Vesicles in Multiple Shapes: Fine-Tuning Polymersomes’ Shape and Stability by Setting Membrane Hydrophobicity

Amphiphilic block-copolymers are known to self-assemble into micelles and vesicles. In this paper, we discuss the multiple options between and beyond these boundaries using amphiphilic AB diblock and ABC triblock copolymers. We adjust the final structure reached by the composition of the mixture, by the preparation temperature, and by varying the time-scale of formation. This leads to the formation of vesicles and micelles, but also internal micelles in larger sheets, lamellar vesicles, and closed tubes, thus broadening the amount of self-assembly structures available and deepening our understanding of them.


Methods
Gel permeation chromatography was carried out using a Malvern Viskotek GPC system (Malvern Instruments, UK) using a Novema Max 100Å Column with a Novema Max Guard Column (both PSS Polymer, Germany) with 0.25 vol % TFA in water as an eluent or a Resipore 100Å Column with a Resipore Guard Column (Agilent Technologies, USA) with a chloroform/methanol (3:1) eluent.
NMR spectroscopy was carried out on a Bruker AV600 spectrometer (14.1 T magnetic field strength, operating at 600 MHz for 1H NMR and 125 MHz for 13C NMR spectra).
Water was used from a TKA water purification system (Thermo Scientific, Germany) Transmission electron microscopy was conducted on a JEM 1010 Microscope (JEOL, Germany) with a 80 kV electron beam or a JEM 2010 Microscope (JEOL, Germany) with a 200 kV electron beam, both using 400 mesh carbon-coated TEM grids cleaned for 45 s with a plasma beam at 25 mA electric current (Elektron Technology, UK, Quorum). Samples were prepared from aqueous solutions at 1 mg/mL. The concentration was obtained by diluting original solutions using water from the TKA water system mentioned above and then stained for 10 s using a 1 M PTA (phosphotungstic acid) solution at pH 7.

Polymer Synthesis
Briefly, initiator (PEG-Br or ME-Br for M-H-D) and monomer was dissolved in ethanol, the solution degassed for 30 min, and the CuBr/bipyridin mixture added. The polymerisation was left until complete conversion to give a highly viscous mixture. The second block was attached by adding the corresponding amount of ethanol-dissolved monomer to the solution. For M-H-D, this was repeated for the third block. After final conversion, the mixture was filtered over silica gel with ethanol and dialysed against chloroform/methanol (3:1, 2x), methanol (2x), and water (2x) before being freeze-dried. The solvents were 500 mL each time and the time between solvent exchanges was at least 4 h each time.

GPC traces and NMR spectra
The molecular structure of each copolymer was given in the main paper (Scheme 1). Molecular compositions were calculated taking characteristic peaks of the respective NMR spectra of the polymers and dividing them by the amount of protons they represent. Molecular weights were calculated from these values and dispersities determined via GPC. Due to either heavy interactions with the column material or solubility issues, PMPC-PHPMA-PDPA could not be analysed via GPC.

Calculating the composition
Reference Peaks taken. Peak assignment as reported previously: PEG: 3.63 ppm. Intensity was divided by 4 so that the final number represents the intensity of one proton per repeating unit.
PDPA: 2.64 ppm. Intensity was divided by 2 so that the final number represents the intensity of one proton per repeating unit.
PHPMA: 3.85 ppm. Intensity was divided by 2 so that the final number represents the intensity of one proton per repeating unit. PMPC: 4.24 ppm. Intensity was divided by 2 so that the final number represents the intensity of one proton per repeating unit.
For PEG-based polymers: PEG2000 has 45 repeating units, so the ratio of intensities can be converted to the final number of units present in the polymer. For PMPC, it is assumed that all initiator reacts and as the reaction is run until completion it is assumed that all initiator molecules reacted with 25 units of MPC on average. This number of 25 units is then the basis for all other repeating units in the block-copolymer.

Typical DLS Traces
In addition to the TEM analysis discussed in the main manuscript, we also checked the DLS traces for the structures, which were relevant. All DLS traces are shown as intensity plots, and the xaxis is the hydrodynamic diameter in nm.
DLS for the micelles observed in Figure  Since larger particles scatter the light more than smaller ones, the DLS appears to show the presence of larger objects as well. However, due to the effect just described, the peak corresponding to agglomerates at 400 nm can be neglected.
Typical DLS trace of the nanobjects observed in Figure 2(b1). Due to the irregular shape of these nanoparticles (no perfect sphere), the DLS shows two peaks and can be misleading.
Disassembled particles for Figure  Comparing the image with the DLS trace shows that the DLS trace can be misleading as no defined objects are present.
Again, given their irregular shape (no perfect sphere, which would be necessary for DLS), the DLS trace recorded is in itself not very conclusive and may be seen as misleading.
Closed "half-moon" structures of Figure  As the structures are disassembling, the distribution becomes even broader and is due to the fact that the present patches of polymer are not conclusive.

Formation of the Self-Assembly Structures
Ryan et al. published a very nice study on how vesicles are formed using thin film rehydration. This figure represents nicely how the self assembly process works, which is why we included it for reasons of understandability.