Reversibly tuning hydrogel stiffness through photocontrolled dynamic covalent crosslinks

By controlling the stability of dynamic covalent crosslinks with adjacent photoswitches, the stiffness of an adaptable hydrogel is tuned reversibly.

Instrumentation. Proton nuclear magnetic resonance ( 1 H NMR) spectra and carbon nuclear magnetic resonance ( 13 C NMR) spectra were recorded on Bruker AVANCE-500 spectrometers at 500 MHz and 125 MHz, and referenced to the solvent residual peaks. Boron nuclear magnetic resonance ( 11 B NMR) spectra were recorded on Bruker AVANCE-400 spectrometers at 128 MHz. BF3•OEt2 in CDCl3 in a capillary was used as a reference for 11 B NMR (0 ppm) in Wilmad Precision NMR tubes (CFQ, 500 MHz, OD: 5 mm, wall thickness: 0.38 mm). NMR data are represented as follows: chemical shift (d ppm), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet), coupling constant in Hertz (Hz), integration. UV-vis spectra were collected on a Cary 5000 UV-vis-NIR spectrophotometer with an Hg lamp; cuvettes were 10-mm path length quartz cells (Starna 23-Q-10). LCMS kinetics experiments was performed on a liquid chromatographymass spectrometry system (Agilent 6120 Quadrupole LC/MS) equipped with a variable-wavelength detector. Infrared spectroscopy was performed on a Thermo Nicolet iS10 with a ZnSe crystal ATR attachment. Size exclusion chromatography (SEC) measurements were performed in stabilized, HPLC-grade tetrahydrofuran using an Agilent 1260 Infinity II system with variable-wavelength diode array (254, 450, and 530 nm) and refractive index detectors, guard column (Agilent PLgel; 5µm; 50 x 7.5 mm), and three analytical columns (Agilent PLgel; 5µm; 300 x 7.5 mm; 10 5 , 10 4 , and 10 3 Å pore sizes). The instrument was calibrated with narrowdispersity polystyrene standards between 640 Da and 2300 kDa (Polymer Standards Service GmbH). All runs were performed at 1.0 mL/min flow rate and 40 ºC. Molecular weight values are calculated based on the refractive index signal. Blue LED strip lights (wavelength = 470 nm, power = 6.6 W) and green LED strip lights (wavelength = 525 nm, power = 5.7 W) were purchased from superbrightleds.com. The UV flashlight (wavelength = 365nm, power = 1.5 W) was purchased from American's Preferred. Light intensity was measured using an EXTECH LT45 LED light meter, and the light intensities for the blue and green LEDs were 900 and 600 lux respectively. The light intensity for the UV flashlight was calculated to be 3.6 mW/cm 2 .

S3
II. Synthesis of azobenzene boronic acids Scheme S1. Synthetic route of small molecules 1 and 2 for kinetic studies and compound SI-2 and SI-5 for synthesis of control polymers P2 and P3. methyl 4-nitrosobenzoate (SI-1) Based on a modified from literature procedure, methyl 4-aminobenzoate (3 g, 20 mmol) was dissolved in 66 mL of DCM into a 500-mL RBF equipped with stir bar. 1 Oxone (20 g, 40 mmol) was dissolved into 250 mL of deionized water and then added to the reaction mixture (capped with rubber septum which was pierced with a needle, exposed to air), and the reaction was aged for 1 hour with high stirring. Over the course of the reaction, the biphasic solution develops into an intense neon green color. The product was extracted against water (1x), 1M HCl (2x), and 1M NaOH (3x). The organic layer was dried over sodium sulfate and concentrated in vacuo to give methyl 4-nitrosobenzoate (2.26 g, 70%). This product was used in the subsequent step without further purification. 1 H NMR (500 MHz, Chloroform-d) δ 8.30 (d, J = 8.5 Hz, 2H), 7.94 (d, J = 8.5 Hz, 2H), 3.98 (s, 3H). (2) SI-1 (4.4 g, 27 mmol) and 2-(4,4,5,5,-tetramethyl-1,3,2-dioxaborolan-2-yl)aniline (6.1 g, 28 mmol) were dissolved in 100 mL of DCM into a 250 mL RBF, to which 20 mL of acetic acid was added. The reaction was stirred for 24 hours, open to air. The solution turned black after several hours. Upon consumption of the starting material, the organic layer wash washed with 1M HCl (3x), and the organic layer was collected, dried with sodium sulfate, and concentrated in vacuo. Residual acetic acid was removed under high vacuum. The collected solids were diluted in hexane (300 mL), and the precipitate which formed (byproduct) was removed with a glass frit. The red filtrate was concentrated in vacuo to approximately 100 mL and then filtered again. The collected filtrate was concentrated in vacuo to yield an orange solid (~ 6 g) which was recrystallized from hexane (30 mL) twice to provide the product as a red crystalline solid (5 g, 50% (E)-(2-((4-(methoxycarbonyl)phenyl)diazinyl)phenyl)boronic acid (1) 2 (5 g, 10 mmol) was dissolved into a 500 mL RBF equipped with a stir bar (capped with a rubber septum which was pierced with a needle, exposed to air). The starting material was dissolved in THF (250 mL), providing a deep clear red solution. Sodium periodate (9 g, 40 mmol) was dissolved in DI water (62 mL) and transferred to the solution. After stirring for 30 minutes, 1M HCl (8 mL) was added. The reaction was stirred overnight at room temperature. In the morning the solution was red and cloudy with precipitate. The THF was removed in vacuo to concentrate the product in water. The resulting suspension was diluted with additional water and filtered to provide an orange/ peach colored solid. The solids were washed with acetonitrile (100 mL) and collected and dried on a glass filter to give the 1 (3.5 g, 90%) as a peach-colored solid. This product was used in the subsequent step without further purification.

(E)-4-((2-boronophenyl)diazenyl)benzoic acid (SI-2)
1 (3 g, 10 mmol) was dissolved in MeOH (282 mL) in a 1-L RBF equipped with a stir bar, providing a bright red solution. Lithium hydroxide (1 g, 40 mmol) was dissolved into water (58 mL) and transferred to the reaction mixture, which was stirred open to air overnight. The methanol was removed in vacuo, providing a red alkaline aqueous solution. This solution was neutralized with 1M HCl to precipitate the desired product, which was collected on a glass filter and dried in a vacuum oven (100 o C) for 1 hr to provide the title product

(E)-(4-((4-(methoxycarbonyl)phenyl)diazenyl)phenyl)boronic acid (SI-4)
SI-3 (510 mg, 1.39 mmol) was dissolved in 25 mL of THF into a 100 mL RBF equipped with a stir bar to give a red colored solution. Sodium periodate (298 mg, 1.39 mmol) was dissolved in water (6.2 mL) and added to the reaction. After 30 minutes, 1 mL of 1 M HCl was added to the solution, which was stirred for a further 6 hours. Upon full conversion, the THF was removed in vacuo leading to the precipitation of product, which was collected on a glass frit. The solids were washed with water (3 x 10 mL) and acetonitrile (10 mL) and dried to provide the title product (320 mg, 81%) as a peach-colored solid. 1

(E)-4-((4-boronophenyl)diazenyl)benzoic acid (SI-5)
SI-4 (1.19 g, 4.19 mmol) was dissolved in methanol (112 mL) in a 250-mL RBF equipped with a stir bar. Lithium hydroxide (0.401 g, 16.8 mmol) was dissolved in water (23 mL) and added to the solution, which was stirred overnight. After 14 hours, the reaction was as an orange solution. The methanol was removed in vacuo and then 1M NaOH was added to the crude mixture to dissolve the solids. The aqueous solution was washed with EtOAc and then neutralized with 1M HCl, leading to the precipitation of the product. The solids were filtered, washed with water, and dried to provide the title product (160 mg, 50%) as a peach-colored solid

4-amino-3,5-difluorobenzoic acid
4-Amino-3,5-difluorobenzoic acid was synthesized by following a previously reported procedure. Characterization data were consistent with literature reports. 2 methyl 4-amino-3,5-difluorobenzoate (SI-6) 4-Amino-3,5-difluorobenzoic acid (500 mg, 2.89 mmol) was dissolved in 20 mL of MeOH to which 1.75 mL of concentrated sulfuric acid was added to. The reaction was heated to reflux for 14 hours. After the reaction the solvent was removed in vacuo and 20 mL of water was added, leading the precipitation of the product, which was collected and dried on a funnel to provide the title product (417 mg, 77%) as a white solid, which was used in the next reaction without further purification. 1 H NMR (500 MHz, CDCl3) δ 7.58 -7.46 (m, 2H), 4.14 (s, 2H), 3.87 (s, 3H).

(E)-(2-((2,6-difluoro-4-(methoxycarbonyl)phenyl)diazenyl)phenyl)boronic acid (SI-8)
SI-7 and 2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)aniline (2.23 g, 10.2 mmol) were weighed into a 100 mL RBF equipped with a stir rod under inert atmosphere. DCM (70 mL) was added, and the dissolution of both starting materials led to the development of a dark-colored solution. Acetic acid (12 mL) was added and the solution was left to react for 24 hours. After 24 hours, the solvent was removed in vacuo and with high vacuum (to remove excess acetic acid), and the crude residue was diluted in 100 mL of hexane and filtered to provide the protected crude (2.17 g) as a red solid. This material was dissolved in 130 mL of THF into a 250 mL RBF equipped with a stir rod. Sodium periodate (3.46 g, 16.2 mmol) was added in 30 mL of water and the solution was left to react for 14 hours. After the reaction, the THF was removed in vacuo, and the remaining crude was diluted with 100 mL of DCM. The product precipitated from solution and was collected on a Buchner funnel, which was washed with an additional 100 mL of DCM to yield the title product (1.12 g, 34%) as a dark orange solid.

Procedure for calibration curves
Stock solutions of (E)-1 or (E)-2 were prepared in acetonitrile at 800 µM. Serial dilutions were performed to obtain a final concentration of 12.5 µM. The solutions were directly injected onto the LCMS (Poroshell 120 column, EC-C18, 2.7 µm, 2.1 x 50 mm). Runs were ten minutes long with solvent gradient from 50:50 to 100:0 ACN:H2O in the first three minutes. Detection was performed at 254 nm with a variable-wavelength detector. The peak area for each concentration was recorded. A calibration curve was created. The Z-isomer calibration curves were prepared by irradiating the series above for 10 minutes with 365 nm UV light and directly injecting the solution onto the LCMS. The concentration of the Z-isomer can be calculated from the known concentration of the E-isomer (based on the previously-generated calibration curve).

Pseudo-first order reversible esterification of (E)-1
500 µL of a stock solution of 800 µM (E)-1 was diluted with 500 µL of 80 mM pinacol in deionized H2O. The concentration of (E)-1 and (E)-2 were monitored as a function of time using the above LCMS method. Because both water and pinacol were present in excess, we used pseudo-first-order approximations. Apparent rate and equilibrium constants were determined using the derivation shown below. Scheme S2. Chemical structures for the pseudo-first order reversible esterification of (E)-1 and/or (Z)-1. The photochemical E/Z isomerization is drawn as irreversible due to the slow thermal relaxation of (Z)-1 and (Z)-2. This also allows for the simplification of the derivation of observed rate constants, shown below.
The data from the concentration vs. time experiments were plotted according to equation (1) for a reversible first order reaction (2) and the slope of the best-fit line was determined in Origin (Scheme S4). The apparent equilibrium constant (K'eq) for the reaction is the ratio of (E)-2 to (E)-1 at equilibrium, which can also be represented as the ratio of apparent rate constants (k') for the forward (esterification, k'1) and reverse (hydrolysis, k'-1) reactions, as shown in equation (2). Equation (2) can be rewritten as equation (3). With the apparent equilibrium constant and slope known, the apparent rate constant for hydrolysis (k'-1) can be determined with equation (5), and subsequently the apparent rate constant for esterification (k'1) can be determined by plugging these values into equation 4.

S12
Definition of equilibrium and apparent equilibrium constants: In Table 1, the k and K values are apparent rate and equilibrium constants (shown as k' and K' in the derivation above). S13 Figure S3.

Pseudo-first order reversible esterification of (Z)-1 and (E)-1
500 µL of a stock solution of 800 µM of (E)-1 was irradiated with a 365 nm flashlight for 10 minutes to produce a mixture of (E)-1 and (Z)-1. The mixture was then diluted with 500 µL of a 80 mM pinacol solution in deionized H2O, and the concentration of (E)-2 and (Z)-2 were monitored as a function of time using the above LCMS method. The observed rate constants of esterification and hydrolysis k1(Z) and k-1(Z) were solved for using the same derivation as shown above for the E isomer. A small increase in the total concentration of (E)-1 and (E)-2 over the course of this experiment is ascribed to thermal relaxation of the (Z) isomers.

Pseudo-first order irreversible hydrolysis of (Z)-2 and (E)-2
500 µL of a stock solution of 800 µM of (E)-2 was irradiated with a 365 nm flashlight for 10 minutes to produce a mixture of (E)-2 and (Z)-2. The mixture was then diluted with 500 µL of DI H2O and the concentration of (E)-2 and (Z)-2 was monitored as a function of time. The observed rate constants of hydrolysis k-1(E) and k-1(Z) were determined using the derivation shown below.    Figure S1. For this reason, we irradiated the samples in ACN prior to mixing with water and observed no deboronation. Given the reversibility of the mechanical properties, we do not believe that the production of boric acid significantly affects the behavior of the gel.

S23
IV. Functionalization of 4-arm PEG P1 was prepared following a literature procedure. 4 P2 was synthesized based on an adapted literature procedure. 5 500 mg of 4-arm PEG amine HCl salt (Mw 5000) was added to a Schlenk flask (50 mL) and melted under high vaccum (3x) to remove excess water. HOBt (0.2 g, 1 mmol), benzotriazol-1-yloxytris(dimethylamino)-phosphonium hexafluorophosphate (0.4 g, 1 mmol), and SI-2 (0.3 g, 1 mmol) were then added as solids along with a stir bar. The vessel was sealed and backfilled with nitrogen three times. 5 mL of DCM and 5 mL of DMF were then added to solubilize the reagents, leading to a clear red solution. Triethylamine (91 mg, 125 µL, 0.90 mmol) was added to the solution, which was stirred at room temperature for 24 hours. The solvent was removed in vacuo and the solution was diluted with deionized water to form an orange precipitate. The solids removed by filtration using a fritted funnel, and the aqueous orange filtrate was subject to dialysis (MWCO = 3.5 kDa) against DI water for 24 hours, changing the dialysate at least 3 times. After dialysis, the sample was lyophilized for 48 hours, yielding 320 mg of an orange powder, P2. Polymers should be stored dry in fridge without exposure to light to preserve integrity. Leaving solutions of P2 in DI H2O in ambient light for 24 hours leads to protodeboronation, as evinced by proton NMR.
P2-F2 was synthesized using the same procedure as P2 using SI-10. After dialysis this solution was concentration and purified using a spin filter (MWCO = 5KDa) After the reaction this was lyophilized for 48 hours (91 mg yield).
P3 was synthesized by the same procedure as P2 using SI-5 (377 mg yield).
P4 was synthesized by same procedure as P2 using (E)-4-(phenyldiazenyl)benzoic acid (217 mg yield).  Figure S18). Unless noted otherwise, oscillatory strain amplitude sweeps were conducted using a frequency of 25 rad/s and oscillatory frequencies were conducted using 10% strain. Gelation profiles were conducted with 10% strain and a frequency of 25 rad/s. Frequency sweeps were performed at 10% strain, with a frequency range of 300 to 1 rad/s. Data were collected at 25 o C. Gels were prepared by mixing 200 µL of P1 (10 w/v% in 0.1M PBS, pH 7.5) with P2, P3, or P4 (10 w/v% in 0.1M PBS, pH 7.5). The setup for photorheology is shown in the photos section.
An consistent observation from our reversible gelation profiles is that the first stiffening event occurs approximately twice as slowly compared to subsequent cycles. We hypothesize this may be due to (a) a slightly elevated temperature of the glass plate (see photos section) of our photorheology setup after continuous irradiation, (b) to diffusion processes in the macroscopic gel, or (c) the photostationary state after blue irradiation is <100% E isomer. While we were unable to control this temperature apart from implementation of a fan, we observed no more than a 3 o C increase in temperature at the sample as measured by as measure by a fluke 62 max IR thermometer.    Figure S16. Stress relaxation profile of gel prepared from P1 and P2 (1:1, 10 w/v% in 0.1M PBS, pH 7.5, 10% strain) after 120, 150, and 260 minutes of UV irradiation. Figure S17. Left: Short-interval cycling data with gels fabricated from P1 and P2 (1:1, 10 w/v% in 0.1M PBS, pH 7.5, 10% strain, 25 rad/s). Each cycle represents 30 minutes of gelation with 365 nm light followed by 2 minutes of 470 nm blue light. Right: Attempted temporal control experiment of gel prepared from P1 and P2 (1:1, 10 w/v% in 0.1M PBS, pH 7.5, 10% strain, 25 rad/s).
There is an induction period prior to the beginning of gelation when the UV light is turned on, and when UV light is removed, the gel continues to stiffen at a slower rate. The induction period may be due to the optical density of the macroscopic gel: some time is required for enough light to pass through the gel to photoswitch enough end groups, and cause a macroscopic response. We hypothesize that the gel continues to stiffen after the light is turned off because of hindered diffusion of the polymer chains through the macroscopic sample, and/or because isomerization and crosslinking are separate steps. The sample is not homogeneous because of its optical density, and more Z azobenzenes exist closer to the bottom of the sample (where light enters) than at the top. When the light is turned off, polymer chains continue to diffuse and the sample becomes more homogeneous, which appears as stiffening. Alternatively, the thermal esterification of the boronic acids to stiffen the gel occurs more slowly than the isomerization photoreaction, so when the light is turned off, the isomerization stops but the crosslinking continues some time. We will test these hypotheses in future work. Figure S19. Left: UV and blue cycle of a mixture of P1 and P3 (1:1, 10 w/v% in 0.1M PBS, pH 7.5, 10% strain, 25 rad/s). From this data, we believe that both UV and blue light are decreasing storage modulus by secondary thermal affect, where the change in G' is no more than 16 Pa using the same irradiation method as for the photocontrolled system. A hand-held IR thermometer measured a maximum 3 ºC temperature increase when the blue light was turned on. Right: Frequency Sweep of sample after irradiation cycle (10% strain, 300 to 0.1 rad/s).

S30
VI. Absorption Spectra Figure S23. Left: Normalized UV-Vis absorbance spectra of (E)-1 (8 µM in acetonitrile) before (black) and after (red) five minutes of irradiation. Right: Normalized UV-Vis of (E)-2 (8 µM in acetonitrile) before (black) and after (red) five minutes of irradiation. The resolution of spectra before and after irradiation suggest why the ratio of Z/E for compound 1 is higher than that of 2 in the kinetic experiments.  Samples were diluted so that UV-Vis detector saturation did not occur. In all cases, the UV-Vis spectra indicate that E-to-Z isomerization is occurring in the polymers.

S62
VIII. Photos Figure S26. Photo of mixtures of P1 and P2 (left), P1 and P3 (middle), and P1 and P4 (right) (1:1, 10 w/v% in 0.1M PBS, pH 9) before and after ten minutes of 365 nm UV irradiation. A qualitative change in appearance is only seen in the mixture of P1 and P2. Figure S27. Pictures of the photorheology setup used to characterize photoresponsive hydrogels. The sample sits between the quartz slide and the cone plate. The 365 nm UV flashlight and 470 nm blue LED can be turned on and off and provide irradiation from below the sample. Green LEDs were installed in the same fashion as blue LEDs.
S63 Figure S28. Photo of a gel formed from irradiation of P1 and P2-F2 (1:1, 10 w/v% in 0.1M PBS, pH 7.5) for one hour with green LED, after standing one week in the dark at 23 ºC. As indicated in by the flow inversion method, the network remains a gel.