Direct Ink Writing of Enzyme‐Containing Liquid Crystal Elastomers as Versatile Biomolecular‐Responsive Actuators

The combination of enzymes and chemically responsive smart materials presents considerable opportunities in anti‐fouling, sensing or medicine. Among stimuli‐responsive polymers, liquid crystal elastomers (LCEs) have emerged as a leading actuator platform due to their inherent programmability, realizing large shape transformations upon exposure to stimuli. Herein, three different hydrolytic enzymes are immobilized into acid‐responsive LCEs, which impart the LCEs with sensitivity to three distinctive classes of biomolecules: lipids, carbohydrates, and peptides. Dye‐doped biocatalytic LCEs readily switch their color and shape upon exposure to the appropriate substrate. Twisted nematic patterning and direct ink writing are used to showcase the versatility of shape changes the enzyme‐containing LCEs may undergo. Multiplexed responsiveness is also demonstrated using a connected array of 3D printed disks, each containing a different enzyme, highlighting the excellent chemical selectivity of the LCEs. This work presents a novel platform of versatile bioresponsive color‐switchable actuators that may have application in a wide range of fields.


Introduction
Stimuli-responsive materials that respond to various signals are at the forefront of modern material science and have served as the foundation for numerous advances in diverse applications, including smart textiles and autonomous systems. [1][2][3] Of particular interest within the field of stimuli-responsive materials is the development of smart materials that undergo color or shape transformations in response to chemical stimuli. [4][5][6] While the limitation, the use of alternative shape reconfigurable materials as supports for the immobilization of enzymes is an attractive direction of research.
As an alternative to traditional hydrogels, liquid crystal elastomers (LCEs) have shown tremendous promise as matrices for enzyme-driven bioresponsive actuators. [21,22] LCEs are soft materials that self-organize at the molecular level and possess intrinsic anisotropic properties that enable them to undergo shape reconfigurations. [23][24][25] Because the final shape is governed by the liquid crystal nematic director, LCEs are readily programmed to acquire different shapes by tailoring the director field during synthesis. [26,27] We have recently reported the immobilization of urease into LCEs that were designed to undergo mechanical actuation in response to the presence of urea. [22] The basis for this actuation was the generation of ammonia by the urease-catalyzed conversion of urea, which resulted in the disruption of supramolecular hydrogen bonding within the material. Nonetheless, the range of feasible morphologies was relatively modest due to the incompatibility of supramolecular LCEs with surface alignment or additive manufacturing techniques. Additionally, given that these materials were designed to respond to base, the scope of enzymes that could be used to induce actuation was rather limited. In contrast, we hypothesized that because a much larger number of hydrolytic enzymes convert substrates into acidic products, the range of biochemical stimuli that can induce actuation in enzyme-containing materials may be greatly expanded by utilizing biocatalytic LCEs that are responsive to acid instead of base. Moreover, the immobilization of enzymes in LCEs that are amenable to surface alignment and additive manufacturing techniques, such as direct ink writing, may further increase the scope of potential applications for bioresponsive LCEs.
In this work, we investigated the versatility of LCEs as a general platform for biomolecular recognition and actuation by immobilizing a diverse group of enzymes in acidresponsive LCEs. The LCEs used in this work were not only responsive to acid but were also compatible with surface alignment techniques and 3D printing, which enabled the engineering of complex shape transformations. Importantly, the acidresponsive nature of the LCEs permitted the use of chemical signal molecules belonging to multiple classifications of biomolecules, including carbohydrates, lipids, and peptide-like signals. The versatility of such materials was specifically demonstrated via the covalent conjugation of glucose oxidase (GOx), lipase (Candida Rugosa; CRL), and trypsin (TRP) within the LCE network. Glutaraldehyde was used as a crosslinker to enable robust covalent conjugation of each enzyme, and the LCEs were doped with a pH responsive dye that provided a visual color change as a complement to the mechanical actuation in response to the production of acid. Using direct ink writing, we further demonstrated the selectivity of both the color change and shape transformation by fabricating monolithic arrays of biocatalytic disks, each of which responded to a particular signal. Ultimately, such materials that change color and shape in response to targeted biological molecules will have widespread utility in diverse applications, including drug delivery, antifouling surfaces, and smart surgical devices.

Enzyme Immobilization and Characterization in Acid-Responsive LCEs
Acid-responsive LCEs were first prepared by aza-Michael oligomerization of the liquid crystal monomer RM82 using the chain extension agent N,N-dimethylethylenediamine (DMEN) ( Figure S1, Supporting Information), as described by Kim and coworkers. [28] Using 1H NMR end-point analysis, we determined that the number of LC monomers per oligomer was approximately six ( Figure S2, Supporting Information). The films were washed after polymerization and their weight did not change, suggesting complete polymerization of the elastomer. The pendant tertiary amines in DMEN may be protonated under acidic conditions, generating a cationic hygroscopic polymer. Given the anisotropic stiffness typical of LCEs, the films exhibited a larger expansion in the direction perpendicular to the nematic director upon swelling. Such anisotropic swelling enables reconfiguration of the shape of the material, with the final shape being governed by the original director field. [29,30] As shown by FTIR analysis of the LCE films, the tertiary amines in DMEN became protonated below pH 5, which could then be reversibly deprotonated when the pH was raised above 7.5 ( Figure S3, Supporting Information). Accordingly, we hypothesized that the production of acid by an enzyme immobilized in the LCE films would lead to actuation in a predictable and programmable manner ( Figure 1A). This was demonstrated using GOx, CRL, and TRP, each of which generate an acidic product via the conversion of their respective substrates ( Figure 1B). These enzymes were specifically chosen given their activity toward different types of biomolecules, including carbohydrates, lipids, and peptides. By exhibiting the responsiveness of the LCE films to a broad range of biomolecules, we sought to demonstrate the versatility of such materials as a general platform for biomolecule responsiveness.
Upon synthesis of the acid-responsive LCE films, GOx, CRL, and TRP were individually immobilized in the films via glutaraldehyde conjugation ( Figure 1C). [31,32] To facilitate reaction with glutaraldehyde, the enzymes were initially physically adsorbed to the LCE network using a slightly acidic buffer (pH 5.5). The enzyme-containing films were then exposed to a mild glutaraldehyde solution, which resulted in the irreversible crosslinking and immobilization of the enzyme. Following treatment with glutaraldehyde, the films were washed vigorously to remove any non-crosslinked enzyme. The results of activity assays indicated that washing removed all of the enzyme that was not covalently immobilized. Notably, enzyme immobilization using glutaraldehyde may involve a variety of mechanisms, and the specific mechanism or mechanisms are not always known for a particular case. [33,34] While it is plausible in this work that the enzyme reacted with residual secondary amines from oligomerization during formation of the network, the enzyme may also have crosslinked with itself. [35][36][37] This in turn could result in a pseudointerpenetrating network whereby the enzyme was intertwined within the LC network, which could explain the apparent irreversible immobilization.
Attachment of the enzymes to the LCE films was confirmed by the appearance of a shoulder at 1660 cm −1 in the ATR-FTIR spectra of the films, corresponding to the amide I protein band ( Figure S4A, Supporting Information). Additionally, confocal imaging of fluorescently tagged GOx confirmed that the incorporation of enzyme was mostly localized to the surface region of the film ( Figure 1C). We found that approximately 75% of the total immobilized enzyme was localized within 5 μm of the LCE surface ( Figure S5, Supporting Information). There are two potential reasons as to why enzyme immobilization was mainly limited to the surface. First, given that the long oligomeric chains can pack in a very efficient manner due in part to pi-pi stacking between the different liquid crystal moieties, it is reasonable to think that the void space inside the network was small and therefore a large macromolecule like an enzyme may not be able to diffuse as well as in other elastomeric networks with larger void spaces (such as those created by a chain transfer mechanism). Second, we believe that the electrostatic interactions that retained the enzymes in place during the first step of the immobilization process happened mostly on the surface of the films due to a surface charge effect, therefore concentrating the enzymes in these regions of the film. We further found that the loading of GOx increased approximately linearly with the enzyme concentration in the loading reaction up to 1 mg mL −1 , after which enzyme loading appeared to saturate ( Figure S6A, Supporting Information). This behavior was likely due to the saturation of available immobilization sites on the surface of the films. Using a concentration of 1 mg mL −1 of enzyme, the loading of the three enzymes in the LCEs ranged from 2 to 6 mg m −2 , which corresponded to approximately one monolayer for each enzyme on the LCE surface ( Figure S6B, Supporting Information). [38,39] Notably, enzyme loading was determined by measuring the biocatalytic activity of the films and assuming that the specific activity of the free and immobilized enzymes was approximately equal.
Having confirmed that the enzyme-containing LCE films were biocatalytically active, the impact of the production of acid on the structure and mechanical properties of the LCEs was investigated. Analysis of the films via ATR-FTIR showed that the tertiary amines within the enzyme-containing LCE films were protonated upon exposure of the films to the substrate for each enzyme ( Figure S4B, Supporting Information). Polarized optical microscopy was used to characterize the GOx-containing LCE films, and showed that the film birefringence was retained after enzyme immobilization and exposure to glucose ( Figure S7, Supporting Information). The retention of birefringence suggested that the nematic alignment of the films was unaffected by either the enzyme incorporation process or protonation of tertiary amines within the films. Additionally, the films were also elastomeric, as shown by the low glass transition temperature (T g = −1.5°C) and the low elastic moduli (E ≈ 15 MPa and E ≈ 3 MPa in the directions parallel and perpendicular to the director, respectively) ( Figure S8, Supporting Information). Compared to glassy, stiffer formulations, LCEs generally have higher liquid crystal chain mobility which can increase the magnitude of the shape change response; however, a lower stiffness also implies that this formulation may generate smaller forces than other densely crosslinked networks. While the GOx-containing LCE films did not exhibit significant changes in their stiffness ( Figure S8A, Supporting Information) or T g ( Figure S8B, Supporting Information) after enzyme incorporation, the stiffness and glass transition of the films did increase after exposure to glucose. A similar effect was observed when control films with-out enzyme were exposed to buffer at pH 3 ( Figure S8A, Supporting Information). This effect was likely due to the restriction of the polymer chain mobility by the cationic repulsion between protonated amines in the films after conversion of substrate (and generation of acid). Given this, it is likely that the films would undergo a reversible softening upon exposure to a basic buffer and successive deprotonation of the amines.
Because the response of the LCE films requires the accumulation of acid to protonate amines in the films, it is crucial that the immobilized enzymes retain their activity under moderately acidic conditions. Given this requirement, the pH stability of the enzyme-containing films was characterized by assaying the relative activity of the biocatalytic LCEs as a function of pH. Figure 2 shows the pH-dependent activity profiles of immobilized GOx, CRL, and TRP. Remarkably, all three enzymes were considerably more stable at low pH upon immobilization in the LCEs relative to the soluble (i.e., non-immobilized) forms of each enzyme. In the case of GOx (Figure 2A), the immobilized form of the enzyme retained approximately 90% of its activity at pH 4 with respect to its pH optimum, while its pH optimum also shifted to approximately pH 5 (from pH 6.5 for soluble GOx). The increase in pH stability was even more dramatic for immobilized TRP ( Figure 2B), which also retained approximately 40% of its activity at pH 4 with respect to its pH optimum despite the soluble form of TRP being mostly inactive below pH 5. Additionally, although the difference between the relative activity of immobilized and soluble CRL was much less than for TRP, immobilized CRL was also active down to pH 4 ( Figure 2C). The increased pH stability of the immobilized enzymes may be attributed to a surface charge effect in which the charged amines in the LCE films effectively buffered the local pH of the enzymes (i.e., by attracting negatively charged hydroxyl ions and repelling protons). [40] Such an effect has been reported previously for enzymes conjugated to positively and negatively charged supports [41][42][43][44] as well as upon modification of enzymes with charged polymers. [45,46] These results confirmed that GOx, CLR, and TRP remained active under the acidic conditions necessary to trigger a shape transformation within the LCE films.

Methyl Red as a Visual pH Indicator in LCEs
Organic dichroic dyes such as methyl red are frequently used in liquid crystal applications due to their affinity to and compatibility with many liquid crystalline networks. [47,48] Because such dyes often preferentially align along the liquid crystal director, they have been used to induce photoalignment [47] and approximate the nematic order parameter. [49] We found that methyl red was highly miscible with acid-responsive LCE films and was readily absorbed into the LCE network. While methyl red is a diazo dye that may undergo cis-trans transitions under UV or near-UV light, wet and dry methyl red-doped films were stable over several weeks exposed to ambient light, showing no color or shape change. Notably, using polarized UV/vis spectroscopy, we found that methyl red aligned with the mesogens in the LCE network; we estimated the nematic order parameter of the films to be 0.46, which is typical of planar uniaxial LC networks ( Figure S9, Supporting Information). Because methyl red is pH sensitive, absorption of the dye into the LCE network also served as local pH indicator within the films. As such, the presence of the dye provided a way to visualize the production of acid by immobilized GOx, CRL, and TRP and thus served as a proxy for enzyme activity. The ability to visualize changes in local color within the films as a function of pH was demonstrated by incubating methyl red-doped LCEs in buffer between pH 2.5 and 6.5. Figure 3A and Figure S10A, Supporting Information, show the change in film color, which varied from bright yellow at pH 6.5 to deep red/pink at pH 2.5. Additionally, if the exposure to acid was not homogeneous throughout the film, the color change was highly confined to the location of the chemical stimulus within the LCE ( Figure  S11, Supporting Information). Interestingly, the color change was also reversible, with the films showing a strong color change going from basic to acidic buffers for at least 10 cycles ( Figure  S10B, Supporting Information). Further analysis of the UV/vis spectra of the methyl red-doped LCEs showed two characteristic absorption peaks at 415 nm and 515 nm, which corresponded to the basic (yellow) and acidic (red) forms of the dye, respectively ( Figure S10C, Supporting Information). The presence of distinct absorption peaks for the basic and acid forms of the dye was particularly useful for quantifying the local pH within the films.
To quantify the pH change inside the films, the parameter was defined as the logarithmic ratio of the absorption peak intensities for the acidic and basic forms of methyl red. Based on this definition, a value of = 1 corresponded to all dye in the acidic state while a value of = 0 corresponded to all dye in the basic state. Therefore, an increase in indicated a decrease in local pH in the films and vice versa. Using this parameter, we measured the apparent pKa of methyl red within the films as 4.2, which corresponded to the buffer pH at = 0.5 ( Figure 3A). Notably, the apparent pKa of methyl red in the films was approximately 1 pH unit lower than that in solution (5.1). This shift in apparent pKa was most likely due to the same surface charge effect that increased the pH stability of immobilized enzymes as shown in Figure 2. Specifically, the buffering effect within the films presumably led to an effective increase in local pH relative to the bulk pH of the solution, which was directly reflected in the behavior of the absorbed methyl red. [50] To illustrate the use of methyl red to visualize the production of acid, the evolution of was monitored upon exposure of GOxcontaining LCE films to glucose. For a fixed response time of 6 h, the value of increased linearly with glucose concentration up to 10 mm glucose after which the response plateaued ( Figure 3B). Notably, the films displayed a response when exposed to as little as 1 mm of glucose. Similarly, the value of increased linearly as a function of response time when exposed to excess glucose (50 mm) until approximately 4 h, after which the response leveled off with time ( Figure 3C). We hypothesized that the saturation of the response beyond 10 mm glucose and after 4 h was due to the surface pH reaching a pH of approximately 3 at which point GOx was likely mostly inactivated. Thus, the adsorbed methyl red provided an optical readout of glucose conversion, both visually and spectroscopically.

Mechanical and Optical Response to Diverse Biomolecules
The combined mechanical and optical response of GOx, CRL, and TRP-containing LCEs to diverse biomolecules was explored using twisted nematic LCE films. The twisted nematic alignment of the films, which were prepared via surface alignment, permitted the 3D shape transformation of the films from a flat sheet into a curl. In this case, the curling behavior was due to an expansion gradient through the thickness of the films upon swelling, [51,52] which happened when the tertiary amines became protonated. By doping the films with methyl red, the color of the films also underwent a transition from yellow to red upon addition of acid such that both a mechanical and optical response was observed. The simultaneous mechanical and optical response of the films upon addition of acid is shown in Movie S1, Supporting Information.
Subsequently, twisted nematic dye-doped LCEs containing GOx, CRL, or TRP were exposed to glucose, tributyrin, or benzoyl-l-arginine ethyl ester, respectively. The films simultaneously curled and turned from yellow to red only in the presence of the appropriate substrate (Figure 4). No response was observed in control films without enzyme. Notably, while all three enzymes triggered both a mechanical and optical response, the speed of the response for the enzyme-containing LCE films was markedly different for each enzyme. This can be seen in Figure S12, Supporting Information, where films containing CRL and TRP appeared to curl fully by 1-2 h after addition of tributyrin and benzoyl-l-arginine ethyl ester, respectively, while films containing GOx did not fully curl until 3-4 h after addition of glucose. This difference was likely due to differential enzyme loading, given that enzyme loading in films with CRL and TRP was approximately three times greater than in the films with GOx ( Figure S6B, Supporting Information). Additionally, the slower response may be explained by the requirement for gluconolactone, the product of the enzyme, to be hydrolyzed to gluconic acid. It is also plausible that differences in the strength or stoichiometry of the acid produced by the enzymes played a role in the variability of response speed of the films. For example, benzoyl-l-arginine (pKa = 3.31) [53] is a stronger acid than gluconic acid (pKa = 3.72), [54] and CRL produces three moles of butyric acid per mol of substrate, while the stoichiometry is 1:1 in the other two cases. For applications in which the response speed is key, the use of alternative immobilization chemistries that allow for the diffusion of the enzyme into the network might increase the total amount of enzyme loading, therefore accelerating the response. Moreover, it is likely that confinement of the enzyme to the surface slowed the deformation, given that the acidic product must diffuse inside the network to trigger the protonation of the amines. However, given the small thickness of the film, this may not be a significant factor compared to others in this particular case.

Direct Ink Writing of Biocatalytic LCEs
4D printing has recently emerged as an additive manufacturing technique in which the 3D printed objects, such as shape memory polymers or hydrogels, evolve over time due to their stimulus response (hence the fourth dimension). [55,56] Notably, LCEs have been shown to be amenable to 3D printing techniques such as direct ink writing, in which an LC oligomer ink is extruded through a nozzle and rapidly crosslinked to   Pictures of methyl red-doped twisted nematic enzyme-containing LCEs after exposure to substrates for 20 h at 37°C. The LCE containing GOx (top) was exposed to 50 mm glucose and 10 mm potassium chloride. The LCE containing CRL (middle) was exposed to 50 mm tributyrin, 10 mm potassium chloride, 5 mm calcium chloride, and 2% gum arabic. The LCE containing TRP (bottom) was exposed to 50 mm benzoyl-l-arginine ethyl ester, 10 mm potassium chloride, and 5 mm calcium chloride. For each substrate, the solution was adjusted to pH 6.0 before use. For comparison, control films without enzyme that remained flat upon addition of substrate are also shown (left side).
Adv. Mater. Interfaces 2023, 10, 2300086   2300086 (8 of 12) www.advancedsciencenews.com www.advmatinterfaces.de retain the mesogenic alignment. [28,[57][58][59] The printing of stimuliresponsive polymers for use in biomedical applications such as tissue engineering and drug delivery has been an important venue for such materials. [60][61][62] For example, previous efforts have demonstrated the use of 3D printed LCEs as urological [63] or intervertebral [64] implants. Because biological processes in the human body often depend on feedback loops which are controlled by small biomolecules such as carbohydrates, lipids or polypeptides, the printing of materials that respond to such chemical signals with high specificity is highly desirable.
As a demonstration of 3D printed biocatalytic LCEs, circular disks, which displayed good nematic alignment as observed via polarized optical microscopy ( Figure S13, Supporting Information), were fabricated and doped with methyl red. Notably, the dye was also preferentially aligned in the direction of the liquid crystal ink, as shown in the dichroism observed when a dye-doped disk was examined under crossed polarizers (Movie S2, Supporting Information). Upon exposure to an acidic buffer, the flat yellow disks transformed into red cones. The disk-to-cone shape reconfiguration can be explained by the anisotropic swelling typical of +1 azimuthal defects. Upon protonation of the tertiary amines, the swollen disks underwent an out-of-plane deformation to accommodate the new perimeterto-radius ratio, which facilitated the formation of a cone. [65,66] To test the biocatalytic responsiveness of 3D printed LCEs, we first immobilized CRL in a 3D printed disk. After exposure to tributyrin, the disk simultaneously changed color and shape, switching from yellow to red and from a flat disk to a cone ( Figure 5A). Interestingly, the color change was fully reversible after immersion in a basic buffer, while the shape change was mostly irreversible ( Figure S14, Supporting Information) as observed previously in swelling-driven disk-to-cone transitions. [66] Considering the adaptability of the LCEs in the response to diverse biomolecules, it was intriguing to investigate an approach in which a single material could respond to more than one biomolecule in a distinct and predictable manner. Taking advantage of the versatility of 3D printing, we designed an array of three disks that were physically connected. After doping the array with methyl red, a different enzyme (GOx, CRL, or TRP) was immobilized in each of the circles by carefully depositing a solution of each enzyme on top of each circle. To test for selectivity, all three disks were simultaneously exposed to benzoyl-l-arginine ethyl ester as shown in Figure S15, Supporting Information. While CRL also has esterase activity, only TRP recognizes peptide-like substrates, which meant that only the disk that contained TRP responded by changing shape and color ( Figure 5B). To further demonstrate the specificity of the response of the 3D printed disks, the remaining two disks in the array were subsequently exposed to glucose. As expected, the CRL-containing disk did not respond, while the disk with GOx did respond. Finally, the disk containing immobilized CRL was exposed to tributyrin, which also elicited a mechanical and optical response. Taken together, these results demonstrate that the enzyme-mediated responsiveness of LCE films is also applicable to 3D printed LCEs, which in turn enables seemingly endless designs that may be imagined using this additive manufacturing technique.

Conclusions
This work demonstrates the ability of immobilized enzymes to sensitize elastomeric actuators to a wide range of biological signals. In applications in which the specificity of the stimuli response of the material is crucial, such as drug delivery and sensing, the immobilization of enzymes into these materials provides a pathway to obtain high specificity and selectivity. Moreover, 3D printed LCEs and other materials amenable to additive manufacturing techniques remain potential candidates for use in highly personalized biomedical applications. Biocatalytic 3D printed LCEs may have utility as implantable smart devices: for example, a coronary stent could be fabricated that responds to the presence of cholesterol and other fatty plaque components by self-expanding, therefore increasing the blood flow only when the material senses the presence of plaque. In another example of a biomedical application, miniaturized drug devices may be fabricated, with the lid being controlled by an enzyme such as GOx. Upon reaching a threshold concentration of glucose, the lid would open, releasing the cargo (i.e., insulin). This platform may also have use in anti-fouling applications. Surfaces fabricated using 3D printed LCEs could be designed to respond to specific quorum signals (e.g., acyl-homoserine lactones) released by biofilmforming bacteria, forming wrinkles or other structures to inhibit biofilm formation while simultaneously changing color. Most importantly, by showing the response could be triggered by signals belonging to three different classes of biomolecules, the high versatility of this platform was demonstrated.

Experimental Section
LCE Synthesis and Enzyme Immobilization: Acid-responsive LCE films for enabling enzyme immobilization were prepared as described by Kim and co-workers. [28] To prepare cells for making LCE films, glass slides were initially plasma treated (Henniker HPT-100) for 10 min. The slides were then spin-coated with a 0.125% weight Elvamide in methanol solution and rubbed 30× with a red velvet cloth either in antiparallel directions for uniaxial cells or in perpendicular directions for twisted nematic cells. Following rubbing, the slides were subsequently adhered using a UV-cured glue with 50 μm spacers and stored at room temperature. For the synthesis of LCE films, C6M (1,4-bis-[4-(6-acryloyloxyhexyloxy)benzoyloxy]−2methylbenzene) and DMEN (N,N-dimethylethylenediamine) were added in a 1.05:1 molar ratio with I-651 (2,2-Dimethoxy-1,2-diphenylethan-1-one, 1.5% wt) as initiator. The components were melt-mixed and added into an LCE cell at 105°C. The cell was then incubated for 24 h at 65°C and finally exposed to 365 nm UV light (Spectroline EA-160 6 W, 30 cm away from samples) for 30 min at room temperature. The resulting LCEs were extracted from the cells and stored dry at room temperature until use.
For enzyme immobilization, glucose oxidase from Aspergillus niger, lipase from Candida rugosa, and trypsin from bovine pancreas were purchased from Sigma-Aldrich and purified using a desalting column (Bio-Scale Mini Bio-Gel P-6 Desalting Cartridge) prior to use. To enable immobilization, 200 μL of a solution of each enzyme (1 mg mL −1 in 25 mm MES, pH 5.5) was incubated with the LCE films for 1 h at room temperature. After washing the films five times on each side with ultrapure water, the LCEs were incubated in 500 μL of a 0.5% wt glutaraldehyde solution in 10 mm potassium phosphate (pH 7.0) at 4°C for 16 h. The films were then washed with DI water (25 mL), incubated in 5 mL of 50 mm HEPES (pH 8.5) at 4°C for 24 h under mild shaking, washed with DI water (25 mL) again, and used immediately. The films were washed until leaching of the enzyme stopped, which was monitored by measuring enzyme activity in the wash solution. No enzyme activity was detected during the final Figure 5. Direct ink writing of biocatalytic LCEs. A) Representative image of an enzyme-containing flat disk transforming into a cone upon addition of substrate. The image was obtained from a 3D printed LCE containing CRL, which was subsequently exposed to 50 mm tributyrin, 10 mm potassium chloride, 5 mm calcium chloride and 2% gum arabic for 20 h at 37°C. The radius of the disk was 5 mm, while the height of the final cone was ≈4 mm. B) Array of 3D printed connected disks with each containing a different immobilized enzyme. The three disks were simultaneously exposed to 50 mm benzoyl-l-arginine ethyl ester, 10 mm potassium chloride, and 5 mm calcium chloride (top row). Following the initial exposure to the substrate for TRP, the remaining two disks were exposed to 50 mm glucose and 10 mm potassium chloride. Finally, the last remaining disk was exposed to 50 mm tributyrin, 10 mm potassium chloride, 5 mm calcium chloride, and 2% gum arabic. For each step, the disks were incubated with substrate for 3 h at 37°C. The radius of each disk and height of the cones was 5 mm and ≈1.5 mm, respectively. washing step. For 3D printed LCEs, the enzymes were immobilized by carefully depositing 250 μL of enzyme solution on top of each disk for 1 h at room temperature, wiping away the excess solution, then dropping 250 μL of glutaraldehyde solution and allowing it to react for 1 h at room temperature. After conjugation of the enzyme, the 3D printed LCEs were washed as described previously.
Enzyme Activity: To measure the biocatalytic activity of the enzymecontaining LCEs, films were cut into 5×5 mm squares and incubated in 4 mL of buffer with substrate at 25°C under vigorous shaking. For glucose oxidase, the assay solution contained 50 mm glucose and 10 mm potassium phosphate (pH 7.0). For lipase, the assay solution contained 10 mm 4-nitrophenyl butyrate, 10 mm HEPES (pH 7.5), and 5 mm calcium chloride. For trypsin, the assay solution contained 10 mm l-BAPNA (N-benzoyl-l-arginine-4-nitroanilide), 10 mm HEPES (pH 7.0), 5 mm calcium chloride, and 10% DMSO. The conversion of substrate was measured by periodically removing 100 μL aliquots from the assay solution for each enzyme. For lipase and trypsin, the absorbance from the aliquots at 400 nm was directly measured without addition of other reagents using a Tecan Infinite M Plex microplate reader. For glucose oxidase, horseradish peroxidase (Sigma-Aldrich, purified using size-exclusion chromatography) and ABTS (2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)) were added to each aliquot to a concentration of 20 nm and 2 mm, respectively. After incubation of the aliquot with horseradish peroxidase and ABTS for 5 min at room temperature, the absorbance of the solution at 415 nm was recorded. Enzyme activity was determined from the slope of absorbance versus time while using standard curves prepared with known concentrations of ABTS· (obtained using a hydrogen peroxide standard), 4-nitrophenol, and 4-nitroaniline. The activity of the soluble form of each enzyme was determined using the same assay buffers.
To measure the activity of the films as a function of pH, citrate buffers (10 mm) ranging from pH 4 to pH 7 were used. For lipase-containing films, the pH of the aliquots was adjusted to 7.5 before measuring absorbance, given that the absorbance of 4-nitrophenol is heavily reduced in acidic pHs. Since l-BAPNA is not compatible with citrate, 10 mm acetate (pH 4-5.5) and MES (pH 5.5-7) buffers were used instead for the trypsin-containing films. The effects of each buffer on enzyme activity were accounted for by normalizing enzyme activity in each buffer at pH 5.5. Notably, the background hydrolysis of the substrate for each reaction was negligible over the pH range used.
Enzyme Loading and Imaging: Enzyme loading into the films was estimated by measuring the biocatalytic activity of the films under the assumption that the specific activity of the soluble and immobilized enzymes was approximately equal. To calculate the specific activity of the enzymes, their activity was normalized by the total amount of enzyme, which was measured using the MicroBCA assay (Thermo Scientific). Additionally, enzyme loading was imaged by laser scanning confocal microscopy. For this, glucose oxidase was labeled with Oregon Green 488 Maleimide (Fisher Scientific) prior to immobilization by adding 15 μL of the fluorophore (20 mm in DMSO) to 500 μL of glucose oxidase at a concentration of 1 mg mL −1 in 50 mm HEPES (pH 7.5). The mixture was incubated at room temperature for 1 h after which any unreacted fluorophore was removed using a 5 mL desalting column. After glutaraldehyde-mediated immobilization of labeled glucose oxidase, the films were placed on a glass slide and imaged with a Nikon A1R Laser Scanning Confocal microscope equipped with a 488 nm EGFP laser and a 20× objective. Z-stacks were obtained by imaging a z-range of approximately 60 μm using a step size of 1 μm.
Methyl Red Absorption and UV-Vis Quantification: To enable quantification of the local pH of the films, LCEs were exposed to a solution of 0.01% methyl red in potassium phosphate 10 mm (pH 7.0) for 1 h at room temperature. Excess dye was removed by incubating the films in fresh phosphate buffer for 1 h with mixing. Using a Cary 7000 UV-vis spectrophotometer, the ratio of absorbance of the acidic ( = 515 nm) and basic ( = 415 nm) forms of methyl red was measured. This ratio was used to define the parameter that ranged from = 0 at pH 7 to = 1 at pH 1.5 using the following equations: In the equation for , ∑ max , and ∑ min correspond to the values of ∑ at pH 1.5 and pH 7.0, respectively. Given that represented the approximate proportion of the acidic form with respect to the basic form of methyl red, the pKa of methyl red in the film could be estimated by fitting versus pH in the pseudolinear region of the curve and calculating the pH value in which = 0.5. To measure an approximate dichroic ratio R and nematic order parameter S, UV-vis spectra of methyl red-doped LCEs were gathered using a polarizer to measure the absorbance at 415 nm in both parallel and perpendicular directions with respect to the director field. The following equations were used, with A ∥ and A ⊥ representing the absorbance at 415 nm parallel and perpendicular to the director, respectively: 4.0.0.1. Printing of LCEs: The LC ink was prepared by melt-mixing C6M and DMEN in a 1.05:1 molar ratio together with I-651 (1.5% wt) and butylated hydroxytoluene (BHT, 0.1% wt), and incubating the mixture at 65°C for 24 h under heavy mixing to allow the monomers to oligomerize. For 3D printing, a Hyrel3D System 30 M printer with a KR2 printhead and 0.4 mm tapered nozzle was used. Samples were printed onto glass slides coated with a thin layer of PVA to serve as a sacrificial release layer. A print speed of 6 mm s −1 , a target nominal layer height of 0.15 mm, and printhead temperature of 60°C was used for all samples. Crosslinking was achieved by exposure to 365 nm light at 3 mW cm −2 during printing followed by exposure to 365 nm light at 75 mW cm −2 for 10 min after printing.
Exposure of LCEs to Biological Signals: Twisted nematic LCEs with and without enzyme were exposed to a solution (200 μL) containing substrate in water at 37°C. The substrate solution contained 50 mm glucose and 10 mm potassium chloride for LCEs with immobilized glucose oxidase; 50 mm tributyrin, 10 mm potassium chloride, 5 mm calcium chloride, and 2% gum arabic for LCEs with immobilized lipase; and 50 mm N -Benzoyll-arginine ethyl ester (BAEE), 10 mm potassium chloride and 5 mm calcium chloride for LCEs with immobilized trypsin. For each substrate, the solution was adjusted to pH 6.0 before use. Pictures of the twisted nematic films were taken with a Canon EOS M50 camera. For 3D printed disks, the LCEs were exposed to the appropriate substrate by carefully pipetting 250 μL of solution on top of the disks. 3D measurements of the circle arrays were taken with a Keyence VR-3000 Wide-Area 3D Measurement System using a 2× height magnification.
Chemical and Mechanical Characterization of LCEs: The chemical structure of LCEs was characterized by attenuated total-reflection Fourier transform infrared spectroscopy (ATR-FTIR) with a Nicolet iS20 spectrometer, using an average of 32 scans. The oligomerization process of the LC ink was characterized by a Bruker AvanceCore 1 H NMR, using 64 scans. For dynamic mechanical analysis, monodomain uniaxial LCE samples were cut into strips ≈15 mm × 3 mm × 0.05 mm (length × width × thickness), loaded with a gap of 5 mm, and strained at a rate of 5%/min at room temperature using a TA Instruments Discovery DMA 850. Samples were cut both parallel and perpendicular to the nematic director for these measurements. Additionally, for analysis of the films via polarized optical microscopy (POM), LCEs were imaged while aligned parallel (0°) and perpendicular (45°) to the crosspolarizers using a Nikon Eclipse Ci-Pol microscope. Furthermore, for differential scanning calorimetry (DSC) measurements, the glass transition of LCE films was measured using a TA Instruments Discovery DSC 2500 with a standard heat-cool-heat cycle in which temperature was cycled between −50 and 150°C with a ramp rate of 10°C min −1 .
Statistical Analysis: Two-tailed t-tests were used to obtain the p-values for all indicated results. Differences were considered significant when p ≤ 0.05 was obtained. ns, *, **, and *** represent p > 0.05, p ≤ 0.05, p ≤ 0.01, and p ≤ 0.001, respectively. Statistical analysis was performed using OriginPro software.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.