DOI: 10.1002/adma.202007758

biology involve diffusible molecular signals, there has also been considerable effort to create materials for the controlled release of molecules in response to localized stimuli (e.g., for targeted drug delivery). [3] While these examples illustrate the use of responsive materials for molecular communication, we are unaware of examples of abiotic interactive materials that can autonomously synthesize a diffusible signal as part of a bidirectional molecular exchange with biology.

Emerging research in redox biology demonstrates that diffusible reactive oxygen species (ROS) serve as signaling molecules for redox-based communication. [4] Initially, ROS were recognized as integral to adversarial immune interactions: host cells generate ROS to defend against pathogen threats while microbes respond to ROS by upregulating antioxidant defense mechanisms. More recently, the biological role of ROS has been expanded beyond adversarial hostpathogen interactions and now ROS are believed to be broadly important in the communication between a host and its microbiome (e.g., between the gut epithelium and its microbiome, [5][6][7] or between the root and its rhizosphere community [8][9][10] ). Interestingly, phenolics are often believed to play an important but incompletely understood, role in redox-based interactions. For instance, plants generate an extensive array of phenolics: some phenolics (e.g., acetosyringone (AS)) are believed to mediate plant-microbe interactions in the rhizosphere; [11] while many plant phenolics appear in our diet and contribute antioxidant activities. One study showed with the spice clove (which is reported to be one of the richest sources of dietary antioxidants) showed that the insoluble fraction retained redox activity and could be repeatedly oxidized and reduced. [12] This observation inspired the current work aimed at understanding if a nonliving (i.e., abiotic) material could participate in redox-based communication and induce biological responses.

In addition to phenolics, thiols are also important moieties in redox biology and there has been considerable effort to create redox-responsive materials that enlist biology's sulfur switching mechanism in which two thiols are converted into a disulfide linkage. [13] Previously we reported that the diffusible plant signaling molecule AS could be electrochemically oxidized to induce the formation of a disulfide-crosslinked hydrogel on an electrode surface. [14] Electrochemical evidence for this disulfide Emerging research indicates that biology routinely uses diffusible redoxactive molecules to mediate communication that can span biological systems (e.g., nervous and immune) and even kingdoms (e.g., a microbiome and its plant/animal host). This redox modality also provides new opportunities to create interactive materials that can communicate with living systems. Here, it is reported that the fabrication of a redox-active hydrogel film can autonomously synthesize a H 2 O 2 signaling molecule for communication with a bacterial population. Specifically, a catechol-conjugated/crosslinked 4-armed thiolated poly(ethylene glycol) hydrogel film is electrochemically fabricated in which the added catechol moieties confer redox activity: the film can accept electrons from biological reductants (e.g., ascorbate) and donate electrons to O 2 to generate H 2 O 2 . Electron-transfer from an Escherichia coli culture poises this film to generate the H 2 O 2 signaling molecule that can induce bacterial gene expression from a redox-responsive operon. Overall, this work demonstrates that catecholic materials can participate in redox-based interactions that elicit specific biological responses, and also suggests the possibility that natural phenolics may be a ubiquitous biological example of interactive materials.

The focus of soft matter research is shifting from static to dynamically interactive materials. This shift is particularly important for life science applications where the goal might be to enable "communication" between the material and biology. There are many examples of materials designed with tailored moduli [1] or specific ligands [2] that enable communication with biology through mechanical modalities or distinct cell-surface receptors. Because many communication mechanisms in crosslinking is demonstrated by the cyclic voltammetry (CV) measurements in Figure 1a(1), where the voltage was first scanned in the oxidative direction and then the reductive direction. The CV curves for: a 4-arm thiolated poly(ethylene glycol) (PEG-SH, 10 × 10 -3 m) shows negligible electrochemical oxidation; an AS (5 × 10 -3 m) solution shows reversible AS oxidation and reduction; a mixture containing both PEG-SH and AS shows an amplification of AS oxidation and an attenuation of AS reduction. This response is characteristic of oxidative redox-cycling mechanism in which the AS Ox generated at the electrode is reduced back to AS Red by the oxidative crosslinking of PEG-SH into the disulfide crosslinked network. [15,16] In addition to forming disulfides, thiols are also nucleophiles that can undergo conjugation reactions. In biology, quinones are common electrophilic conjugating moieties [17,18] and we thus investigated the interaction between PEG-SH and the quinone-forming species catechol. [19] The CV curves in Figure 1a(2) show that: the control catechol solution undergoes reversible electrochemical oxidation and reduction; the mixture of PEG-SH and catechol shows comparable catechol oxidation and significantly attenuated catechol reduction. This observation suggests the oxidized quinone (Cat Ox ) that was generated at the electrode was consumed by reaction with PEG-SH. [20] Conceivably, Cat Ox reacted with PEG-SH to form Michael-type thiol-quinone adduct. [21,22] Chemical evidence that thiols react with AS and catechol through different mechanisms were provided by analyzing the products of cysteine/AS and cysteine/catechol oxidation reactions. Experimentally, we electrochemically oxidized AS or catechol (5 × 10 -3 m) in the presence of cysteine (5 × 10 -3 m; + 0.7 V, 10 min) and then analyzed the product using single quadrupole electrospray ionization mass spectrometry (MS). As shown in Figure 1b, a new peak (m/z 241.07) corresponding to the molecular weight of the cystine dimer was observed from reaction between AS and cysteine. Two new peaks were observed from the reaction between catechol and cysteine, in which the peak at m/z 230.45 is consistent with molecular weight of (mono)thiol-quinone adduct and the peak at m/z 349.2 suggests that one catechol may form adducts with two cysteines. [23,24] These MS results provide independent evidence that catechol can crosslink thiols through the formation of thiol-quinone adducts.

As previously reported, the electrochemical oxidation of AS with PEG-SH over a long time leads to the electrodeposition of a crosslinked hydrogel film at the electrode surface. [14] Experimentally, a gold-coated silicon wafer was immersed in solutions containing PEG-SH (10 × 10 -3 m) and AS (5 × 10 -3 m) and a constant voltage (+0.7 V for 10 min) was imposed to induce gel formation. The left images in Figure 1c show: a colorless hydrogel film was formed when electrodeposition was cued by AS oxidation; the gel shows no color change upon contacting with a moderately strong oxidant K 2 IrCl 6 (Ir 4+ ; 5 × 10 -3 m); and the gel dissolves upon treatment with the reducing agent dithiothreitol (DTT, 5 × 10 -3 m). The redox-responsive dis-assembly of the AS-electrodeposited film further indicates that AS mediates the formation of reducible disulfide bonds. [25] When electrodeposition was cued by catechol oxidation, the right images in Figure 1c show a colorless hydrogel film was also formed. However, the color of the catechol-electrodeposited film changed to brown upon Ir 4+ -oxidation, and switched back to colorless upon DTT-reduction. The fact that the catechol electrodeposited hydrogel is not dissolved by DTT treatment suggests that at least some irreversible quinone-thiol crosslinks are formed. [26] Further, the observation that the catechol-PEG-SH gel has different colors upon oxidation or reduction suggests the conjugated catechol moieties remain redox-active and can be switched between reduced (catechol) and oxidized (quinone) states. [27,28] In the second electrodeposition experiment, we used transparent indium tin oxide (ITO)-coated electrodes to allow spectroscopic comparison of the films (as illustrated in Figure 1d). The catechol-and AS-deposited films were first generated onto the ITO electrodes (+0.9 V, 30 min) and then different redox states were set by treatment with either Ir 4+ or DTT (note: we used stronger oxidative conditions for electrodeposition because of the lower electrocatalytic activity of ITO versus gold). The spectra at the right in Figure 1d show the oxidized form of the catechol-deposited film (designated Cat Ox ) exhibits higher absorbance than the reduced form (Cat Red ), while AS-deposited film shows little absorbance. The difference spectrum between the Cat Ox and Cat Red film shows a peak near 420 nm, which is consistent with the characteristic peak of thiol-quinone adduct. [29] In summary, electrodeposition by catechol (but not by AS) leads to conjugation and crosslinking with redox-active catecholic moieties.

We further characterized the redox-activity of the catecholdeposited film using mediated electrochemical probing. [30,31] Hydrogel films were first deposited onto standard gold electrodes (+ 0.7 V, 1 min) from solutions containing PEG-SH (10 × 10 -3 m) and either catechol or AS (30 × 10 -3 m). After deposition, the redox activities were characterized by immersing the film-coated electrodes in mediator solutions and imposing cyclic oscillating voltages as illustrated in the left scheme in Figure 2a. The mediators serve as electron shuttles between film and electrode as they can: be electrochemically oxidized or reduced; diffuse into films; engage in electron-transfer redox reactions in the films; and then diffuse back to electrode to undergo electrochemical reactions to generate electrical output currents that can be interpreted to characterize the film's redox activity. Specifically, we cycled the voltage input between -0.5 V and +0.68 V to sequentially engage two redox mediators: Ru(NH 3 ) 6 Cl 3 (Ru 3+ ) (Ru 3+ , 50 × 10 -6 m, E 0 = -0.2 V) and ferrocene dimethanol (Fc, 50 × 10 -6 m, E 0 = 0.25 V). The scheme in Figure 2a illustrates Ru 3+ engages in reductive redox-cycling reactions to transfer electrons from electrode to the film. The second mediator, Fc, engages in oxidative redox-cycling to transfer electrons from film to the electrode.

The redox activity of catechol-deposited film was examined by comparing its current output and CV response against those observed for a control uncoated gold electrode as shown by the plots in Figure 2a. When the voltage was cycled into the reducing region, the Ru 3+ -reduction current was significantly amplified for the film-coated electrode consistent with a reductive redox cycling mechanism. When the voltage was cycled into the oxidative region, the Fc-oxidation current was significantly amplified for the film-coated electrode consistent with an oxidative redox cycling mechanism. Furthermore, the CV curves from 10 consecutive scans of the electrode coated with the catechol-deposited film are approximately superimposable, indicating the catechol film can repeatedly accept and donate electrons. (Note: additional experimental evidence of this film's redox activity is provided in the Supporting Information.)

In contrast, the output and CV response for AS-deposited film are similar to that of control gold electrode. The small currents observed were associated with oxidation and reduction of the diffusible mediators. This indicates that the AS-deposited film is permeable to the mediators, but this film possesses no redox-activity.

Taken together, the mediated electrochemical probing of AS-and catechol-deposited films demonstrates that: i) the AS-deposited film shows no redox activity within the redox window probed (the E 0 range between Ru 3+ and Fc); and ii) the catechol-deposited film shows reversible redox-activity in this voltage range indicating that the grafted catechol moieties can repeatedly accept and donate electrons.

While CV allows dynamic measurement of the switching between oxidized and reduced redox states, open circuit potential (OCP) measurements provide a non-destructive means [32][33][34][35] to detect the redox state for the catechol-deposited films. We illustrate OCP measurements using films that were first electrodeposited on standard gold electrode (10 × 10 -3 m PEG-SH, 30 × 10 -3 m catechol or AS, + 0.7 V, 1 min). We then used the experimental procedure illustrated in Figure 2b in which a catechol-deposited film was poised in a specific redox state by incubation in a mediator solution and imposing a voltage for 10 min (either -0.5 V to poise a reduced state or +0.6 V to poise an oxidized state). The electrodes were then disconnected from the power source for 5 min to allow the films to equilibrate, after which they were re-connected to allow measurement of OCP as a function of time. The plots in Figure 2b show the film poised in the reduced state maintained a reducing OCP, while the film poised in its oxidized state maintained an oxidizing OCP. Controls of an uncoated gold electrode or an electrode coated with an AS-deposited film show that their OCP's have minimal dependence on their prior treatment. These OCP results illustrate the ability to switch and detect the redox state of the catechol-deposited films.

Next, we tested whether the catechol-deposited film can exchange electrons with biologically relevant redox-active molecules. In the first study, we electrochemically poised a catechol-deposited film in either an oxidized or reduced state (as illustrated in Figure 2b), incubated the film-coated electrode in the presence of the reductant ascorbate (0.5 × 10 -3 m) and monitored changes in OCP. The schematic in Figure 3a illustrates a putative mechanism in which ascorbate donates electrons to the film to reduce the oxidized quinone moieties to their reduced catechol state. The plot in Figure 3a shows the measured OCP for the initially oxidized film becomes increasingly negative over the course of 4 min and approaches the OCP of the initially reduced film. The controls (gold electrode with or without an AS-deposited film) show no change in OCP upon ascorbate treatment. This experiment provides initial evidence that the films can accept electrons from reductants in the external aqueous solution and this film-reduction occurs over time scales of minutes.

As illustrated in Figure 3b, we used a second experimental approach to demonstrate that the catechol-deposited film can accept electrons from the biological reductants ascorbate and nicotinamide adenine dinucleotide phosphate (NADPH). Experimentally, the film was deposited onto standard gold electrode (+0.7 V, 5 min) from a solution containing 10 × 10 then converted to an oxidized state by Ir 4+ treatment (designated Cat Ox ), and then the ability of the film to accept electrons was tested by incubation in solutions containing ascorbate or NADPH. Visual evidence for such electron transfer is provided by the bottom-left image in Figure 3b which shows the dark-colored oxidized film loses its color upon incubation with 500 × 10 -3 m ascorbate, consistent with the redox-responsive color change we observed in Figure 1c. Further electrochemical analysis was performed by incubating these films in solutions containing ascorbate (5 × 10 -3 m) or NADPH (5 × 10 -3 m) for 10 min, and then transferring the film-coated electrodes into a solution containing Fc (50 × 10 -6 m) to allow redox-probing of the film's redox state. The CV plots in Figure 3b show that the films that had been incubated with ascorbate and NADPH show amplified Fc-oxidation currents (e.g., greater abilities to donate electrons to Fc) compared to controls (gold electrode or untreated film-coated electrode). These observations indicate that the catechol-deposited film can accept electrons from common biological reductants.

We then examined if the electrodeposited films can donate electrons to O 2 to generate H 2 O 2 [36][37][38] (Figure 3c). In the first study, a catechol-deposited film was immersed in a mediator solution that had been bubbled with N 2 , and then the film's redox state was electrochemically poised (as illustrated in Figure 2b). After 5 min of equilibration, the OCP was monitored initially (first min) with an N 2 gas stream, and then with a compressed air stream. Figure 3c shows that the OCP measurement for the initially reduced catechol film is rather stable under N 2 , but becomes increasingly more oxidative upon air purging approaching the OCP of the initially oxidized film within 3 min. The controls (catechol-deposited film in an initially oxidized state, an AS-deposited film and an uncoated gold electrode) show no significant change in OCP upon the introduction of air. This experiment provides initial evidence that the films can donate electrons to air over a time scale of minutes.

The donation of electrons from catechol-deposited films to O 2 is expected to lead to the generation of ROS (e.g., H 2 O 2 ). Experimentally, we electrodeposited films onto a gold-coated wafer (+0.7 V, 20 min) and then set the film's redox state by treating with Ir 4+ or DTT. As illustrated in Figure 3d, these films were then incubated in 1 mL deionized water (room temperature), and the H 2 O 2 concentrations in the solutions were then analyzed at pre-determined intervals using the ferrous mediated oxidation of xylenol orange which generates a purple colored product with characteristic absorbance peak at 595 nm. [39] We should note that a relatively large electrode surface area to liquid volume (SA/V ≈ 1.8 cm -1 ) was used in these experiments to allow the generation of a more concentrated H 2 O 2 solution (i.e., to facilitate measurement). The inset images in Figure 3d were taken after 24 h incubation, and quantitative results show that the reduced film (designated Cat Red ) progressively generates H 2 O 2 (approximately 200 × 10 -6 m was generated after 24 h incubation), while the control films show minimal H 2 O 2 production (additional evidence for H 2 O 2 -generation is provided in the Supporting Information). These results demonstrate that the catechol-deposited film in its reduced state is able to donate electrons to O 2 for H 2 O 2 -generation.

Because H 2 O 2 is an important biological signaling molecule, [40] we reasoned that a H 2 O 2 -generating film could interact with living cells through this native biological communication modality. Figure 4a illustrates our hypothesis that a Cat Red film can generate H 2 O 2 to activate the redox-responsive regulon oxyRS in Escherichia coli (E. coli). [41][42][43] To test this hypothesis, we engineered an E. coli reporter cell that responds to H 2 O 2 by expression of the reporter protein, superfolder green fluorescent protein (sfGFP) and measured gene expression using fluorescence-activated cell sorting (FACS). Cloning details and characterization of the H 2 O 2 -response of these reporter cells is provided in the Supporting Information. The first histogram in Figure 4b shows a large population of E. coli expressed relatively high levels of sfGFP after 6 h incubation with 200 × 10 -6 m H 2 O 2 . The second histogram shows the addition of the H 2 O 2 -degrading enzyme catalase (250 µg mL -1 ) significantly diminished the expression level of sfGFP.

Next, we investigated whether Cat Red film can induce H 2 O 2dependent gene expression. Experimentally, we electrodeposited the films using 10 × 10 -3 m PEG-SH and 30 × 10 -3 m catechol (+0.7 V, 20 min) onto gold-coated wafer, set the redox state (using DTT), and then incubated the films with 1 mL E. coli reporter cells (OD 600 = 0.4) for 6 h. Note the SA/V in these experiments is purposefully high (approximately 1.8 cm -1 ) to allow the generation of a more concentrated H 2 O 2 solution and thus induce a more pronounced gene response. As shown by the histogram in Figure 4b, a high level of gene expression was observed when the cells were incubated with the Cat Red film, while expression was significantly attenuated when the H 2 O 2degrading catalase was included in the incubation mixture. Incubation with control films (Cat Ox or AS-deposited films) also showed negligible expression from this redox-responsive E. coli construct after 6 h. These results demonstrate that a catecholdeposited film that is poised in its reduced state can communicate with biology by synthesizing the diffusible H 2 O 2 signal that can induce redox-responsive gene expression (e.g., to alter bacterial phenotype).

We then tested for communication in the opposite direction: the bacterial-mediated transfer of electrons to an oxidized Cat Ox film. Bacterial extracellular electron transfer is well known for processes that include energy metabolism, [44] biofilm development, [45,46] antioxidant protection, [47] and virulence. [48] The

Adv. Mater. 2021, 33, 2007758 www.advmat.de www.advancedsciencenews.com Adv. Mater. 2021, 33, 2007758

simplest illustration of electron transfer from bacteria to the film is to observe a change in the film's redox-state dependent optical properties upon contact with a bacterial culture. Experimentally, we prepared catechol-deposited film (10 × 10 -3 m PEG-SH, 30 × 10 -3 m catechol, +0.7 V, 5 min), poised the film's redox state (by Ir 4+ treatment) and transferred this film-coated electrode to 800 µL of a high density E. coli culture (OD 600 = 3; SA/V ≈ 0.8 cm -1 ) as illustrated in Figure 4c. The first photograph in Figure 4c shows the bottom portion of the initially brown Cat Ox film lost color after 10 h contact with the E. coli culture, while the second photograph shows no such color change for the film that was exposed to cell-free Luria broth medium. These results are consistent with the redoxresponsive color change of catechol-deposited film shown in Figures 1c and 3a, indicating the oxidized Cat Ox film can accept electrons from E. coli culture. Additional experimental evidence that the bacterial culture can transfer electrons to the film is provided in the Figure S12, Supporting Information.

The bidirectional interactive communication between the films and bacteria was demonstrated by a study of the biological response of E. coli. Briefly, films were electrodeposited (10 × 10 -3 m PEG-SH, 30 × 10 -3 m catechol, +0.7 V, 20 min) and poised to specific redox state (by DTT or Ir 4+ treatment), and these films were incubated with 1 mL E. coli culture (OD 600 = 0.4, SA/V ≈ 1.8 cm -1 ) under shaking for 24 h. At predetermined intervals, the expression of sfGFP was analyzed and quantified by FACS as the percentages of the population responding (i.e., green cells).

Figure 4d(1) shows the representative FACS histogram of E. coli incubated with the Cat Ox film: expression was negligible at 6 h; small fluorescent population was observed at 10 h; and a larger fluorescent population was observed at 24 h. When the H 2 O 2 -degrading catalase was included in the incubation mixture, the gene expression was significantly suppressed over the full 24-h time course. As illustrated by the scheme in Figure 4d(1), these results provide biological evidence for interactive communication between the Cat Ox film and E. coli: E. coli must provide the source of electrons that reduce the Cat Ox film enabling it to transfer these electrons for the generation the H 2 O 2 and induction of the H 2 O 2 -specific response in the E. coli culture.

The representative FACS histograms of E. coli incubated with the Cat Red film are shown in Figure 4d(2) and compared with the response of a control culture exposed to 200 × 10 -6 m H 2 O 2 (the concentration generated after 24 h from the Cat Red film of Figure 3d). These histograms show both cultures respond rapidly while the cultures exposed to the Cat Red film show a stronger and more prolonged response (compared to the H 2 O 2 -control). As illustrated by the scheme in Figure 4d(2), this more prolonged response in the presence of the Cat Red film suggests a redox-cycling between the bacterial culture and the film which provides additional evidence for interactive communication.

Figure 4d(3) shows the time course which summarizes the observations from this experiment. The culture exposed to the Cat Red film shows a large and progressively increasing H 2 O 2 -induced response (compared to the H 2 O 2 -control), while the culture exposed to the Cat Ox film shows a significant but delayed response.

In summary, we report a hydrogel film capable of interactively communicating with biology through a native redox modality. Specifically, a redox-active catechol conjugated/ crosslinked thiolated PEG film is formed by a one-step electrodeposition. We demonstrate that this film can accept electrons from biological reductants and donate electrons to O 2 to generate the diffusible H 2 O 2 signaling molecule. In the presence of an E. coli reporter culture, this catechol-conjugated gel can accept electrons from E. coli culture and autonomously synthesize H 2 O 2 to induce redox-responsive gene expression. This demonstration that catecholic materials can communicate with biology through a native redox modality has relevance both for the development of interactive materials [36] and characterization of the participation of non-living components (e.g., melanin) in redox biology. [49,50] It is also interesting to speculate on two broader potential implications of this work. First, phenolics and thiols are each known to be important in redox biology and in some biological contexts these species are co-localized (e.g., mucin and dietary antioxidants in the gut). [51,52] This study shows that phenolics and thiols can undergo important redox interactions that appear to be selective (AS oxidatively crosslinks while catechol conjugates) and lead to different molecular structures and functional characteristics (reversible disulfide crosslinking versus reversible redox-activity). Second, catechols are among nature's most abundant redox active organics and results from this study indicate that they can engage in interactive redox communication (accepting and donating electrons) that cue biological responses. Emerging research indicates that redox-based communication involves redox reaction networks (i.e., redox interactomes) [53] and the results from this study suggest catechols may be hubs in such reaction networks. Thus, we speculate that redox-based interactive materials could become an important component of redox-biology and redox-medicine. [54,55]