeric systems,s uch as the propagating fronts of RNAr eplication [15] and small DNAo ligonucleotides. [16] Although fascinating systems,t hey do not readily lend themselves to practical applications, as they either involve harsh, oxidative chemicals or cannot easily be coupled to other processes.
To create as ystem that operates under mild conditions,w eu sed an autocatalytic enzyme-catalyzed reaction:t he urea-urease reaction. Urease is awell-studied enzyme that is present in many natural systems, including various plants,s oil, and microorganisms,s uch as the bacterium Helicobacter pylori,which uses the reaction to raise its local pH, thereby protecting itself against the acidic environment of the stomach. Theurease-catalyzed hydrolysis of urea has been exploited in enzyme-based logic gates, [17] biocement for crack healing, [18] and chitosan gels for cell delivery; [19] its autocatalytic nature, however, was not used in these applications.
Theu rea-urease reaction displays rate acceleration as ar esult of its bell-shaped rate-pH curve coupled with the production of abase ( Figure 1a). If the initial pH value is low (pH % 4), as low increase in pH occurs,f ollowed by ar apid conversion to the high-pH state (pH % 9), because the formation of ammonia leads to an increase in the rate of production of ammonia. Theinduction period of the reaction can be taken as the time for the reaction to reach pH 7and has aw ell-defined dependence on the initial concentrations of urease,u rea, and acid, as well as the temperature. [20] The reaction displays useful features inherent to autocatalytic reactions,i ncluding the ability to respond to as mall amount of base with at ransition from the low-pH "off" state to the high-pH "on" state,a nd the potential for oscillations and propagating pH fronts. [21] Thehigh-pH state of the urea-urease reaction can be used to drive ab ase-catalyzed thiol-Michael addition reaction ( Figure 1b). Thiol-acrylate chemistry has aw ide range of applications,owing in part to the mild conditions the reactions need, the presence of thiols in biological systems,a nd the great variety of monomer options. [22] We used water-soluble monomers,e thoxylated trimethylolpropane tri(3-mercaptopropionate) (Thiocure ETTMP 1300) and poly(ethylene glycol) diacrylate (PEGDA7 00), to create ao ne-pot aqueous-phase system. As imilar reaction was previously characterized in phosphate buffers and displayed an exponential dependence of the gelation rate on the pH value with ag elation time of seconds when the pH value was above 8 at T = 25 8 8C. [23] When asolution of urease was added to asolution of urea/ ETTMP/PEGDA ( Figure 2a), the initial pH value was around 4a saresult of the small amount of 3-mercaptopropionic acid (3-MPA) present in ETTMP (see the Supporting Information for further details) and increased to more than 8 after an induction period (Figure 2a). Thes igmoidal charac-

Angewandte Chemie
Zuschriften teristic of the pH-time curve was preserved with different initial concentrations (Figure 2b). TheM ichael addition reaction had no discernible effect on the change in the pH value,a sd etermined in control experiments performed with water in place of the PEGDA. Gelation took place rapidly above pH 8a nd was accompanied by ac essation in motion of the magnetic stirrer.
Thefinal pH value ranged from 8.5 to 9.5;lower values of about 7a re also possible for smaller urea concentrations (Figure 2c). [21a] Thef inal pH value was determined by the ammonia/ammonium ratio,w hich depended upon the initial amount of urea and the acid. Thus,t he final pH value increased as the concentration of urea increased and decreased as the concentration of ETTMP or 3-MPA increased, but remained approximately constant with changes in enzyme concentration. Thei nitial pH value after mixing was determined by both the pH value of the ETTMP stock solution and the production of ammonia during the mixing period. Hence,the initial pH value decreased with increasing ETTMP concentration and increased with increasing urea and enzyme concentrations,even though the pH value of the ETTMP stock solution remained the same in the latter cases (Figure 2d-f).
Thei nduction period depended upon the rate of production of ammonia and increased with decreasing concentrations of urea and urease.T he time to reach pH 7w as thus inversely correlated with the initial pH value after mixing. Thetrends of the dependence of the induction period on the initial concentrations agree well with those found in our earlier study without ETTMP when the pH value of the stock solution was adjusted with sulfuric or acetic acid. [21a] The induction period depends on the nature of the acid, as weak acids can buffer the pH change, thus reducing the reaction rate.Adifferent acid (or base) could be added to the stock solution to tune the induction period independently of the ETTMP concentration.
Thus,t he time before gelation can be controlled by tuning the initial composition of the reaction mixture with three variables:the substrate,e nzyme,a nd acid. Reproducible induction times of several minutes to hours were observed under the conditions specified. Theoretically,i nduction times of months are possible in ao ne-pot system, but in practice the reaction is limited by the eventual loss of enzyme activity in solution (days to weeks,see the Supporting Information). Increased enzyme stability and longer induction times are possible at lower temperatures.
An additional degree of control over the time to gelation is made possible by the ability of the reaction to support propagating pH fronts.I nathin layer (1 mm) in ap etri dish, the enzyme-catalyzed reaction was initiated locally by the addition of ab ase,t hus giving rise to ar eaction-diffusion front that converted the medium from acid (yellow) to base (blue), as visualized by the use of ap H indicator. Thei ncrease in pH locally catalyzed the Michael addition reaction, thus resulting in polymerization fronts that converted the mixture from al iquid into ag el before the corresponding induction period in aw ell-stirred mixture was complete.T he polymerization front was imaged by shadowgraphy (see the Supporting Information for further details). [24] With this technique,t he position of the polymerization front was visible as adark band surrounding an expanding blue disk (Figure 3a).
Theintensity profiles along ahorizontal slice in images at three different times are shown in Figure 3b.Unlike diffusive processes alone,i nw hich ac hemical becomes progressively more dilute in space,t he amplitude of chemical change associated with autocatalytic fronts is constant. [21a, 25] In this case it corresponds to the pH change observed in the experiments with well-stirred mixtures.T hus,a te ach point in space the rate of increase in pH, the final pH value,and the rate of conversion from al iquid into ag el is the same,b ut there is aphase lag between points corresponding to the time at which the front passes.
Autocatalytic reaction fronts propagate with constant velocity;hence,alinear space-time plot was obtained for the polymer front (Figure 3b). Fronts propagated with speeds that ranged from 0.02 to 0.2 mm min À1 ,d epending on the initial concentrations (Figure 3c). Thev elocity of autocatalytic reaction-diffusion fronts is dictated by the reaction rate and diffusion coefficient of the autocatalytic species,i nt his case,t he base.T hus,t he front speed can be related to the induction period of the well-stirred reaction:ashorter induction period resulted in af aster front (Figure 3c). The fronts observed in this study were slower than corresponding

Angewandte
Chemie fronts obtained in the absence of polymerization in earlier aqueous-phase experiments,p robably as ar esult of the increased viscosity of the reaction mixture. [21a] Convective effects can lead to front speeds that depend on the orientation of the reaction vessel and enhance mixing or even extinguish the front, thus limiting cure-on-demand applications. [26] To determine whether the fronts were subject to convective effects,w ec ompared the velocity of ascending and descending fronts in small vials.Intwo vials with identical reaction mixtures,adrop of NaOH (pH 8) was added to the side of the vial, which was then slowly tilted until the drop of NaOH met the surface of the liquid containing components 1-4.Agel immediately formed where the base was added. One of the vials was inverted, and the thin gel layer was able to support the liquid above it (Figure 3d). The polymerization fronts propagated with ac onstant velocity until the entire reaction mixture switched to ahigh pH value at the end of the induction period (Figure 3d-f). Ascending fronts (Figure 3g,black line) had aslightly higher speed than the descending fronts (Figure 3g,green line).
Thus,t he liquid can be injected into ac avity in any orientation, and ar eaction can be initiated to form ap lug at the open end and is then propagated as afront to cure the rest of the mixture.T he coupling of the base-catalyzed thiol-Michael addition reaction with the autocatalytic enzyme reaction results in an ew method for frontal polymerization that does not require ac onstant external perturbation to be maintained or involve large temperature gradients.
Considerable interest in PEG-acrylate hydrogels revolves around their use as degradable biomaterials,f or example,i n drug and cell delivery and as scaffolds in tissue repair. [27] A similar PEG-based hydrogel to the one constructed in this study was demonstrated to undergo hydrolytic degradation over the course of weeks,t hus resulting in controlled drug release. [23] More recently,amethod for the pH control of the self-assembly and disassembly of peptide hydrogels was proposed for fluidic guidance in channels and self-erasing rapid prototyping. [28] Thet hiol-acrylate hydrogels formed in this study have both an inbuilt time to gelation and an inbuilt time before complete degradation (Figure 4a). Degradation proceeded slowly,a nd the mixture returned to the liquid state.B asecatalyzed ester hydrolysis provides ac onvenient method for the control of gel degradation. While the time before gelation is mainly governed by the components of the urea-urease reaction, the degradation time also depends on the gel strength and hence the precursor concentrations.Gel strength was followed by dynamic rheometry.Arapid increase in the storage modulus G' was observed after al ag phase (Figure 4b). With an increase in the ETTMP concentration, the maximum G' value increased (Figure 4b,b lack and red curves), and the degradation time increased. An increase in the ETTMP concentration increases the cross-linking density but also decreases the final pH value,t hus additionally slowing the rate of base-catalyzed hydrolysis.Adecrease in urea concentration also resulted in an increase in G' and as lower degradation rate (Figure 4b,r ed and green curves) because of the lower final pH value and higher polymer conversion associated with the longer induction time.
Thet ime for the gel to return to the liquid state varied from 5h to over 20 weeks (Figure 4c,d). Fast degradation times were favored by ah igh final pH value and low gel strength:hence,high urea and low thiol concentrations.Inthe examples shown, the degradation time was correlated with the induction period;h owever,i tm ay be possible to independently vary these characteristic timescales through simultaneous variations in two of the control variables: enzyme,substrate,and acid.
Herein we have shown how the amplification of achemical signal might be translated into ap hysical response:a n autocatalytic enzyme reaction was used to drive time-lapse gelation and frontal polymerization. Thegel lifetime was also controlled through the initial concentrations of the components of the enzyme reaction and the thiol. Thec oupling of autocatalytic reactions with physical processes has generated pulses of precipitates, [29] bioinspired chemomechanical devices, [30] thiol-acrylate microparticles, [31] and periodic nanoparticle aggregation; [32] however, these systems involved harsh chemicals that limit their use in applications.W eused an enzyme-catalyzed reaction with aw ater-soluble thiol and acrylate to create ag elation process that operates under ambient, aqueous-phase conditions.
Our system does not require radical initiators or ah igh temperature but operates on the basis of an inbuilt pH switch. Other autocatalytic enzyme reactions,s uch as the glucoseoxidase reaction, involve base-to-acid switches that might be used in conjunction with acid-catalyzed polymerization. [33] This systems-chemistry approach to transient gelation has numerous attractive features for bioinspired, biocompatible materials applications.