Molecular Engineering of pH-Responsive Anchoring Systems onto Poly(ethylene glycol) Corona

An adaptive surface that can sense and respond to environmental stimuli is integral to smart functional materials. Here, we report pH-responsive anchoring systems onto the poly(ethylene glycol) (PEG) corona of polymer vesicles. The hydrophobic anchor, pyrene, is reversibly inserted into the PEG corona through the reversible protonation of its covalently linked pH-sensing group. Depending on the pKa of the sensor, the pH-responsive region is engineered from acidic to neutral and basic conditions. The switchable electrostatic repulsion between the sensors contributes to the responsive anchoring behavior. Our findings provide a new responsive binding chemistry for the creation of smart nanomedicine and a nanoreactor.


Instruments
NMR is carried out on a Bruker AVANCE HD nanobay console with a 9.4 T Ascend magnet

S4
PEG corona is constructed with PEG44-b-PS50 using CHARMM-GUI interface (Polymer Builder feature) 5,6 . The types, charges, and parameters of atom are assigned with the program of CHARMM General Force Field 7-10 . PEG44-b-PS50 is configured with CHARMM36 all-atom force fields 11 . A configuration of 400 polymers is constructed by replicating PEG44-b-PS50 for 20 times along X and Y direction. The configuration is energy minimized with steepest descent method (10 fs time steps, 5000 steps). An NPT equilibration is implemented (10 ns, 2 fs time steps). Temperature (283 K) and pressure (1 bar) are maintained with velocity rescaling method and semi-isotropic pressure coupling by Berensden method (coupling constant: 5 ps), respectively 12,13 . Coulombic and van der Waals interaction are cutoff at 1 nm. Particle mesh Ewald summation (grid size: 0.12 nm) is used to correct long-range interactions [14][15][16] . Hydrogen bond is constrained with LINCS algorithm 17 . Periodic boundary condition is used for all directions. As-equilibrated configuration is fully hydrated with 123825 water molecules by SPC water model 18,19 . An NPT equilibration for the solvated configuration is implemented (100 ns, 2 fs time steps).
For molecular probe, the same set of run input parameters is used as described in section 3.1.
After equilibration, molecule probe in the most stable configuration is loaded onto the solvated

Preparation of polymer vesicles with PEG corona
Polymer vesicles were prepared by solvent switch method. Briefly, 10 mg PEG44-b-PS178 was dissolved in 1 mL mixture of 1,4-dioxane and THF (1:4 v/v). After stirring for 0.5 h, water was injected (1 mL/h, 1 mL). As-obtained polymer vesicles were quenched by adding 10 mL water.
After centrifugation and water washing (3 times), the polymer vesicles were dispersed in water.
The polymer vesicles were evenly divided into 10 portions for the loading experiments.
NanoSight LM10 was used to measure the concentration of polymer vesicles.

The loading of molecular probes onto PEG corona under different pH
Molecular probes of Py-EG4-OH, Py-EG4-Im, Py-EG4-COOH, and Py-EG4-NH2 were dissolved in water. The probe solution was added to the dispersion of polymer vesicles under different pH values. The pH was adjusted by adding H2SO4 or NaOH. In the loading solution, the probe concentration was around 20 uM, and the concentration of polymer vesicles is 2.72×10 -4 μM. After incubating under room temperature for 10 min, the polymer vesicles were removed by centrifugation. The supernatant was used for UV-vis absorbance measurements.
As-obtained polymer vesicles were redispersed in 1 mL water for fluorescence measurements.

Loading and unloading of molecular probes
A molecular probe, for instance, Py-EG4-Im, was loaded onto polymer vesicles according to section 5.2. H2SO4 was added into the loading solution to induce the unloading of molecular probes. For multiple loading and unloading cycles, H2SO4 and NaOH were alternatively added into the loading solution to adjust the pH. The distance between loaded probes is calculated according to the loading amount and size of the polymer vesicles 1 .  To evaluate our loading protocol, the absorption spectra for the aqueous solution of Py-EG4-Im (~20 uM) before and after centrifugation were measured. As shown in Figure S5a Figure S5c). Specifically, PEG22-b-PBD37 (1.25 mg) dissolved in chloroform was dried with nitrogen stream. After vacuum drying for 12 h, water (3 mL) was S15 added. The resulting mixture was incubated for 24 h under 60 o C. As-obtained vesicles were incubated with Py-EG4-Im (33 μM) in a chamber slide. After 10 min, the vesicles were visualized with Leica DMi8 widefield microscope (excitation: 395 nm). The fluorescence of Py is used to visualize the loading and distribution of the molecular probes on vesicles. As shown in Figure S5c, the non-fluorescent polymer vesicles start to emit fluorescence after incubated with Py-EG4-Im. This confirmed the successful loading of Py-EG4-Im onto polymer vesicles with PEG corona. The even fluorescence indicated the homogenous distribution of Py-EG4-Im on the vesicular membrane. To clarify the contribution of the interaction between Py and PS on the loading of molecular probes, we investigate the loading of Py-EG4-Im onto PS and PEG microparticles. Due to the fluorescence of Py, the loading of Py-EG4-Im onto microparticles can be visualized by fluorescence microscope (Figure S6a). Here, PS microparticle is fabricated by interface precipitation as reported previously, which has no surfactants on their surface, allowing us to investigate the interaction between Py-EG4-Im and PS 28 . PEG microparticle is fabricated with a microfluidic setup according to the procedures of our previous work 29 .
After incubated with Py-EG4-Im in the chamber slide for 10 min, PEG and PS microparticles  and imidazole (7.8-6.5 ppm). The peaks are slightly shifted due to the change in solvent (CDCl3 to D2O), as shown by the arrows. The incline from 5.4 ppm is due to the water peak. Polymer vesicles keep intact during the experiment, which is attributed to the glassy PS core. Figure S12. The unloading of molecular probes when pH is decreased from 7.4 to 6.5 and 6.0. 7.4 represents physiological pH, 6.5 represents the pH in early endosome, and 6.0 represents the pH in late endosome. 26% and 47% of the loaded Py-EG4-Im are unloaded when pH is decreased from 7.4 to 6.5 and 6.0, respectively.  Figure S14. Schematic distribution of Py-EG4-Im on the surface of PEG corona during the loading and unloading cycle. The red hexagon represents the loaded -Im. By assuming a 2-fold increase over the distance between the loaded probes during unloading, each probe would result in 3 neighboring probes being released into the solution by electrostatic repulsion. This corresponds to an unloading amount of 75%, which is close to our experimental results.

S25
By assuming the homogenous distribution of probes on the vesicular membrane, the distance (d) between loaded molecular probes can be calculated with equation S1: where N is the loading amount of probes, n is the number of polymer vesicles, and r is the radius of polymer vesicles. In multiple cycling experiments, n is 2.07×10 -7 μmol as measured by Nanosight, and r is 229 nm as measured by DLS. With N obtained from the UV-vis spectrometer, d can be calculated.