Memory Effect in Thermoresponsive Proline‐based Polymers

Abstract We report that synthetic polymers consisting of L‐proline monomer units exhibit temperature‐driven aggregation in water with unprecedented hysteresis. This protein‐like behavior is robust and governed by the chirality of the proline units. It paves the way to new processes, driven by either temperature or ionic strength changes, such as a simple “with memory” thermometer.


Table of Contents
All chemicals and solvents in this work were purchased from Sigma Aldrich, Fluorochem, Acros, TCI, Strem and VWR, and were used without any purification unless otherwise described. Acetonitrile (ACN), was obtained from a solvent system purifier (PureSolv, Innovative Technology), kept under argon atmosphere and freshly used. Milli-Q water was obtained from a Purelab Prima purification system (ELGA) with a resistivity of 18.2 MΩ cm -1 . The N-carboxyanhydride monomers were purchased from PMC Isochem, stored at -20 °C under argon atmosphere and weighed in the glove box Jacomex GP13 no. 2675 at the Laboratoire de Chimie des Polymères Organique (LCPO, Bordeaux, France). Dialysis membranes were purchased from SpectrumLabs. Hexylamine, benzylamine, allylamine and 2-(2-amino ethoxy) ethanol were all freeze-thawed and cryo-distilled on the Schlenk line prior to use.

Size Exclusion Chromatography (SEC) Analyses
Polymer molar masses were determined by Size Exclusion Chromatography (SEC) using an aqueous solvent as the eluent. Measurements were performed on an Ultimate 3000 system from Thermoscientific equipped with diode array detector DAD. The system also includes a multi-angle light scattering detector (MALS) and differential refractive index detector (dRI) from Wyatt technology. Polymers were separated on two TOSOH TSK Gel G4000PWXL and G3000PWXL columns (300 × 7.8 mm) (exclusion limits from 200 Da to 300 000 Da) at a flowrate of 0.6 mL/min. Aqueous solvent composed of acetic acid (AcOH) 0.3 M, ammonium acetate 0.2 M and ACN (H2O/ACN: 6.5/3.5, v/v) was used as the eluent. Column temperature was held at 25°C. dn/dc measurements were performed on the differential refractive index detector dRI from Wyatt technology by injecting 500 μL of each sample dissolved in the aqueous phosphate buffer at 0.5-5 mg mL −1 . The chromatograms were recorded with Chromeleon 7.2 software and Astra 7.1.0 software and analyzed using the latter. dn/dC values calculated were of 0.1616 mg/ml for L-Proline homopolymers and copolymers with D-Proline and 0.1444 mg/ml for block copolymers with PEG.

UV−Vis Spectroscopy
The cloud point temperature (TCP) and the clearing point temperature (TCL) of PLP solutions in pure water were determined by measuring the turbidity at 550 nm between 10°C and 85°C at a 1 °C.min -1 scan rate at different concentrations of polymer and salt. Data were collected on a Cary 100 UV−Vis spectrophotometer equipped with a multicell thermoelectric temperature controller from Agilent Technologies. TCP was defined as the temperature at the inflection point of the transmittance-temperature curves during heating ramps and concordantly for TCL during the cooling ramps. A typical experiment followed the following heating program: 1-10 min isotherm at 10°C 2-1°C/min ramp to 85°C 3-10 min isotherm at 85°C 4-1°C/min ramp to 10°C Absorbance was plotted in function of temperature, with isotherms' data not collected. Subsequently, absorbance was converted to transmittance for a better visualization of the degree of turbidity and to discriminate between complete and incomplete aggregations. The y-axis in the results is displayed as 1-Transmittance (%) to help the reader relate the soluble and the aggregation form to low and high values respectively.

FTIR spectroscopy
The IR spectra were recorded using the FTIR spectrometer (Vertex 70, Bruker) equipped with a temperature control cell, and the samples were measured with the ATR (GladiATR, Pike Technologies) from Fisher technologies performing 64 scans at 2 cm -1 in solution at the LCPO (Bordeaux, France). The plate and pressure system of the ATR were equipped with a liquid sample set (Pike Technologies) bought from eurolabo to keep the solvent from evaporating during the heat treatment and thus analyse the sample in the liquid state as other experiments. The raw data were obtained with the Opus7.5 software and processed using the Originlab 2016 software. Samples were dissolved in water and analysed as such on the ATR plate. A background with pure water was run before each sample in the same manner so as not to mask the vibration bands of the C=O bonds between 1500 cm -1 and 1700 cm -1 .

MALDI-MS measurements
MALDI-MS spectra were performed at CESAMO facility (Bordeaux, France) on an Autoflex maX TOF mass spectrometer equipped with a frequency tripled Nd:YAG laser emitting at 355 nm. Spectra were recorded in the positive-ion mode with an accelerating voltage of 19 kV. For MALDI-MS analyses, polymers' deposits were prepared by dissolving 10 mg of polymer in 1 ml of Milli-Q water, 10 mg of NaI in 1 ml methanol and 22 mg of sinapinic acid in methanol. Subsequently 2 µL of polymer solution was mixed with 2 µL of salt solution and added to 20 µL of matrix solution. 1 µL of this mixture was deposited on the MALDI plate and dried.

CD spectroscopy
The CD measurements were performed on a JASCO J-815 spectropolarimeter between 190 nm and 260 nm (far-UV). A quartz cell of 1 mm path length (type: 21/10/Q/1) was purchased from Starna Scientific, Ltd. Spectra were recorded at desired temperatures: 20 °C for standard measurements, or a temperature gradient between 10 °C and 90 °C. The measurement parameters were optimized as follows: sensitivity between 5 and 200 mdeg, 0.01 mdeg resolution, 8 seconds response time (Digital Integration Time), 1 nm bandwidth and 10 nm min −1 scanning rate. Polymer solutions at a monomer unit concentration of 2.5 µM in Milli-Q water were used for the measurements. For the turbidity tests, the concentration of the samples had to be adapted for this analysis to a concentration of 0.25 mg/mL to avoid flocculation of the polymers. Thus, an optimum concentration was found that neither saturated the signal nor brought the cloud point temperature above reactional conditions.          Polymers 7-9 having small molar masses, SEC analysis was deemed unfit for the calculation of Mn and thus MALDI-MS was used to estimate the Mn and to confirm the initiation by the amine via the end chain( Figure S9). For diglycol-PLP75 10 and allyl-PLP75 11, GPC showed two populations ( Figure S10). The second (smallest population) was determined to be polymer chains initiated by water by MALDI-MS via the end-chain group determination ( Figure S11). Nonetheless, this did not have any considerable effect on the TCP and the overall thermal behaviour (see Table S2). Only in the case of low DP polymers (7-9) we can observe that the hydrophobic hexylamine end chain (hexyl-PLP20 7) decreases slightly the TCP. Polymers 10-11 to be compared with hexyl-PLP70 4 ( Table 1 and Table S1). [a] Absolute number-average molar mass (Mn) and dispersity (Đ) determined by SEC using a multi-angle static light scattering detection. [e] Transitions that were observed to have started before the 10 min stabilization at 85°C, but it was impossible to calculate the derivative of the curve to obtain the temperature value at the inflexion point giving the TCP.
[d] The shouldered peaks shown in Figure S10 were treated by the softwre as two different peaks and thus their Mn and Đ are presented separately here.   All polymers containing a mixture of both D-and L-Proline repeat units showed no thermoresponsive behaviour, confirming our hypothesis of the conformational role on this property. The LCST behaviour of hexyl-PDP100 20 was comparable to that of its L counterpart; polymer 5 ( Figure S22).  . CD spectra of polymers 15-19 and hexyl-PLP 5 at room temperature in milliQ water at a 0.25mg/ml concentration. All polymers show a PLPII conformation except for hexyl-P(50L/50D)P 19 as the signal is canceled out by the presence of an equal quantity of both enantiomers. Figure S24. Evolution of CD peaks of hexyl-PLP50 1 during heating and cooling of a 0.25mg/ml in milliQ water sample. The peak at 208nm (typical of PLPII soluble) sharply decreases in intensity at higher temperatures (80°C) during the heating ramp, while the peak at 208nm (attributed to the aggregated form) sharply increases in its intensity around the same temperature. Both peaks have a constant value in the cooling ramps from 90°C to around 30°C, at which time the tendency inverses showcasing the conformation hysteresis. Wavelength (nm) hexyl-PLP100 5 hexyl-P(90L/10D)P 15 hexyl-P(80L/20D)P 16 hexyl-P(70L/30D)P 17 hexyl-P(60L/40D)P 18 hexyl-P(50L/50D)P 19 hexyl-PDP100 20 Figure S25. FTIR spectra of polymer solutions in H2O. hexyl-PLP20 21 (polymerized in dry ACN) is in PLPI conformation and hexyl-PLP50 1 in PLPII conformation. hexyl-PLP20 21 was then precipitated by heating and brought to room temperature (spectra is shown in red). The amide bond stretching wavelength clearly indicate the presence of both conformations.