Precision Anisotropic Brush Polymers by Sequence Controlled Chemistry

The programming of nanomaterials at molecular length-scales to control architecture and function represents a pinnacle in soft materials synthesis. Although elusive in synthetic materials, Nature has evolutionarily refined macromolecular synthesis with perfect atomic resolution across three-dimensional space that serves specific functions. We show that biomolecules, specifically proteins, provide an intrinsic macromolecular backbone for the construction of anisotropic brush polymers with monodisperse lengths via grafting-from strategy. Using human serum albumin as a model, its sequence was exploited to chemically transform a single cysteine, such that the expression of said functionality is asymmetrically placed along the backbone of the eventual brush polymer. This positional monofunctionalization strategy was connected with biotin–streptavidin interactions to demonstrate the capabilities for site-specific self-assembly to create higher ordered architectures. Supported by systematic experimental and computational studies, we envisioned that this macromolecular platform provides unique avenues and perspectives in macromolecular design for both nanoscience and biomedicine.


Synthesis of brush-1 to brush-4
Preparation of the stock solution of CuBr2/TPMA catalyst CuBr2 (4.47 mg, 0.02 mmol) and TPMA (46.5 mg, 0.16 mmol) was dissolved in 1 mL mixture solution of water:DMF (1:1 v/v) and stored at 4 °C prior to use. Therefore, the concentration of the effective Cu 2+ in the stock solution was 20 nmol μL -1 .

Preparation of the ascorbic acid solution
To a 10 mL Schlenk flask, L-ascorbic acid (4.4 mg, 25 μmol) was firstly added under argon flow. Degassed deionized water (5 mL) was then added to dissolve L-ascorbic acid. The solution was then stirred under argon flow for 40 min. Therefore, the concentration of Lascorbic acid was 5 nmol μL -1 .

Atom-transfer radical polymerization (ATRP)
Protein-templated brush polymers (brush-1 to brush-4) were synthesized via activator regenerated by electron transfer (ARGET) ATRP under different conditions. As a typical protocol, PEG-dcHSA-Br61 (0.449 mg, 3 nmol) was firstly dissolved in 0.449 mL of deionized water in a 5 mL flask. MOEGMA (52.3 μL, 183 μmol), the CuBr2/TPMA catalyst solution (11 μL, 220 nmol of Cu 2+ ) and 0.5 mL of deionized water were then added. Subsequently, the mixture solution was degassed through three freeze-pump-thaw cycles and the flask was refilled with argon. The L-ascorbic acid solution was then added with a syringe pump at a speed of 0.6 μL min -1 at room temperature for two hours. The brush polymer solution was purified with deionized water for five times by ultracentrifugation using a Vivaspin 6 concentrator (MWCO 30 kDa).

Preparation of the protein-templated brush polymer site-specifically functionalized
with a dye (brush-5)

Synthesis of fluorescent HSA (f-HSA) via site-specific labeling
The site-specific labelling of HSA with a dye was conducted under dark. HSA (10 mg, 0.15 μmol) was dissolved in 10 mL of degassed PBS (pH 6.5

Synthesis of PEGylated f-cHSA (f-PEG-cHSA)
f-cHSA (8.1 mg, 0.11 μmol) was firstly dissoleved in 8 mL of degassed phosphate buffer (50 mM, pH 8.0). NHS-PEG (6.9 mg, 3.45 μmol) was dissolved in 0.1 mL of DMSO and then added into the cHSA solution. After stirring overnight at room temperature, the reaction solution was purified with deionized water for eight times by ultracentrifugation using a Vivaspin 6 concentrator (MWCO 30 kDa). The resulting solution was lyophilized to obtain the product as a light blue fluffy solid (11.6 mg, yield: 96%). The MALDI-ToF MS indicates a molecular weight of 107.9 kDa which means on average 18 PEG chains were conjugated into each f-cHSA backbone.

Synthesis of denatured f-PEG-cHSA (f-PEG-dcHSA)
The reaction and purification were conducted under dark. To a 25 mL flask, 10 mL of urea-PB atmosphere. The obtained reaction solution was purified with urea-PB for three times and with deionized water for five times by ultracentrifugation using a Vivaspin 6 concentrator (MWCO 30 kDa). The product was obtained as a light blue fluffy solid after lyophilizing (11.6 mg, yield: 99%, MALDI-ToF MS: 110.9 kDa).

Synthesis of brush-5
The brush polymer site-specifically functionalized with AF647 (brush-

Preparation of protein-templated brush polymers site-specifically functionalized
with biotin (biotin-brush-6 and biotin-brush-7) and their assembly S9
After stirring for two hours at room temperature, acetate buffer (0.1 mL, 4 M, pH 4.75) was added to terminate the reaction. The obtained reaction solution was purified twice with acetate buffer (100 mM, pH 4.75) and thrice with deionized water by ultracentrifugation using a Vivaspin 6 concentrator (MWCO 30 kDa). The resulting solution was lyophilized to afford the product as a white fluffy solid (6.2 mg, yield: 97%, MALDI-ToF MS: 76.6 kDa).
NHS-BiB (160 mg, 0.61 mmol) dissolved in 1.6 mL of DMSO was then added dropwise into the biotin-PEG-dcHSA solution. The reaction solution was stirred overnight at room temperature and then washed with deionized water for eight times by ultracentrifugation using a Vivaspin 20 concentrator (MWCO 30 kDa). The purified solution was then lyophilized to afford the product as a white fluffy solid (6.3 mg, yield: 64%, MALDI-ToF MS: 131.7 kDa).

Synthesis of biotin-brush-6 and biotin-brush-7
Protein-templated brush polymers site-specifically functionalized with biotin (biotin-brush-6 and biotin-brush-7) were synthesized by ARGET ATRP. In a typical example for the synthesis solution was then degassed through three freeze-pump-thaw cycles and the flask was refilled with argon. The L-ascorbic acid solution was then added with a syringe pump at a speed of 0.6 S11 μL min -1 at room temperature for one hour. The biotin-functionalized brush polymer was purified with deionized water for five times by ultracentrifugation using a Vivaspin 6 concentrator (MWCO 30 kDa).

Assembly with biotinylated somatostatin (biotin-SST) and streptavidin (SA)
SA or AF568-SA (32 µg, 0.6 nmol) was firstly dissolved in Milli-Q water to 1 mg mL -1 and then mixed with biotin-brush-8 (2 mL, 0.3 µM) in a 5 mL tube. After stirring at room temperature for 8 h, biotin-SST (2.964 µL, 1 mg mL -1 ) was added into the tube, and the mixture was stirred further for 10 h. The final solution was purified with PBS for five times by ultracentrifugation using a Vivaspin 6 concentrator (MWCO 30 kDa) and the final construct concentration was tuned to 2 µM.

Matrix-assisted laser desorption/ionization (MALDI) mass spectrometry
MALDI time-of-flight (MALDI-ToF) mass spectrometry was performed on Bruker rapifleX spectrometer. Saturated solution of sinapinic acid dissolved in a 50:50 water/acetonitrile with 0.2% TFA (trifloroacetic acid) was used as the matrix solution.

Gel permeation chromatography (GPC)
The molecular weight and molecular weight distribution were determined by GPC. Deionized water containing 0.1 M NaNO3 was used as eluent at a flow rate of 1 mL min -1 . Shodex RI 101 detector was used. Linear PEG standards were used for calibration.

Dynamic light scattering (DLS)
Dilute where α is the amplitude, τ the relaxation time and β ≤ 1 the stretching parameter. ILT was employed in the case of more than one relaxation process in C(q,t), whereas Equation S1 can amount for the presence of one process deviating from the ideal single exponential shape. In dilute solutions, the relaxation rate Γ(q) = 1/τ(q) is usually diffusive defining the diffusion For species with small size R i.e., both the scattering intensity I(q) and S14 D = D0 are q-independent with I ~ cM and 0 = B 6 0 h where c, M, Rh, η0, kB, and T are the probed species concentration, its molecular weight and hydrodynamic ratio, the solvent viscosity, the Boltzmann constant and the absolute temperature, respectively. For qR ~ 1, both I(q) and D(q) depend on q defining the probing length (2π/q). The former, known as the form factor, yields (at low qRg) the radius of gyration Rg, whereas the effective D is given by, with A is a parameter characterizing the shape of the diffusing species.

Stability test against trypsin
Trypsin was dissolved in deionized water to prepare a 3 mg mL -1 solution and then passed through a syringe filter with the pore size of 220 nm. The solution (1 mL) was then mixed with 2 mL of brush polymer solution (0.3 μM) in a clean hood at room temperature. Therefore, the final concentrations of trypsin and the brush polymer were 1 mg mL -1 and 0.2 μM, respectively. S15 DLS was then used to track the size change of the mixture solution immediately and after different time intervals until six hours.

Cell viability test
A549 mammalian lung adenocarcinoma cells were cultured in standard T-75 flasks using high glucose DMEM fortified with 10% fetal bovine serum, 1% penicillin/streptomycin and 1% MEM non-essential amino acids. The cells were split at near confluency and incubated at 37 °C, 5% CO2 prior to each experiment.

Transmission electron microscopy (TEM)
TEM samples were prepared by adding 4 μL of the brush polymer solution onto a carbon-coated copper grid. After drying in air for 10 min, the remained solution was removed by a filter paper.
The samples were then stained with 2% uranyl acetate solution and dried in air. The measurement was conducted on a JEOL JEM-1400 TEM operating at an accelerating voltage of 120 kV. ImageJ software was used for the analysis of length and width of brush polymer samples.

Fluorescent spectroscopy
Fluorescent emission spectra were collected using a TECAN (Spark 20M) microplate reader at room temperature. The excitation wavelength was set as 594 nm and the emission wavelength S16 was monitored from 600 to 800 nm. Excitation and emission bandwidths were both maintained at 20 nm and the emission wavelength step size was 1 nm.

Fluorescent correlation spectroscopy (FCS)
Fluorescence correlation spectroscopy (  where kB is the Boltzmann constant, T is the temperature, and  is the viscosity of the solvent. Furthermore, FCS yielded also the fluorescent brightness (FB) of the studied species defined as the ratio between the detected average fluorescent intensity and the mean number of fluorescent species in the observation volume, FB=〈I( )〉/N. As the value of r0 depends strongly on the specific characteristics of the optical setup a calibration was done using the reference value 5 of the AF647 diffusion coefficient 3.310 -10 m 2 /s at 25°C.

Confocal laser scanning microscopy (CLSM)
Cell culture was performed with human lung adenocarcinoma cell line, A549, pre cultured in high glucose DMEM, supplemented with 10% Fetal Bovine Serum, 1% Penicillin/Streptomycin, 1% MEM. Cell passages used within the experiment are between 7-

12.
Cells were seeded at a density of 15,000 cells/well in an Ibidi ® µ-slide 8-well confocal microscopy chamber and left to adhere for 24 h. Stock solutions, prepared at 2 µM, were diluted with DMEM to a final concentration of 500 nM. From the cells, DMEM was aspirated and the samples were introduced into each individual well. The cells were then incubated for 24 h.
Subsequently, the cell nuclei were stained with Hoechst for 20 min before imaging live on a Leica SP5 confocal microscope system.
For the time lapse studies at 2 h, 6 h and 24 h, the cells were seeded identically as above with the samples introduced into the cells in such a way that all three time intervals end at the same time for imaging. Each well was likewise washed and stained with Hoechst for 20 min before imaging live on a Leica SP5 confocal microscope system.

SDS-PAGE and Western blot
Electrophoresis was carried out using 10% Bis-Tris polyacrylamide resolving gels with 6% Bis-    Figure S3. Size distribution of brush-1 to brush-4 after storage at 4 °C for two months measured by DLS.

Stability of brush-1 to brush-4
As shown in Figure S3, the sizes of brush polymers (brush-1 to brush-4) did not change during storage and they remained stable even in water solution for more than two months at 4 °C suggesting that they could be stored at low and elevated temperatures without the formation of stable aggregates and precipitates. S27 Figure S4. Digital images showing the thermal responsiveness of brush-3.

Thermo-responsiveness of the brush polymer
Brush-3 reveals high solubility in water at 25 °C. Upon temperature increase to 65 °C, the solution turned turbid immediately ( Figure S4). This lower critical solution temperature (LCST) is consistent with other PMOEGMA polymers based on the monomer with a molecular weight of 300 g mol -1 . 6 Importantly, this behavior is fully reversible and the solution became transparent when the temperature was again decreased to 60 °C underlining that individual brush polymers were recovered and no permanent aggregation occurred. Such reversible responsiveness of the brush polymers also provides interesting future opportunities, e.g. to control their self-assembly. S28 Figure S5. Additional TEM images for the statistical analysis of brush-1.

Simulation details and further analysis
To study the structural behavior of brush polymers with different side chain lengths and grafting densities (as shown in Table S2)  We use simplified replica of the experimental brush polymer system. Here the experimental backbone sequence of hydrophobic and hydrophilic residues (see Figure S14) are modeled with the standard Lennard-Jones (LJ) interactions, where interaction between two hydrophobic residues is chosen as attractive with interaction strength of 2kBT and a cutoff of 2.5σ. The interaction between hydrophilic units is modeled as the repulsive LJ with a cutoff of 2 1/6 σ and interaction strength of kBT. This ensures that the native structure of bare backbone is well reproduced in our generic simulations, as known from experimental backbone structure. The backbone sequence used as an input in our simulations is shown in Figure S14. To model the S36 hydrophilic side chains, we again use repulsive LJ interaction. Note that in our simplified model, a one-to-one monomer mapping is done for the backbone chain. Figure S14. The hydrophobic-hydrophilic pattern of amino acid sequence of HSA for molecular simulation. Hydrophilic amino acids are represented as number "1" and hydrophobic acids are number "0" in the simulation.
We consider a chain of length Nl = 585, the same as number of residues in the backbone of the experimental system. The side chain lengths and the grafting densities are again taken to be the same as the experimental system. The equations of motion are integrated using a velocity Verlet algorithm with a time step δt = 0.01τ. The simulations were usually equilibrated for 10 7 MD time steps. The measurements are typically observed over another 5 × 10 9 MD steps. These values are at least one order of magnitude larger than the typical S37 chain end-to-end relaxation time. During this time, observables such as the gyration radius Rg and static structure factor S(q) are calculated. The temperature is set using a Langevin thermostat with damping constant γ = 1.0τ −1 and the temperature is set to 1 ε/kB. Figure S15. Polymer gyration radius Rg as a function of side chain length Ns for a given grafting density where a backbone is grafted with 61 side chains (see Table S2). Results are shown for bare backbone and also for the full brush polymer chain.
In Figure S15 we show simulation results of Rg values as a function of side chain length Ns for a given grafting density, where a backbone is grafted with 61 side chains. The data is shown for only bare backbone size and also for the full chain. As expected, a chain becomes more swollen with increasing Ns.
S38 Figure S16. A comparative plot of gyration radius Rg obtained from experiments and simulations for all four brush polymer samples (brush-1 to brush-4). Inset shows the same Rg with changing side chain length Ns for a given grafting density where the backbone is grafted with 61 side chains.
In the main panel of Figure S16 we present a comparative experimental and simulation data of Rg for four different bottle brush systems. It can be appreciated that rather simplified simulation model reproduces the correct trend observed in experiments. Not only that we have the same trend, simulations also give same scaling of total Rg with the side chain length Ns (see the inset of Figure S16). Interestingly, a simple shifting of the simulation data (green triangle) by a factor of ~1.33 gives a rather convincing master curve, i.e. Rg * ~ 1.33 Rg. While we do not have a direct evidence of this shifting parameter, we use it as a match between chemical specific experimental and chemically independent generic simulation data. S39

Schematic illustration for the synthesis of brush-5
Figure S17. Schematic illustration for the synthesis of the brush polymer site-specifically functionalized with AF647 (brush-5).      to detect biotinylated molecules. The cyclic peptide SST consist of only 14 amino acids, which were not enough for a sufficient staining with coomassie dye. Additionally, the small size of the SST prevented retention on the nitrocellulose membrane with a pore size of 0.2 µm.