Adaptable Fast Relaxing Boronate‐Based Hydrogels for Probing Cell–Matrix Interactions

Abstract Hydrogels with tunable viscoelasticity hold promise as materials that can recapitulate many dynamic mechanical properties found in native tissues. Here, covalent adaptable boronate bonds are exploited to prepare hydrogels that exhibit fast relaxation, with relaxation time constants on the order of seconds or less, but are stable for long‐term cell culture and are cytocompatible for 3D cell encapsulation. Using human mesenchymal stem cells (hMSC) as a model, the fast relaxation matrix mechanics are found to promote cell–matrix interactions, leading to spreading and an increase in nuclear volume, and induce yes‐associated protein/PDZ binding domain nuclear localization at longer times. All of these effects are exclusively based on the hMSCs' ability to physically remodel their surrounding microenvironment. Given the increasingly recognized importance of viscoelasticity in controlling cell function and fate, it is expected that the synthetic strategies and material platform presented should provide a useful system to study mechanotransduction on and within viscoelastic environments and explore many questions related to matrix biology.


Supporting Figures Cited in the Main Text
. Rheological characterization showing the impact of catechol oxidation on the viscoelastic properties of hydrogels synthesized from octa-arm PEG polymers functionalized with CAT and FPBA. The shift of G" max to the low-frequency end and the decrease in the normalized G" suggest that gels lose energy dissipation capacity over time and become more elastic, likely due to catechol oxidation. Hydrogel concentration was at 10% (w/v), and the temperature was 37°C.  After the two macromolecular gel precursors were mixed and time-sweep test was set up, which took approximately 20 s, G' had been larger than G" at the onset of the experiment. This suggests that gelation of boronate-based hydrogels is very rapid. Gel concentration was at 10% (w/v). Experiment was performed at 37°C, with a fixed angular frequency of 10 rad s -1 and 1% strain.     Figure S11. Immunofluorescence images of spread hMSCs cultured in the stress-relaxing hydrogels for 7 days: F-actin (orange), nucleus (blue) and YAP/TAZ (magenta). Note that the scale bars in these images are different. Figure S12. Scatter plots showing correlations between cell volume, nuclear volume, and YAP/TAZ nuclear to cytosolic ratio for hMSCs encapsulated in elastic and stress relaxing gels. Figure S13. Quantification of (a) cell volume, (b) cell sphericity, (c) nuclear volume, (d) nuclear sphericity, and (e) YAP/TAZ nuclear to cytosolic intensity ratio at the corresponding condition. Statistical analysis from one-way ANOVA test, ** and **** indicating p < 0.01 and 0.0001, respectively. Three replicates per condition, and over 30 cells analyzed per replicate. Compared to Figure 5 shown in the main text, the cell and nuclear volumes are smaller in all conditions, perhaps due to the differences in donor race and gender.

Materials
Octa-arm PEG-hydroxyl (M n = 20 kg mol -1 , tripentaerythritol core) and octa-arm PEG-amine (M n = 20 kg mol -1 , hexaglycerol core or tripentaerythritol core) were purchased from JenKem Technology USA. DBCO (dibenzylcyclooctyne) acid, 3-azidopropinoic acid, and azido propylamine were purchased from Click Chemistry Tools. All chemicals for peptide synthesis were purchased from Chem-Impex International Inc. All other chemicals were purchased from Sigma Aldrich or Fisher, and used as received unless otherwise noted. Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectra were obtained on a Voyager DE mass spectrometer (Applied Biosystems). α-Cyano-4-hydroxycinnamic acid (CHCA) was used matrix. High resolution mass spectra (HRMS) were acquired on a Waters Synapt G2 high definition mass spectrometer.

Instrumentation and Characterization
Rheology experiments were performed on a DHR-3 rheometer (TA instrument) with an 8-mm stainless steel upper plate, and the temperature was controlled by a Peltier bottom plate. Inertial calibration and motor adjustment were performed before each experiment. Frequency sweep experiments were performed at 1% strain (within the linear viscoelastic regime as pre-determined from strain sweep experiments) from 100 to 0.01 rad s -1 . For time-temperature superposition, experiments were performed at three different temperatures, namely 10, 25 and 37°C. For measurements of the mechanical properties of swollen hydrogels, adhesive-back sandpaper (600 grit) was applied to both the upper and lower plates. Data with torque values larger than 500 times the minimal torque were reported.

Synthesis of Nitrodapamine
The synthesis of nitrodapamine was performed according to a published procedure [1] with some modifications. Dopamine (500 mg, 2.64 mmol) and NaNO 2 (630 mg, 9.13 mmol) were dissolved in MilliQ water (15 mL) and cooled with an ice bath. Then 20% (v/v) sulfuric acid (2.5 mL) was carefully added to the previous solution under vigorous stirring. After ca. 15 min, precipitates were collected by vacuum filtration and washed with cold water three times and then with cold MeOH three times to afford orange solid (400 mg, yield: 51%). 1

Synthesis of hydroxyethylaminomethylphenylboronic acid
Similar to procedures reported elsewhere [2] , to a flame-dried round bottom flask was added 2formylphenylboronic acid (100 mg, 0.67 mmol), anhydrous MgSO 4 (600 mg), and anhydrous MeOH (10 mL). The reaction was cooled to 0°C under an argon atmosphere. Ethanolamine (80.5 μL, 1.33 mmol) was dropwise added to the mixture. The reaction was stirred at 0°C for 30 min and then at room temperature for 3 h. The reaction was again cooled to 0°C, and NaBH 4 (100 mg, 2.63 mmol) was added in four portions at 10 min intervals. After being stirred at room temperature for another 2 h, the reaction was quenched with 1 mL water. The mixture was then filtered through a celite pad and eluted with MeOH. The crude product was further purified by reverse phase HPLC using a Biotage SNAP ultra C18 column with a gradient of 5-100% MeCN in water with 0.1% TFA. Lyophilization of the combined pure fraction afforded a slightly yellow liquid (60 mg, yield: 46%). 1
a) Synthesis of 3-bromo-4-methylbenzonitrile [3] To a stirred solution of 4-methylbenzonitrile (5.0 g, 42.7 mmol) in H 2 O (35 mL) and sulfonic acid (35 mL), N-bromosuccinimide (7.60 g, 42.7 mmol) was added drop wise in dark and the mixture was allowed to react at 25°C for 12 h. The reaction mixture was quenched by sodium bicarbonate (100 mL) and then washed with ethylether (3 × 50 mL). The collected organic layer was then dried over anhydrous MgSO4, filtered and concentrated to afford the 3-bromo-4-methylbenzonitrile as a white solid ( b) Synthesis of 3-bromo-4-(bromomethyl)benzonitrile [4] To a stirred solution of 3-bromo-4-methylbenzonitrile (7.13 g, 36.4 mmol) in chloroform (73 mL), Nbromosuccinimide (7.38 g, 41.5 mmol) was added, and the mixture purged under nitrogen system for 30 min. Then, azobisisobutyronitrile (418 mg, 2.55 mmol) was added to the reaction mixture and refluxed for 12 h. The reaction mixture was then quenched by brine (100 mL) and extracted with dichloromethane (3 × 50 mL). The collected organic layer was then dried over anhydrous MgSO 4 , filtered and concentrated to afford the crude mixture and then purified by flash column chromatography to afford the desired product as a white solid (5.82 g, 58% c) Synthesis of 3-bromo-4-(hydroxymethyl)benzonitrile [5] To a stirred solution of 3-bromo-4-(bromomethyl)benzonitrile (5.82 g, 21.2 mmol) in water (106 mL) and 1,4-dioxane (70 mL) co-solvent system, calcium carbonate (921 mg, 9.20 mmol) was added and reacted at 100°C for 12h. The reaction mixture was diluted with ethyl ether (50 mL) and washed by ddH 2 O. The collected organic layer was then dried over anhydrous MgSO 4 , filtered and concentrated to afford the crude mixture and then purified by flash column chromatography to afford the desired product as a white solid (3.6 g, 80%

Synthesis of Octa-Arm PEG-Azide
The synthesis of octa-arm PEG-azide was performed according to a published procedure [7] for the synthesis of tetra-arm PEG-azide. In the first step, octa-arm PEG-hydroxyl (5.00 g, 0.25 mmol, M n = 20 kg mol -1 ) was dried at 80 °C under vacuum overnight. After the compound was cooled down to room temperature, anhydrous DCM (150 mL) was added to completely dissolve the polymers. Then triethylamine (TEA, 2.50 mL, 18 mmol) and mesyl chloride (1.24 mL, 16 mmol) were added to the solution. The reaction was stirred under argon overnight. The mixture was then diluted with 300 mL DCM and washed with brine (4 × 50 mL). The organic phase was dried over Na 2 SO 4 , filtered, and concentrated under vacuum. Polymers were recovered by precipitation into diethyl ether twice to give slightly yellow solid (5.01 g, yield: 97%). 1  In the second step, octa-arm PEG-mesylate (2.50 g, 0.125 mmol), NaN 3 (0.39 g, 6 mmol) and anhydrous DMF (25 mL) were added to a round bottom flask. The reaction was stirred at 60°C for 2 days. The solution was then dialyzed again water extensively (MWCO 8 kDa), and then lyophilized to give white (2.38 g, yield: 95%). The complete disappearance of the mesyl peaks suggests that the end group conversion was quantitative.

Synthesis of Octa-Arm PEG-DBCO
To a reaction vial equipped with a stir bar was added DBCO acid (91.5 mg, 0.3 mmol), HATU (114.1 mg, 0.3 mmol) and anhydrous DMF (1.5 mL). The solution was stirred at room temperature for 10 min and then transferred to a separate flask containing octa-arm PEG-amine HCl salts (500 mg, 25 μmol, M n = 20 kg mol -1 , tripentaerythritol core). The vial was rinsed with DMF (1.5 mL), and the solvent was added to the flask. DIPEA (279 μL, 1.6 mmol) was added to the mixture. The reaction was stirred under argon for 1.5 days at room temperature. Polymers were precipitated into cold diethyl ether, air dried, dialyzed against water extensively (MWCO 8 kDa), and then lyophilized to give yellow power (487 mg, yield: 87%). 1

Synthesis of Octa-Arm PEG-Boroxole
To a reaction vial equipped with a stir bar was added boroxole carboxylic acid (45 mg, 0.25 mmol), HATU (96 mg, 0.25 mmol), and anhydrous DMF (1.5 mL). The solution was stirred at room temperature for 10 min and then transferred to a separate flask containing octa-arm PEG-amine HCl salts (316 mg, 15.8 μmol, M n = 20 kg mol -1 , hexaglycerol core). The vial was rinsed with DMF (0.5 mL), and the solvent was added to the flask. DIPEA (200 μL, 1.15 mmol) was added to the mixture. The reaction was stirred under argon for 2 days at room temperature. Polymers were precipitated into cold diethyl ether twice, air dried, dialyzed against water extensively (MWCO 8 kDa), and then lyophilized to give white power (296 mg, yield: 88%). 1

Synthesis of Octa-Arm PEG-o-Aminomethylphenylboronic Acid
The synthesis of octa-arm PEG-o-aminomethylphenylboronic acid was performed according to a published procedure [8] . To a reaction vial equipped with a stir bar was added octa-arm PEG-amine HCl salts (250 mg, 12.5 μmol, M n = 20 kg mol -1 , hexaglycerol core), 2-formylphenylboronic acid (37.5 mg, 0.25 mmol), triethylamine (TEA, 200 μL, 1.44 mmol), and anhydrous MeOH (2.5 mL). The reaction was stirred under argon at room temperature for 3 days. After this time, the mixture was cooled to 0°C with an ice bath, and NaBH 4 (80 mg, 2.13 mmol) was added to the mixture in four portions at 10 min intervals.
The reaction was purged with Ar(g) again and left at room temperature for another 20 h. The reaction mixture was then dialyzed against water extensively (MWCO 8 kDa), filtered and lyophilized to give white powder (235 mg, yield: 92%). 1

Synthesis of Octa-Arm PEG-p-Nitrophenylcarbonate
Octa-arm PEG-hydroxyl (2.00 g, 0.1 mmol, M n = 20 kg mol -1 ) was dried at 80 °C under vacuum overnight. After the compound was cooled down to room temperature, anhydrous THF (60 mL) was added to completely dissolve the polymers. Then p-nitrophenyl chloroformate (1.29 g, 6.4 mmol) and TEA (1.00 mL, 7.2 mmol) were added to the solution. The reaction was stirred under argon at room temperature for 2 days. The purified product was obtained by precipitation into cold diethyl ether three times and drying under vacuum to yield off-white solid (1.95 g, yield: 98%). 1

Synthesis of Octa-Arm PEG-Azide/Fluorophenylboronic acid
To a reaction vial equipped with a stir bar was added 3-azidopropionic acid (4.3 mg, 37.5 μmol), HATU (14.3 mg, 37.5 μmol), and anhydrous DMF (1.5 mL). The solution was stirred at room temperature for 10 min and then transferred to a separate flask containing octa-arm PEG-amine HCl salts (500 mg, 25 μmol, M n = 20 kg mol -1 , tripentaerythritol core). The vial was rinsed with DMF (0.5 mL), and the solvent was added to the flask. DIPEA (279 μL, 1.60 mmol) was added to the mixture. The reaction was stirred under argon overnight at room temperature. Polymers were recovered by dialysis against water and then lyophilization. By 1 H NMR analysis and TNBS (2,4,6-trinitrobenzene sulfonic acid) assay, on average 2 out of the 8 arms were functionalized into azide.
In the second step, 4-carboxy-2-fluorophenyl boronic acid pinacol ester (100 mg, 0.38 mmol), HATU (142 mg, 0.38 mmol) were dissolved in anhydrous DMF (2 mL). The solution was stirred at room temperature for 10 min and then transferred to a separate flask containing the previously obtained polymers and DIPEA (200 μL, 1.15 mmol). The reaction was stirred under argon overnight at room temperature. Polymers were recovered by dialysis against water and then lyophilization. 1 H NMR analysis shows that the other 6 arms were functionalized with fluorophenylboronic acid.

Synthesis of Octa-Arm PEG-Azide/Nitrodopamine
To a round bottom flask was added octa-arm PEG-p-nitrophenylcarbonate (250 mg, 12.5 μmol), azidopropylamine (1.84 μL, 18.8 μmol), DIPEA (70 μL), and DMF (2 mL). The reaction mixture was stirred under argon at room temperature for 1.5 days. Then nitrodopamine (59.5 mg, 0.30 mmol) was added to the reaction mixture, which was purged with argon and left to stir at room temperature of 1.5 days. The solution was then dialyzed against water extensively (MWCO 8 kDa) and lyophilized to afford yellow powder. 1 H NMR analysis suggests that the ratio of azide to nitrodopamine in the final polymers was 1:3.

Determination of the small-molecule equilibrium constants from titration
The small-molecule boronate ester equilibrium constants were determined by UV-vis titration following published procedures [9] . All measurements were performed in 100 mM phosphate buffer (pH 7.4), identical to the condition used for preparing hydrogels. The dissociation constants K D were calculated based on the following equation: where ∆A = A -A 0 , the difference in absorbance at 425 nm, [ND] 0 is the initial concentration of nitrodopamine, [B] is the initial concentration of different boronic acids. The range of [B] varies to maximize the changes in absorbances for different boronic acids. In eq. S1, the two fitting parameters are K D and ∆A max . All fittings were performed in MATLAB. Representative titration and fitting results are shown in Figure S9 and S10. The derived values of K D are tabulated in Table S1.

Estimation of the high-frequency plateau modulus of hydrogels from modified phantom network theory
The phantom network theory estimates the network elastic modulus G ph as follows, where v is the number density of the elastically effective network strands, kT is the thermal energy, f is the junction functionality (f = 8 for octa-arm PEG polymers). In the case of reversible networks with crosslinks based on dynamic association and dissociation of boronate esters, v is affected by the equilibrium constants of the boronates. By assuming that the small-molecule boronates and the polymerconjugated boronates have the same equilibrium constants K D , the fraction of associated bonds ζ under each hydrogel condition can be calculated.
where [P] is the concentration of polymers at equilibrium (note that both boronic acid and nitro-dopamine functionalized polymers are at the same concentration), and f B and f ND are the number of functional groups on boronic acid and nitro-dopamine functionalized polymers, respectively (f B ≤ f ND ≤ 8 for the polymers synthesized in this work).
The fraction of associated bonds ζ can be solved from eq S3, which gives, where r is the stoichiometric imbalance, defined as r = f B /f ND .
In addition, where N A is the Avogadro constant.
By combining eqs S2, S4 and S5, the high-frequency plateau modulus of the reversible gels is estimated to be

Boronate equilibrium in hybrid gels in PBS and DMEM
Here, the concentration of FPBA-ND boronate pair is calculated in the following two scenarios, one in PBS without the presence of glucose, and the other in DMEM for cell studies, with a glucose concentration of approximately 5 mM.
As shown in Table S2, the concentrations of FPBA and ND were both 9 mM in the hybrid gels containing two types of crosslinks. Therefore, the equilibrium between FPBA and ND in PBS can be described as where the concentration of the FPBA-ND complex is the only unknown in eq S7. Solving eq S7 gives [FPBA-ND] = 7.67 mM , without glucose (S8) For the hybrid gels immersed in DMEM, two equilibrium equations should be solved simultaneously, one for FPBA-ND and the other for FPBA-glucose, which are where K D,2 is estimated to be 20 mM, according to the work by Yan et al. [10] The two unknowns in eq S9 and S10 are [FPBA-ND] and [FPBA-Glucose]. Solving S9 and S10 gives [FPBA-ND] = 7.54 mM , with glucose (S11) [FPBA-Glucose] = 0.28 mM (S12) Therefore, the presence of low concentration glucose only introduced a 3% difference in the concentration of adaptive FPBA-ND pairs.

Rheology data processing and analysis
The detailed method of fitting the rheology data to a stretched exponential function is described in detail elsewhere [11] . Briefly, a continuous relaxation spectrum H(τ) was first constructed based on the dynamic moduli in the master curve obtained from time-temperature superposition. Then, the relaxation modulus G(t) was calculated based on the relaxation spectrum using the following equation: The experimentally derived stress relaxation curve was then fit to a stretched exponential function, or the Kohlrausch-Williams-Watts (KWW) model, which provides an empirical formula to mathematically capture the time dependent stress relaxation modulus of the covalent adaptable networks: million cells per milliliter. For non-relaxing gels, stock solutions of octa-arm PEG-azide, azido-GRGDS, and the hMSC suspension were gently mixed, and then the octa-arm PEG-DBCO solution was quickly added. The mixture was pipetted several times and quickly transferred to a syringe barrel. After 10 min, the construct was equilibrated in pre-warmed media. For relaxing gels, stock solutions of octa-arm PEG-DBCO, azido-GRGDS, and the cell suspension were gently mixed. The mixture was transferred to a syringe barrel. Then octa-arm PEG-azide/FPBA and octa-arm PEG-azide/ND were sequentially added. The mixture was quickly triturated several times after adding each component. Constructs were allowed to gel at 37°C for 15 min, and were then placed into pre-warmed media. The final concentration for all components in non-relaxing and relaxing gels are shown in Table S2. Media was changed the first day after encapsulation and every other day afterwards.

Live/Dead Assay
Hydrogels were incubated in phenol red free DMEM media (supplemented with 10% FBS) with 0.5 μM calcein AM and 4.0 μM ethidium homodimer for 30 min per manufacturer's protocol. Hydrogels were imaged using an Operetta high-content imaging system. Viability was determined using this membrane integrity assay, and by counting live cells (green) and dead cells (red) in images stacked from maximum intensity projections (step size 5 μm, total thickness ~ 200 μm).

Immunostaining
Samples were fixed in 10% buffered formalin for 1 h at room temperature, washed with DPBS twice, and then permeabilized with 0.1% (v/v) Triton X-100 in PBS for 2 h at room temperature. Subsequently, constructs were incubated with 5% (w/v) BSA in PBS to block nonspecific binding, after which primary antibodies anti-YAP (mouse monoclonal, Santa Cruz Biotechnology, sc-101199, dilution of 1:250; note that this antibody was reported to stain both YAP and TAZ) in 5% (w/v) BSA were added to the constructs and incubated at 4°C overnight. Hydrogels were then rinsed with 0.5% (v/v) Tween-20 in PBS five times. Secondary antibody goat anti-mouse AlexaFluor® 647 (Invitrogen, dilution of 1:300), rhodamine phalloidin (Invitrogen, dilution of 1:300), and 4',6-diamidino-2-phenylindole (DAPI, sigma, 3.3 μg mL -1 ) were added to the constructs and incubated at 4°C overnight. Samples were finally rinsed with 0.5% (v/v) Tween-20 in PBS five times and stored in DPBS before imaging.

Image acquisition and analysis
Immunofluorescence images were taken using a laser scanning confocal microscope (Zeiss, LSM710) with a Plan-Apochromat 20× (NA = 1.0) water objective. Images were quantified using Imaris software (Bitplane). Three dimensional views were first reconstructed from z-stack images (step size 1.5 μm). The cell and the nucleus were determined from the rhodamine and the DAPI channels, respectively ( Figure S15). The volume and surface area of cell and nucleus are obtain from the Imaris software after 3D reconstruction.
The cell sphericity ψ is defined as the ratio of the surface area of a sphere with the same volume as the given cell to the surface area of the cell. The mathematical form is   (S18) The value of sphericity ψ ranges from 0 (a 2D object) to 1 (a perfect sphere). The nuclear sphericity is defined similarly.
The YAP/TAZ nuclear to cytosolic ratio was calculated based on the following formula, (S19) where I nuc and I cell are the total intensities of the YAP/TAZ signal inside the nucleus and the cell, respectively, and V nuc and V cell are the volumes of the nucleus and the cell, respectively.

Statistical analysis
Experiments were performed with three independent hydrogels, and more than 30 cells were analyzed per gel. One-way analysis of variance (ANOVA) tests were performed on all data sets, followed by Tukey post-hoc tests. Tukey box-whisker plots of single cell data contained boxes corresponding to the 25 th percentile and the 75 th percentile. Significance levels were indicated by *, **, *** and **** for p < 0.05, 0.01, 0.001, and 0.0001, respectively.