Bienzymatic Generation of Interpenetrating Polymer Networked Engineered Living Materials with Shape Changing Properties

The synthesis of a porous shape‐changing interpenetrating network (IPN) bioink for the fabrication of large‐scale (cm) reversibly thermosensitive structures is described. The poly(N‐isopropylacrylamide) (PNIPAm) IPN is generated in situ within an ionically crosslinked alginate hydrogel at room temperature and under aerobic conditions using a horseradish peroxidase (HRP)/glucose oxidase (GOx) bienzymatic initiation system. Mechanical testing assessment of the IPN hydrogels confirm mechanical reinforcement via covalent single network interdigitation. Furthermore, the thermosensitive bioink can be used to print biohybrid reactors containing genetically engineered phosphotriesterase‐expressing E. coli capable of hydrolyzing toxic organophosphorus compounds. Herein, increasing the bioink pore size using the contractile‐thermosensitive response of the IPN improves the temperature‐dependent theoretical mass‐transfer‐limited enzyme catalyzed reaction rate, providing a plausible route to externally regulated enzymatic catalysis within bioprinted structures.


Introduction
[3][4][5][6] The 3D printing of shape changing (or shape Figure 1.Schematic showing the 3D printing of a shape-changing enzyme-mediated interpenetrating network (IPN) bioink.a), The bioink contains the pre-gel elements NIPAm, MBA, and sodium alginate, and is cured via crosslinking with a solution containing CaCl 2 and glucose.The alginate chains are ionically crosslinked by the dissolved Ca 2+ ions and the glucose triggers the enzyme-mediated polymerization of NIPAm and MBA to produce the interdigitating PNIPAm network.b), The bienzymatic quaternary initiation system (HRP/GOx/AcAc/glucose) consumes glucose and molecular oxygen to produce acetylacetone radicals which in turn initiate PNIPAm polymerization.c), A representative palm-sized (35 × 35 × 5 mm) 3D printed interpenetrating network (IPN) hydrogel.The resulting structure is robust, flexible, and self-supporting.d), The 3D printed IPN also retains the contractile thermosensitivity of PNIPAm single networks.Scale bar = 1 cm.HRP protein structure from PDB ID: 1HCH. [54,55]GOx protein structure from PDB ID: 3QVP. [56,57]plored due to their excellent cell supporting potential and research maturity. [18]However, many rudimentary single-network hydrogels have limited utility due to poor mechanical performance.This mechanical weakness arises predominantly because classical polymer network formation will generate hydrogels that are randomly crosslinked. [19]This leads to heterogenous gels with domains of densely and sparsely crosslinked regions.Therefore, upon loading, the reception of stress will be localized to the weaker failure zones inside the gel.One can try to increase the level of crosslinking inside a gel, but this simply lowers the configurational entropy of the overall network resulting in gels that are brittle and less extensible. [20] hydrogel can often also be made tougher by increasing its polymer mass fraction or molecular weight, but this also tends to interfere with the microarchitecture and cytocompatibility of hydrogels by limiting nutrient diffusion or restricting space for proliferation. [21,22]Accordingly, IPNs have emerged as a promising alternative strategy for the fabrication of tough hydrogels.IPN hydrogels are reinforced by interdigitating polymer networks that are not covalently linked, but are mechanically interwoven. [23]26][27] Herein, extrusion additive manufacturing was employed to 3D print high fidelity IPN hydrogel structures that utilize a double enzyme-mediated polymerization (EMP) curing step to facilitate aerobic ambient-temperature reinforcement.The IPN bioink's components comprised GOx, HRP, N-isopropylacrylamide (NI-PAm), N, N'-methylenebisacrylamide (MBA) and sodium algi-nate.The enzymatic cascade, comprising an acetylacetone (AcAc) radical mediator, initiated the radical EMP of the PNIPAm/MBA network upon the addition of glucose, and scavenged molecular oxygen to reduce radical polymerization inhibition.The bioink exhibited excellent printability (accurate overhangs and arches) which allowed sophisticated print feature fabrication.The crosslinked IPN prints displayed reversible shape-changing contractile thermosensitive (CTS) properties and gave rise to robust and flexible structures on the centimeter length scale.The HRP in the bienzymatic initiation system could also be substituted with the artificial oxidoreductase C45, demonstrating synthetic enzyme incorporation.Furthermore, the bioink was cytocompatible with lab strain BL21 E. coli, which could be readily incorporated to produce a 3D bioprinted engineered living material that could detoxify the pesticide paraoxon, and the temperature-dependent mass-transfer-limited rate could be manipulated using the contractile-thermosensitive response of the IPN.

3D-Printing an Enzyme-Mediated IPN
The 3D constructs were fabricated layer-by-layer by extrusion from a pneumatic printhead and were then crosslinked through immersion in a glucose and calcium chloride solution to yield the IPN (Figure 1).No delamination was observed in crosslinked prints which indicated that interlayer adhesion was high.3D prints on the centimeter scale were readily achievable and maintained structural integrity for weeks.This demonstrated that the bienzymatic system successfully initiated radical polymerization in a biocompatible, aerobic environment, whilst maintaining low enough concentrations of hydrogen peroxide to avoid enzyme denaturation.The standard bioink alginate:PNIPAm mass fraction ratio used was 3.5:15, but it was possible to tune the bioink properties by changing these values (Figure S1, Supporting Information).For example, network compositions favoring alginate promoted printability but attenuated CTS, and vice versa.4D hollowconstruct printing was also possible by adjusting the initiator concentrations to promote surface curing (Figure S2, Supporting Information).GOx-mediated production of gluconic acid has the potential to decrease the pH of the hydrogel, which may impede the calcium crosslinking of alginate.However, this issue was addressed by incorporating a 5 mM, pH 5, NaAcetate buffer into the bioink.Additionally, successful alginate crosslinking was confirmed through comparison of selectively crosslinked bioinks (Figure S3, Supporting Information).
The bioink's initiation system was also compatible with synthetic enzymes.C45 (Supplementary Figure 4a,b) is a highly efficient maquette peroxidase, developed recently by the Anderson group, that is readily expressible in E. coli systems and has one of the best reported catalytic efficiencies among synthetic enzymes (k cat = 1200 s −1 , k cat /K m (ABTS) = 3.2 × 10 6 M −1 s −1 ). [28]nitial experiments demonstrated that C45 could successfully crosslink NIPAm/MBA single networks when utilized in a C45/GOx/AcAc/Glucose quaternary initiation system (Figure S4c, Supporting Information).The C45-based initiation system was also successfully incorporated into the PNIPAm/Alginate IPN bioink print protocol (Figure S4d,e, Supporting Information).The resultant constructs were robust and flexible like their HRP-mediated analogues and demonstrated the CTS properties indicative of PNIPAm single networks.

IPN Bioink Printability
The IPN bioink displayed high printability performance and was evaluated for deposition uniformity and filament bridging capability (Figure 2).Uniformity was assessed by printing a single layer regular crosshatch structure (Figure 2a) and analyzing the uniformity of its pore areas and line widths (Figure S5, Supporting Information).Multiple prints demonstrated excellent uniformity with narrow pore area and filament width distributions (Figure 2b,c), where the average relative standard deviation within all the samples was 8 and 4% for pore area and filament width, respectively.Filament collapse test results (Figure 2d,e) confirmed that the filament structures were self-supporting.For comparison, a commercial-bioink was tested alongside (CELLINK Bioink, CELLINK's universal cellsupporting bioink formulation).The test pillar platforms had heights of 10 mm, widths of 5 mm, and the increasing gap distances: 1.0, 2.0, 4.0, 8.0, and 16.0 mm.The IPN bioink was capable of bridging the final 16.0 mm gap and displayed no visually observable collapse up to the 4.0 mm gap.Furthermore, the IPN bioink's overall collapse area was not statistically significantly different from CELLINK Bioink's.This demonstrates the bridging capabilities of the IPN bioink are closely comparable to optimized bioink formulations.The IPN bioink then should be able to achieve analogously complicated print structures without the need for support materials (Figure S6, Supporting Information).

Structure and Mechanical Properties of IPN Hydrogel
Cryo-scanning electron microscopy (Cryo-SEM) was performed on IPN prints, as well as PNIPAm and alginate single networks, to investigate their microstructure (Figure 3a-c).Micrographs from the IPN print showed highly networked microporous structures (Figure 3c).Interestingly, micrographs from the PNIPAm (Figure 3a) and alginate (Figure 3b) SNs showed smaller, and larger pore sizes to the IPN, respectively.In the absence of the alginate SN, the PNIPAm forms a more densely crosslinked hydrogel network.This is consistent with the alginate network act-ing as a template for the larger pore formations in the IPN, which are advantageous for strong CTS, as largerpores allow increased contraction. [29]SEM micrographs of the IPN treated with ethylenediaminetetraacetic acid (EDTA) supported a templating mechanism, as selectively dissolving the IPN's alginate SN left behind a PNIPAm network with larger pores (Figure S7, Supporting Information).Cryo-SEM was also performed on IPN prints that had been contracted via thermal actuation (Figure 3d).As anticipated, IPN contraction was concomitant with a reduction in hydrogel pore size, with a pore diameter decrease from 2.13 ± 0.52 to 1.05 ± 0.50 μm (Figure 3e).
Compression studies on the IPN hydrogel and its constituent SNs confirmed the mechanical reinforcement provided by network interdigitation (Figure 3f).The IPN yielded a higher compressive stiffness (63 ± 15 kPa) than either of the PNIPAm (26 ± 6 kPa) or alginate (40 ± 6 kPa) SNs.Furthermore, compression cycles show the IPN also retained the majority of the elastic recovery properties of PNIPAm SNs (Figure S8, Supporting Information), i.e., IPN samples were able to fully recover from 40% strain (vs 56% for the PNIPAm SNs).Mechanical compressionto-failure experiments performed on the IPN samples found an average ultimate compressive strength of 54 ± 3.6 kPa at an average strain of 46 ± 3%.

Printed IPNs Demonstrate Reversible Thermosensitive Contractile Behavior
Having established a material synthesis protocol to 3D print robust alginate/PNIPAm IPNs, the CTS properties were benchmarked against the PNIPAm single networks.Standardized 16 × 3 × 3 mm prints were used for all thermosensitivity assays (Figure S9, Supporting Information).When heated to 60 °C for 30 min in an oven, IPN prints demonstrated a remarkable first contraction to 38 ± 5% of their initial volume (Figure 4a).However, these then recovered to only 90 ± 10% of the original volume, signifying hysteresis.This decrease in recovered volume continued by an average of 6% per cycle until after the third heat cycle, where the structures settled into fluctuating between 76 ± 2% and 31 ± 2% of the original printed volume.This is likely due to IPN network reorganization over repetitive cycles, resulting in the formation of a stable network configuration.[32] Despite hysteresis, this stabilized contraction still represented a 2.5-fold reduction in volume.To measure the IPN's contraction speed, prints were placed in a 60 °C water bath and their volume was measured at regular time intervals (Figure 4b).Over 30 min, the IPN prints contracted considerably (22 ± 3% their original volume).The initial contraction rate was fast, with a contraction to 73 ± 4% of the initial volume in the first minute, followed by a contraction to 33 ± 2% after 5 min.50% volume contraction was achieved-within 3 minutes.These data fit an exponential decay with rate constant -0.01 and a pre-exponential-factor of 75.9.This overall contraction profile behavior is typical of PNIPAm-based networks and is a function of their thermal conductivity (as the transfer of heat from the gel surface to its bulk is a determinative factor). [33]However, the specific function constants are dependent on the IPN print dimensions, as surface-to-volume ratio and the distance interstitial water would have to travel are key parameters in a PNIPAm gel's contraction speed.To achieve complete cycling, a temperature of 60 °C was employed for acellular experiments.36][37] The IPN bioink could also be co-printed with excellent continuity to non-CTS alginate-based 3D inks to fabricate structures with composition/property asymmetries and anisotropic actuation behavior (Figure 4c,d).This supports IPN bioink variant co-printing to pattern subtle performance dissimilarities into a single construct, resulting in 4D shape changes such as the formation of valve and gripper structures, a standard demonstration structure for many soft robotics materials, [38] or the realization of targeted, bioinspired structures. [39]Complex physical information processing could also be introduced by coupling this thermoresponsive IPN to soft actuating ELMs that have been designed to respond to other diverse stimuli including relative humidity [40] and small molecule signals [41] through mechanisms such as hydromorphic swelling or cell division, respectively.

Bioprinting Bacteria to Produce Dynamic ELMs
BL21 E. coli cells were dispersed in the IPN bioink immediately prior to bioprinting.E. coli viability (69 ± 4%) and their homogenous distribution within the bioink was confirmed via confocal fluorescence microscopy (Figure S10, Supporting Information).To demonstrate the potential utility and cytocompatibility of the IPN bioink, E. coli cells that were transformed to express the phosphotriesterase arPTE were incorporated.arPTE is a metalloenzyme that catalyzes the hydrolysis of the triester linkages found in organophosphate (OP) groups, present in many insecticides and nerve agents.arPTE is especially efficient at hydrolyzing the-organophosphate paraoxon, with turnover numbers (kcat) and catalytic efficiency (kcat/KM) values close to the diffusion-controlled limit (2280 s −1 and 6.2 × 107 M −1 s −1 respectively). [42]The paraoxon hydrolysis activity of the 3D printed living structures was assayed by monitoring the change in absorbance at 405 nm, where the paraoxon hydrolysis product (4-nitrophenolate) has its absorbance maximum (Figure 5).Significantly, the arPTE-E.coli laden bioreactors degraded 0.2 mM paraoxon extremely rapidly, with complete hydrolysis of 0.2 mM paraoxon to 4-nitrophenol observed in <15 min (Figure 5b).Furthermore, the detoxifying activity of the arPTE-E.[45] These results also demonstrate that encapsulated E. coli survived the room temperature bioprinting process and remained viable in the bioreactors overnight (incubated at 29 °C).This was expected, as most of the bioink components are known to be compatible with bacteria, and E. coli are known to be robust organisms capable of surviving in environments of non-physiological tonicities, including even sterile, distilled water. [46]Control experiments performed using either printed acellular and non-arPTE expressing E. coli laden bioreactors both demonstrated negligible absorbance and displayed no yellow coloration from paraoxon hydrolysis at any time after the addition of paraoxon.This confirmed that the production of 4nitrophenol was catalyzed by the heterologous expressed arPTE and demonstrated the IPN bioink's potential as a printable and programmable biohybrid material platform.
To explore whether the bioreactor rate could be modulated by the IPN's CTS properties, reaction rates were measured at 37 °C and 22 °C (Figure 5d; Figure S11, Supporting Information).For comparison, rates for aqueous arPTE enzyme and non-contracting (n-CTS) hydrogel bioreactors (6.5 wt.% alginate, 13 wt.%Poloxamer 407) were also measured.The aqueous enzyme demonstrated only a 6% decrease in reaction rate when temperature was lowered from 37 °C to 22 °C, but the n-CTS hydrogel bioreactors decreased in rate by 70%.This reduction in rate aligns with previous observations for immobilized enzyme hydrogel reactors, where reaction rates are mass transfer limited. [47,48]Significantly, the IPN bioreactors only displayed an 18% decrease in rate on temperature reduction.Here, although the lower temperature reduced mass transfer rates within the network, the IPN expanded as it transitioned through its lower critical solution temperature (LCST) (Figure S12, Supporting Information), increasing pore size, thereby countering the reduction in effective diffusion due to temperature change.
The effective diffusion coefficient for a small molecule through bulk water at 22 °C was calculated to be only 31% less than that for the same molecule and solvent at 37 °C, estimated using the Stokes-Einstein equation (Eqution (1)), where D eff is the effective diffusion coefficient of a particle through a given solvent, k B is the Boltzmann constant, T is temperature in Kelvin,  the dynamic viscosity of the solvent, and r the radius of the particle.This suggests that for the non-contracting gel, where a 70% reduction in hydrolysis rate occurred on cooling, other phenomena were affecting the mass transfer rates through the hydrogel.CryoSEM imaging of the n-CTS hydrogels at 37 °C and 22 °C revealed a more open pore structure within the gel at 37 °C (Figure S13, Supporting Information), potentially caused by improved wash out or phase change of residual Poloxamer 407 at the higher temperature.This in turn may have driven a reduction in tortuosity (the ratio between actual flow path length and a straight line path between two points) and constrictivity (a bulk property depending on the ratio between particle size and pore diameter), resulting in the large difference in rates observed for the n-CTS hydrogel.This was in contrast to the IPN, demonstrating that the IPN's contracting properties could be coupled to bioreactor rates, allowing for the design of bioreactors with functional thermosensitive reaction rates.
A shift in the hydrophobicity of the IPN network, caused by the change of interaction energies between polymer and solvent through the LCST, may also have affected the mass transfer rates of the solutes with the IPN.Here, network expansion is accompanied by an increase in the dielectric constant of the polymer, reducing solute-polymer interactions, and allowing for faster diffusion through the network.As the reactant ethyl-paraoxon and product 4-nitrophenol are both hydrophobic (LogP values of 1.97 and 1.61, respectively), this effect may have been significant.Whilst current diffusion models exclude interaction effects [49] or incorporate them into an empirical quantity, with solute diffusivity sufficiently explained in terms of hydrodynamic, free volume, and obstruction effects, [50] complex systems such as the CTS-DN described here will require further investigation to fully understand the effect of the LCST transition on diffusion and mass transfer.

Conclusion
In conclusion, the novel IPN-producing bioink described herein can be readily used to fabricate robust, flexible, microporous structures at high-resolution under mild and aerobic conditions.We have demonstrated that the printed IPN's structures maintain rapid and repeatable contractile thermosensitive properties, as well as their ability to be cohesively co-printed with other alginatebased inks to produce constructs with patterned property asymmetry.Post-printing viability and bioremediation activity of the cell-laden bioink also demonstrates the potential for this technology to be a platform to produce functional biohybrid materials.
Additionally, we have shown that printed IPN bioreactor rates are coupled to their thermosensitive material properties, producing a unique temperature dependent relationship.This represents a significant advance in bacteria-laden reinforced 3D printable stimuli-responsive hydrogel materials, and has potential applications in soft robotics, smart microreactors, responsive sensors, as a unique functional bio-scaffold, and in medical devices.The scope for genetically engineered bacterially derived functionality is also expansive, especially as synthetic biology continues to improve and augment the capabilities of natural proteins.
Preparation of the IPN bioink primed with a C45-based initiation system followed the exact same preparation but with a molar equivalent of C45 substituted for HRP.
4.0.0.1.Printing: All prints were performed in an aerobic environment and at room temperature with the Cellink INKREDIBLE+ bioprinter.Unless otherwise stated, print moves were set to 50 mm s −1 , extrusion pressure to 22 kPa, infill to 100%, and layer height equal to the used nozzle diameter.After deposition, the bioink was cured via immersion in a 300 mM CaCl 2 and 5 mg Ml −1 glucose solution overnight.The prints were then re-immersed in 10 mM CaCl 2 solution for storage.
Printability Assays: The printing uniformity assay was based from examples in the literature and analyzed the performance of the bioink via measuring the variance in pore sizes and line widths for a standard print (a 30 × 30 mm, 1-layer high cuboid STL file sliced to a 60% square mesh infill). [51,52]The print was performed using the standard IPN bioink formulation and extruded through an 18G plastic tapered needle (Metcal; Farnell).The print was then brightfield imaged on a LEICA DMI6000 inverted epifluorescence microscope.Print pore areas were calculated using image analysis with ImageJ software.Images were initially processed via thresholding to make them binary.After thresholding the pore areas were measured via calculating the volume and distribution of connected white pixels.Mean pore areas and standard deviations were calculated from 24 pores from each print, and 3 prints were performed.Line widths were measured manually via ImageJ.Mean values and standard deviations were calculated from 15 measurements from each print, and 3 prints were performed.Line widths were taken from halfway between crosshatch vertices.
The filament collapse assay was also based on previous examples from the literature. [53,54]The assay was conducted on a 3D printed platform (printed using a Formlabs Form 2 SLA printer loaded with a V4 clear resin cartridge) with pillar heights of 10.0 mm, widths of 5.0 mm and incrementally increasing gap distances of 1.0, 2.0, 4.0, 8.0, and 16.0 mm.Bioinks tested include the standard bioink formulation and Cellink's CELLINK Bioink (IK1020000303).Bioinks were printed at the lowest pressure that the bioink would bridge all sections, rounded to the closest multiple of 5 kPa, and print moves were set to 50 mm s −1 .All Bioinks were extruded through a 25G tapered plastic nozzle (Metcal; Farnell).The filament collapse areas were manually quantified using ImageJ software and calculated as the sum of the areas bellow the pillar heights the filament had collapsed.Mean values and standard deviation for the IPN filament collapse areas were taken from 5 repeats.Mean values and standard deviation for the CELLINK bioink's filament collapse areas were taken from 3 repeats.
Contractile Thermosensitivity Assays: The contractile TS assays were all performed on 16 × 3 × 3 mm cuboid prints.All volume measurements were calculated from measurements taken with electronic calipers.Volume changes were calculated as percentages of construct's initial volumes (v/v I ).In the repeatability assay printed IPN constructs were subjected to multiple heat cycles to study how elastically the IPN would contract.Here, IPN prints were heated by being placed in an oven at 60 °C for 30 min.Volume was measured immediately post-oven.Volume was remeasured after the print was allowed to relax at room temperature overnight.Mean values and standard deviation for print constructs volumes were taken from 6 repeats.In the kinetics assay, printed IPN constructs were heated via immersion in a water bath set to 60 °C and volume measurements were taken every minute for the first 10 min.Two further volume measurements were then taken after 20 and 30 min of immersion.Mean values and standard deviation for print constructs volumes were taken from 3 repeats.inoculate 30 mL of LB medium with 1 mg Ml −1 glucose.Inoculated broth was grown in a shaking incubator at 37 °C, 180 rpm.The next morning the culture was collected via centrifugation (3500 x g for 15 min) and resuspended in 5 mL of Mq water.This suspension was then used to inoculate the bioink immediately, which was then mixed in the DAC at 3500 rpm for 1 min.A typical 5 g preparation of bioink would contain 900 μL of this bacterial suspension as well as 50 μg L −1 carbenicillin.E. coli-laden prints were crosslinked post-printing via immersion in a 100 mM CaCl 2 + 5 mg mL −1 glucose solution for 1 hour.Prints were re-immersed in a 20 mM CaCl 2 , 5 mg mL −1 glucose + 1 mM IPTG solution.
For the assay, 2 of each print was placed per well in a 6-well plate.Prints were then washed three times with 10 mL, before being finally re-immersed in 4 mL of HEPES buffer.Paraoxon of 10Mm of 80 μL, an organophosphate which has a yellow hydrolysis product (4-nitrophenol), stock solution was then added to each well (0.2 mM final concentration).The assay was conducted with a BioTek SYNERGY neo2 multi-mode plate reader at 22 °C.Absorbance measurements at 405 nm were taken from a single point at the center of each well every 30 s for 30 min post-paraoxon addition.3D printed plastic spacers (designed in Tinkercad and printed using a Formlabs Form 2 SLA printer loaded with a V4 clear resin cartridge) were added to each well to ensure the hydrogels didn't float into the center and interfere with absorbance measurements.The negative control E. coli experiment was performed using BL21(DE3) E. coli that had only been transformed with a plasmid encoding for carbenicillin-resistance.The negative acellular control experiment was performed using uninoculated bioreactors.
C45 Expression and Purification: The artificial c-type cytochrome maquette C45 was purified from E. coli using established procedures. [55,56]ll of the steps were performed under sterile conditions.The sequence for C45 was placed on a pMal-p4x+ plasmid containing the sequences for the cytochrome itself, and CARB resistance, along with a Cterminus His 6 tag (to allow later purification via immobilized metal ion affinity) with a Tobacco Etch Virus (TEV) recognition sequence between them (to allow later removal of the purification tag).This plasmid was then transformed into E. coli T7 Express competent cells (NEB) following a standard protocol. [56]45 was co-transformed with plasmid pEC86, which contains the sequences for the relevant maturation machinery (the cytochrome I biogenesis system I (Ccm) required for incorporation of a c-type heme into the C45 scaffold) [57] as well as for chloramphenicol resistance.To ensure both plasmids were successfully transformed into cells, carbenicillin (50 μg L −1 ) and chloramphenicol (34 μg L −1 ) were added to media and plates.
Single colonies were selected from streaked plates of transformed E. coli and used to inoculate 5 mL of LB broth.Inoculated broth was grown at 37 °C overnight in an incubator with shaking/rotation.The next morning, 2.5 ml of the overnight culture would be used to inoculate 1 L of LB broth.This was then cultured until its optical density at 600 nm (OD 600 ) was between 0.6 and 0.8.The culture was then induced with 1 mL of 1 M IPTG to overexpress C45.The 1 L cultures would then be returned to the incubator for a further 4.5 h before being spun down (4000 x g for 30 mins) to a red paste.
If purifying isolated C45, the cell paste was dissolved in a small amount of 300 mM NaCl, 50 mM sodium phosphate and 40 mM imidazole at pH 8.0 (lysis buffer) and lysed by sonication (Sonics Vibra-Cell).C45 protein purification was then achieved by following a protocol closely based on the original by Watkins et al. [28] The soluble lysate was clarified by removing the insoluble fraction via ultracentrifugation (40000 g for 30 mins).The soluble lysate was then filtered (0.22 M syringe filter, Millipore) and loaded onto a HisTrap FF IMAC column (Cytiva) that had been equilibrated with lysis buffer.The bound protein was then washed with lysis buffer before being eluted with 300 mM NaCl, 50 mM sodium phosphate and 250 mM imidazole at pH 8.0 (elution buffer).Fractions containing the C45 protein were combined and dialyzed for 18 h against 5 L of 0.5 mM EDTA and 20 mM Tris at pH 8.0 in a 14 kDa semi-permeable dialysis membrane.Then, under anaerobic conditions, TCEP (1 mM) and TEV protease (1 μM) was added to the dialyzed protein to cleave the N-terminal hexa-histidine tag.This reaction was allowed to stand in anaerobic conditions overnight to ensure it progressed to completion.The protein was then filtered to remove precipitated TEV protease (0.22 μM syringe filter, Millipore) and concentrated to ≈5 mL.This was then loaded onto a Superdex 200 pg 26/600 size exclusion column (Cytiva), which had been equilibrated in redox buffer, and underwent isocratic elution at 2 mL min −1 .The sample eluted as 2 distinct peaks.The first peak corresponds to apo-C45 and aggregated proteins, these fractions were discarded.The second peak has two initial minor peaks likely corresponding to improperly folded C45.The major peak was pure C45 and eluted as a red solution.This was collected, concentrated to ≈50 μM and stored at 4 °C.
Confocal Microscopy of E. coli-Laden Bioink: Inoculated bioink samples (170 μL) were first washed in a sterile saline solution (1 mL, 0.9 w/v% NaCl) and then stained with SYTO 9 (10 mM) and propidium iodide (60 mM) in sterile saline solution (0.5 mL, 0.9 w/v% NaCl) stationary in a dark room for 45 min.The bioink samples were then rewashed twice with sterile saline solution (2 × 1 mL, 0.9 w/v% NaCl).The saline solution was then aspirated off the samples before imaging.All confocal imaging was performed on a Leica SP8 confocal laser scanning microscope.To calculate cell viability, images were first thresholded to make them binary.Cells were then counted via pixel volume, and viability calculated as the volume of live cells divided by the volume of live and dead cells.Viability percent-ages were reported as the average from 10 images taken at random from the bioink bulk.
Casting Isolated-C45 Initiated Single Networks: For isolated-C45 (1.14 μM) experiments, NIPAm was added as a solid (15 wt%) and AcAc was added neat (97 mM).All other reaction components were added from stock solutions, MBA in Mq (20 mg Ml −1 ), GOx in 50 mM NaAcetate pH 5 buffer (3.7 mg mL −1 , 23.125 μM), and glucose in Mq (10 mg mL −1 , 55.5 mM), to final reaction concentrations of 1 mg mL −1 , 3.12 μM, and 5.55 mM respectively.The final reaction volume was 1 mL.All casts were performed in 2 mL microcentrifuge tubes with the lid open to allow the influx of molecular oxygen.

Figure 2 .
Figure 2. Bioink printability.a), Photograph of a printed crosshatch and a composite image of the same printed crosshatch made up from a tile-scan of optical microscopy images.Scale bar = 5 mm.b), Histogram for pore areas of three repeats of the crosshatch print.c), Histogram for line widths of three repeats of the crosshatch print.d), Photographs of representative filament collapse tests conducted with the interpenetrating network bioink (above) and CELLINK Bioink (below).Scale bars = 10 mm.e), Collapse area values for the different bioinks.Collapse area values represent mean averages ± standard deviation for n = 3 experimental repeats.

Figure 3 .
Figure 3. Cryo-SEM and compression testing of the interpenetrating network (IPN) hydrogel and its individual constituent single networks (SNs).Cryo-SEM micrographs of: a), a 15 wt.%:7.5 mg mL −1 PNIPAm/MBA SN. b), A 3.5 wt.% alginate and 0.75 wt.% xanthan gum SN, c), The IPN hydrogel in the expanded state.d), The IPN hydrogel in its contracted state.scale bars = 20 μm.e), Average pore diameters from the micrographs of the IPN in the expanded IPN contracted state.Values represent mean averages ± standard deviation for 150 measurements.f), Strain modulus for the different hydrogels as measured by unconfined compression testing.Values represent mean averages ± standard deviation for n = 5 experimental repeats.

Figure 4 .
Figure 4. Printed IPN's contractile thermosensitivity (CTS).a), Plot demonstrating 16 × 3 × 3 mm rectangular print construct's reversible contraction during multiple thermal cycles.Samples were heated by being placed in a 60 °C oven for 30 min.Values represent mean averages ± standard deviation for n = 6 experimental repeats.b), Plot for volume contraction of a 16 × 3 × 3 mm rectangular printed construct versus time spent in a 60 °C water bath.Values represent mean averages ± standard deviation for n = 3 experimental repeats.Co-printing the IPN bioink (orange) with a non-thermosensitive alginate-based ink (green) to generate constructs with asymmetric smart properties: c), a hydrogel gripper.d), a hydrogel valve.The inks demonstrate excellent continuity.Scale bars = 10 mm.

Figure 5 .
Figure 5. 3D bioprinting organophosphate-degrading bacterial bioreactors with bioink.a), Photographs of a six-well plate well with two 3D printed bacterial bioreactors before (top) and 30 min after (below) paraoxon addition.The bioink has been inoculated with E. coli transformed to express phosphotriesterase, an organophosphate hydrolyzing enzyme.The yellow coloration is due to the formation of 4-nitrophenol, (a) paraoxon hydrolysis product that absorbs at 405 nm.The clear plastic construct is a 3D printed porous plastic scaffold designed to keep the bioreactors away from the center of the well where the absorbance is being measured.Scale bars = 10 mm.b), Paraoxon degradation was monitored over time using the absorbance at 405 nm.Plots are for arPTE-E.coli (blue), acellular (yellow), and E. coli (green) bioreactor wells.Complete remediation for arPTE-E.coli occurred within 20 min.Values represent mean averages ± min/max values for n = 3 experimental repeats.c), Bioreactor activity in Au/s for the first 600 s.Values represent mean averages ± standard deviation for n = 3 experimental repeats.d), Paraoxon hydrolysis activity decrease for aqueous arPTE (19 nM), non-thermosensitively contracting (n-CTS) hydrogel bioreactors (6.5 wt.% sodium alginate, 13 wt.%Poloxamer 407, crosslinked in 100 mM CaCl2 (aq)), and IPN bioreactors at 22 °C as compared to 37 °C.Values represent mean averages ± min/max values for n = 3 experimental repeats.