Non-equilibrium dissipative supramolecular materials with a tunable lifetime

Many biological materials exist in non-equilibrium states driven by the irreversible consumption of high-energy molecules like ATP or GTP. These energy-dissipating structures are governed by kinetics and are thus endowed with unique properties including spatiotemporal control over their presence. Here we show man-made equivalents of materials driven by the consumption of high-energy molecules and explore their unique properties. A chemical reaction network converts dicarboxylates into metastable anhydrides driven by the irreversible consumption of carbodiimide fuels. The anhydrides hydrolyse rapidly to the original dicarboxylates and are designed to assemble into hydrophobic colloids, hydrogels or inks. The spatiotemporal control over the formation and degradation of materials allows for the development of colloids that release hydrophobic contents in a predictable fashion, temporary self-erasing inks and transient hydrogels. Moreover, we show that each material can be re-used for several cycles.


Reaction 4 (k 4 and k 5 )
Hydrolysis of anhydride proceeded with a (pseudo)-first order rate as determined by HPLC. We found that for Fmoc-D, this rate was slowed down two orders of magnitude if assemblies were formed. When small amounts of fuel were added and no assemblies were formed, the HPLC data could be fitted with k 4 . However, when a greater batch of fuel was added (3mM or more), we could not fit the data and had to implement a 1 st order k 5 for the hydrolysis reaction in the models that acts when the concentration anhydride was >0.08 mM. In other words, the assemblies exerted negative feedback on their own degradation (they slowed down hydrolysis).
For Fmoc-E, similar behavior was found. At low fuel concentrations, the model could fit the data well with k 4 . As soon as we added 5 mM or more fuel, the concentration anhydride reached values higher than 0.13 mM and the data could not be fit. Here, the behavior was somewhat different than for Fmoc-D. We needed to implement a 0 th order k 5 that only acted when the concentration anhydride was > 0.13 mM in order to fit the data well. This can be rationalized by a mechanism where the assemblies exclude water from the anhydrides, protecting them from degradation. They do that such that hydrolysis is limited by the solubility of the anhydride.

Supplementary Figure 13: Representative HPLC traces of Fmoc-AAD after EDC addition. HPLC traces of 10 mM Fmoc-AAD after 11 minutes of EDC addition (10 mM). EDC consumption was monitored at 220 nm (a) while changes in precursor and anhydride concentration were monitored at 254 nm (b).
Chromatogram represents a linear gradient water: ACN from 40:60 to 2:98.

Supplementary Figure 14: Representative HPLC traces of Fmoc-AVE after EDC addition. HPLC traces of 10 mM Fmoc-AVE after 11 minutes of EDC addition (10 mM). EDC consumption was monitored at 220 nm (a) while changes in precursor and anhydride concentration were monitored at 254 nm (b).
Chromatogram represents a linear gradient water: ACN from 40:60 to 2:98.

Supplementary Figure 15: Representative HPLC traces of Fmoc-AVD after EDC addition. HPLC traces of 10 mM Fmoc-AVD after 11 minutes of EDC addition (10 mM). EDC consumption was monitored at 220 nm (a) while changes in precursor and anhydride concentration were monitored at 254 nm (b).
Chromatogram represents a linear gradient water: ACN from 40:60 to 2:98.
Synthesis was performed on a CEM Liberty microwave-assisted peptide synthesizer. The first amino acid coupling to the resin was accomplished by using symmetrical anhydride methodology. Briefly, a 0.2 M solution of the Fmoc-amino acid symmetrical anhydride was prepared by allowing the corresponding Fmoc-protected amino acid (FmocE(OtBu)OH or FmocD(OtBu)OH, 12 mmol) and N,N′-diisopropylcarbodiimide (DIC, 6 mmol) to react in 30 mL N,N-dimethylformamide (DMF) for 40 min. The solution was placed in the freezer for 15 min and the solid urea formed was filtered out before next step. Loading of the resin was performed using the automated peptide synthesizer. The symmetrical anhydride solution (0.2 M, 12 mL) and 4-(dimethylamino)pyridine (DMAP) solution in DMF (20 mM, 2.5 mL) were added to the pre-swollen Wang resin (0.5 mmol, 1.1 mmol/g) and heated in the microwave (30 min, 75 ºC). The coupling was repeated twice. The resin was then washed with DMF (2x10 mL). Following couplings were achieved using 4 equivalents (eq.) of Fmoc-protected amino acid in DMF, 4 eq. of DIC and 4 eq. of ethyl (hydroxyimino)cyanoacetate (Oxyma). The resin solution was then heated in the microwave (1x2 min, 90ºC). Fmoc removal was accomplished using a solution of 20% piperidine in DMF (1x2 min, 90ºC). The resin was washed with DMF (3x7 mL) between different steps. Tripeptides were cleaved from the resin using a mixture of 95% trifluoroacetic acid (TFA), 2.5% water, and 2.5% triisopropylsilane (TIPS). The solvent was removed by co-distillation with ether by rotary evaporation and dried under reduced pressure. The product was purified using reversed-phase high-performance liquid chromatography (HPLC, Thermofisher Dionex Ultimate 3000, Hypersil Gold 250x4.8 mm) in a linear gradient of acetonitrile (ACN, 40% to 98%) and water with 0.1% TFA. Purified product was lyophilized and stored at -20 °C until further use. The purity of the tripeptides was analysed by electrospray ionization mass spectrometry in positive mode (ESI-MS) as well as analytical HPLC (Thermofisher Dionex Ultimate 3000, eluted with a gradient of 0.1% TFA in water: ACN from 40:60 to 2:98 in 10 min, see below for results).
Sample preparation. Stock solutions of the precursor were prepared by dissolving the precursor in 200 mM MES buffer, after which the pH was adjusted to pH 6.0. Stock solutions of EDC were prepared by dissolving the EDC powder in MQ water. Typically, stock solutions of 1.0 M EDC were used freshly. Reaction networks were started by addition of the high concentration EDC to the peptide solution.
HPLC. The kinetics of the chemical reaction networks were monitored over time by means of analytical HPLC (HPLC, Thermofisher Dionex Ultimate 3000, Hypersil Gold 250 x 4.8 mm). A 750 µL sample was prepared as described above and placed into a screw cap HPLC vial. Every 10 minutes, samples of these solutions were directly injected without further dilution, and all compounds involved were separated using a linear gradient water: ACN from 40:60 to 2:98. In order to avoid aggregation problems during the injection, samples of Fmoc-D were continuously stirred. In the case of the gelators, gels were just slightly broken manually by shaking the vial before the injection took place. Calibration curves for fuels (λ= 220 nm) and precursors (λ= 254 nm) were performed in triplicate in order to quantify the compounds over time. Calibration was not possible for the anhydrides due to their intrinsic instability. Instead, the absorption coefficient of their corresponding precursor was used. Measurements were performed at 25 ºC.
UV/Vis Spectroscopy. The UV/Vis measurements were carried out using a Genesys 10S (ThermoFisher) UV-VIS spectrophotometer. Samples were prepared as described above and placed in a 10 mm quartz cell (HellmaAnalytics) and stirred while acquiring data. Data was measured at 600 nm with an interval of 7 seconds.

DLS. DLS measurements on
Fmoc-E solutions were performed using a DynaPro NanoStar from Wyatt with a laser wavelength of 658 nm. The MES buffered samples were measured using the disposable cuvette for DLS from Wyatt. Each measurement consisted of 5 acquisitions with an acquisition time of 10 s. For measurements and analysis the software Dynamics V7 was used. Measurements were performed at 25 ºC.

Rheology.
Rheological measurements were carried out on a stress controlled rheometer (MCR 302, Antor Paar, Graz, Austria) using a plate-plate geometry (PP25, Anton Paar, Graz, Austria) and a plate separation of 0.3 mm. Samples were prepared as previously mentioned and placed between the two plates. Frequency and strain values were fixed to 1 Hz and 10 % respectively. A solvent trap was placed around the sample holder to avoid evaporation. Data was recorded at 25 ºC. The reusability experiments were performed by first preparing gels in an Eppendorf tube and, after at least 24 hours, the second, third or fourth gel was prepared in the rheometer using the above described procedure.
Confocal Fluorescence Microscopy. Confocal fluorescence microscopy was performed on a Leica SP5 confocal microscope using a 63x oil immersion objective. Samples were prepared as described above, but with 25 µM Nile Red as dye. 20 µL of the sample was deposited on the glass slide and covered with a 12mm diameter coverslip. Samples were excited with 543 nm laser and imaged at 580-700 nm.
Fluorescence Microscopy. For samples that used both Nile Red and Pyrene as dye, we used conventional fluorescence microscopy. Sample preparation was performed as described above and micrographs were acquired on a Leica DMi8 microscope using a 100x oil immersion objective.
In order to measure the cumulative release of Nile Red and pyrene, we used the solvatochromic properties of these dyes when free in water. Over time, as the particles hydrolysed, the Nile Red and pyrene signal decreased as it was expelled into the water phase. The signal was normalized as a measured for release of dye. Nile Red was excited at 550 nm and its emission was traced over time at 630 nm, while pyrene was excited at 330 nm and its emission traced at 384 nm. Measurements were performed at 25 ºC.
The Nile Red assay was performed on the same spectrofluorimeter as described above. Samples were directly prepared in the 10 mm quartz cuvette (Precision Cells Inc.) by mixing different concentrations of precursor (from 2.5 to 15 mM in MES 0.2 M) with Nile Red (5 µM). The fluorescence intensities were measured at 635 nm with and excitation at 550 nm.
The THT assay was performed on the same spectrofluorimeter as described above. Samples were directly prepared in the 10 mm quartz cuvette (Precision Cells Inc.) by mixing precursor (10 mM in MES 0.2M) with EDC (100 mM) and THT (5 µM). The fluorescence intensities were measured over time, every 2 minutes, at 485 nm with excitation at 450 nm.
The Fmoc assay was performed on the same spectrofluorimeter as described above. Samples were directly prepared in the 10 mm quartz cuvette (Precision Cells Inc.) by mixing precursor (10 mM in MES 0.2M) with EDC (100 mM). The fluorescence intensities were measured over time, every 5.8 minutes, from 300 to 600 nm with excitation at 285 nm.
Circular Dichroism Spectroscopy. Circular dichroism measurements were performed on a Jasco (Jasco J750) equipped with a Peltier temperature Control. Samples were prepare as described above and placed into a 1 mm quartz cell (Hellma Analytics) cuvette. Spectra were recorded from 325 to 280 nm with 0.2 nm step, 1 nm bandwidth. For single measurements, final spectra are the average of 10 accumulations at scan speed of 50 nm/min. For kinetic measurements spectra were recorded every 2 minutes with no accumulation at scan speed of 100 nm/min. Measurements were performed at 25 ºC.
Spray Coating. For the preparation of the self-erasing medium, Fmoc-D was immobilized in a 30% polyacrylamide hydrogel. Briefly, a solution of 30% (29:1) acrylamide: bisacrylamide was prepared. In this solution, 10 mM Fmoc-D and 200 mM MES was dissolved. To prepare a gel, 4mL of this stock solution was deposited in a petridish and 8µL tetramethylethylenediamine (TEMED) was added. The gelation was started by addition of 60µL 10w/w/% ammonium persulfate in water. The reaction was allowed to proceed for at least 2 hours, before EDC was spray coated or painted.
The fuel deposition via spray coating was performed using a spray gun (Harder & Steenbeck GmbH & Co. KG, Grafo T3) and oil-free nitrogen as carrier gas. The spray gun was fastened on a mount above a vertically adjustable sample stage to fixate the nozzle-to-sample distance at 15 cm. Applying a constant nitrogen pressure of 2 bar, the flow rate of the solution was set to about 15 µL s -1 by tuning the nozzle diameter accordingly. The spray time and thereby the amount of deposited material was defined by an electronic controller. About 300 µL of the respective solution were deposited on the substrate using 2 spray shots of 10 s duration with a short pause. 3D-printed masks were placed onto the substrates in order to create reproducible, detailed images. After the completed spray deposition, the substrates were immediately transferred to an imaging station for further analysis.
ESI. ESI-MS measurements were performed using a Varian 500 MS LC ion trap spectrometer. The samples were diluted in acetonitrile and injected into an acetonitrile carrier flow (20 µL/min).

Cryogenic-Transmission Electron Microscopy (cryo-TEM).
Samples for TEM were prepared as described above. Shortly before imaging the samples were diluted a 10-fold to decrease the density of fibers in the micrographs. Cryo-TEM imaging was performed on a Jeol JEM-1400 plus operating at 120 kV. The images were recorded in a low-dose mode on a CCD camera. Quantifoil R2/2 on Cu-grid 400 mesh were used. The grids were freshly glow-discharged for 30 seconds prior to use. Preparation of the grids was performed in a FEI Vitrobot at 21 °C with the relative humidity set to 100% and the blotting force was set to -5. The sample (5 µL) was incubated for 30 seconds, blotted twice for 3.5 seconds and then directly plunged into liquid ethane that was pre-cooled by liquid nitrogen. The cryo-EM grids were transferred and stored in liquid nitrogen, and when needed, placed into a Gatan 625 cryo-specimen holder to insert into the microscope. The specimen temperature was maintained at -170 °C during the data collection.