H2S gasotransmitter-responsive polymer vesicles

A type of new polymeric vesicle self-assembly by o-azidomethylbenzoate-containing diblock copolymer can respond to the cell signaling molecule hydrogen sulfide (H2S). The intracellular H2S gasotransmitter can trigger biomimetic polymersome disruption for targeted drug delivery.


Gas Injection.
In all the gas experiments we employed commercial high-pure H 2 S gas (Newrada Gas Ltd. Co., H 2 S ≥ 99.9%, 8 L) as a gas source, and equipped with an electronic gas micro-flowmeter (Smit Instrument LWE-10, critical pressure: P c ≤ 4.0 MPa, flow rate: 0.01 ≤ v ≤ 150 mL min -1 , ±∆v ≤ 1.5%, operating temperature: -30 o C-80 o C) as a gaseous flow controller. When we added H 2 S gas into the polymer systems, we fixed the gas flow rate at 0.1 mL min -1 and modulate the aeration time to control the total molar amount of H 2 S gas. To further measure the concentration of H 2 S in polymer solution, we also designed a set of gas collection instrument to calculate the amount of unreactive or leaked H 2 S gas. The brief sketch of this equipment depicted as follows: H 2 S gas (0.1 MPa) was first passed through a decompressor for reducing pressure and flow rate, and then we tuned the flowmeter controller to obtain a required constant flow rate (0.1 mL/min). According to different conditions, we changed the aeration time and injected gas into our polymer solution for inducing chemical reactions. The total H 2 S concentration (C total ) can be obtained from flow rate, aeration time and the volume of our polymer solution. In fact, not all of gas can react with the polymer and a small amount of gas will leak out, we calculated this part of leaked gas via draining oil method (C leak ). Finally, we could obtain the actual reactive H 2 S concentration with our polymer by the formula of C reactive = C total -C leak . The oil collection was finished by high-precision miniature measuring cylinder (Sujing Lab Apparatus Co., V total =1.0 mL, scale division is 0.01 mL, error volume is ±0.01 mL, the outside and inside diameter of the cylinder is 7.0 and 2.8 mm respectively, and the height is 145 mm). Using this high-precision measuring cylinder can obtain the volume of the excluded oil.

Synthesis of H 2 S-sensitive monomer (AGMA).
A typical reaction was as follows: 2-Azidomethyl benzoic acid (3.542 g, 0.02 mol) was dissolved in 1.25 equiv. thionyl chloride (SOCl 2 , 3.570 g, 0.03 mol) and reacted in neat under reflux for 1 h, followed by evaporation under reduced pressure. The acid chloride was used in situ since all attempts to purify it caused considerable decomposition. The resulted compound (3.51 g, yield ~90%) was reacted with glycerol methacrylate (GMA, 2.645 g, 0.016 mol) in pyridine (50 mL). The mixture was stirred at room temperature for 4 h, and then the crude products were distillated by reduced pressure to remove the excess of pyridine. Afterward the solids were dissolved in CH 2 Cl 2 and poured into water with vigorous stirring for 1 h. The organic phase was collected and the water phase was extracted twice using CH 2 Cl 2 . The combined organic solution was further washed with 1.0 M NaHCO 3 solution, dried over anhydrous MgSO 4 overnight and the solvent was then removed by reduced distillation. The product was dissolved in 2 mL of CH 2 Cl 2 and purified by column chromatography with ethyl acetate/CH 2 Cl 2 (5/1, v/v). Yield: 3.164 g, 62%. 1

Synthesis of poly(ethylene oxide)-based macro-initiator (PEO-Br).
The PEO-Br macro-initiator was prepared by the reaction of PEO with excess of 2bromo-isobutylryl bromide (BiBB) in the presence of triethylamine (TEA) according to the previous literature. 2 First, anhydrous PEO (3.000 g, 1.0 mmol) was dissolved in 60 mL of CH 2 Cl 2 , followed by adding TEA (0.202 g, 2.0 mmol). BiBB (0.570 g, 2.5 mmol, 2.5 equiv.) dissolved in anhydrous CH 2 Cl 2 (20 mL) was finally injected dropwise over 2 h to the solution at 0 o C. Then the solution was stirred for 24 h and treated with activated charcoal. After filtration, the filtrate was collected and the solvent was removed by rotary evaporation. The crude product was dissolved in CH 2 Cl 2 and then poured into NaHCO 3 aqueous solution for 1 h. The extracted organic phase was further washed with 1.0 M HCl (10 mL), 1.0 M NaOH (10 mL) and deionized water (2×10 mL) successively, dried over anhydrous MgSO 4 overnight and the solvent was then removed by reduced distillation. The product was dissolved in 10 mL CH 2 Cl 2 and precipitated into 100 mL of cold diethyl ether. This dissolution/precipitation cycles was repeated twice and obtained the solid. Yield: 2.810 g, 86%. M n, NMR = 3.16 kDa, M n,GPC = 3.12 kDa, M w /M n = 1.04. FT-IR (KBr,

Synthesis of poly(ethylene oxide)-b-poly(o-azidomethylbenzoyl glycerol methacrylate) via atom transfer radical polymerization (PEO-b-PAGMA).
PEO-Br macro-initiator (0.325 g, 0.1 mmol), the monomer of AGMA (1.914 g, 6.0 mmol), CuBr (14 mg, 0.1 mmol), PMDETA (20 μL, 0.1 mmol), and 5.0 mL of anhydrous THF were added into a round bottom flask, followed by three freezevacuum-thaw cycles. The flask was reacted at 75 o C with magnetic stirring. After reaction for 8 h, the flask was immersed to liquid nitrogen in order to stop the radical polymerization. Then the solution was diluted to 30 mL of THF and passed through a neutral alumina column twice to remove the copper catalysts. The filtrate was concentrated to ~8 mL, and then precipitated into 300 mL of cold diethyl ether/methanol (1/1, v/v) mixture for three times. The product was collected and dried in vacuum oven at 25 o C for 24 h, yielding a white solid of 0.980 g (conversion: 35%).

H 2 S-induced cascade reaction of PEO-b-PAGMA copolymer
When we added H 2 S (20 μM) into the PEO-b-PAGMA copolymer solution (9×10 -3 g L -1 , containing ~20 μM AzMB group), the polymer began to remove the AzMB groups to yield PEO-b-PGMA, and the solution generated a new kind of cyclic benzolactam byproduct. First, we used NMR spectroscopy to survey whether PEO-b-PAGMA can be decomposed to PEO-b-PGMA by H 2 S treatment. As shown in Fig.  S3a, we designed and synthesized a water-soluble diblock copolymer PEO 67 -b-PGMA 35 , and it showed typical glycerol group proton peaks [δ = 3.82 (t), 4.26 (m) and 4.63 (m)]. As compared to the decomposed product by H 2 S gas (Fig. S3b), its NMR proton shifts are almost consistent with that of PEO 67 -b-PGMA 35 . Thus, we can deduce that H 2 S can cleave the AzMB groups and convert PEO-b-PAGMA to watersoluble PEO-b-PGMA. Next, according to the H 2 S-induced cascade reaction as Eq.(1), 1 we deduced that the byproduct is possible to be indonlin-1-one. To survey this mechanism, we used standard indonlin-1-one (purchased from Sigma-Aldrich) as a reference and carried out control experiments by means of NMR and UV-Vis spectroscopy.  1H)]. In the other hand, after we applied H 2 S gas stimulus to the PEO-b-PAGMA, the decomposed copolymer was collected by precipitation method. Except for the copolymer, there was still a kind of unknown byproduct existing in the filtrate. As we purified this byproduct, its 1 H NMR spectrum (Fig. S4b) was found to be consistent with that of the indonlin-1-one, which indicates that low concentration of H 2 S stimulus can induce our copolymer to undergo the cascade reactions for removing the AzMB groups and further forming the PEO-b-PGMA copolymer. Furthermore, we employed UV-Vis spectroscopy to monitor the difference between the standard indonlin-1-one and the byproduct from H 2 S-sensitive decomposition reaction. As shown in Fig. S5, it is noted that standard indonlin-1-one displayed a double-peak absorption at λ = 248 and 259 nm (red curve), which is consistent with that of the decomposed unknown byproduct (blue curve). This result further confirmed the above conclusion that H 2 S can induce a cascade reaction and cut off the PEO-b-PAGMA to yield indonlin-1-one, like the reaction process as Eq. (1).

Fig. S6. FT-IR spectra of the PEO-b-PAGMA in different stimulus conditions: (a) no stimulus, (b) 6 min of H 2 S stimulus and (c) 30 min of H 2 S stimulus.
At last, we also employed FT-IR spectroscopy to survey the spectral changes of the polymer system before and after H 2 S gas treatment. In the IR experiments, we fixed the H 2 S gas flow rate at 0.1 mL min -1 and modulated the aeration time from 0 to 30 min. We prepare three copies of PEO-b-PAGMA solution (THF/H 2 O, 5/1, 0.2 g L -1 , 10 mL) for H 2 S stimulation. The first sample was in the absence of H 2 S and removed the solvent by lyophilization for obtaining solid sample to perform IR characterization. The second sample was passed through H 2 S for 6 min, and then stopped the gas flow and removed the solvent to gain solids for IR characterization. The third one was injected H 2 S gas for 30 min, and then removed the solvent to characterize. Before H 2 S stimulus, it is clear that the PEO-b-PAGMA showed typical symmetric stretching vibration of benzene ring and azido group at 1614 cm -1 and 2102 cm -1 , respectively (Fig. S6a). However, after H 2 S has passed through the polymer system for 6 min, the vibration band ascribed to azido group was strongly depressed but the band of aromatic group kept constant, concomitantly, a new shoulder peak at 3465 cm -1 ascribed to the typical benzylamine group appeared (Fig. S6b). It is demonstrated that at the early stage of this reaction, the azido group can be fast reduced by H 2 S to convert into benzylamine group. Further prolonging the H 2 S stimulation time to 30 min, the three vibration bands of benzylamine group, azido group and aromatic group all vanished, whereas a broad absorption band belonged to hydroxyl group (3240-3640 cm -1 ) was strengthened (Fig. S6c), which results from the removal of AzMB group from the polymer main chain. These findings confirmed that H 2 S can fast transform benzylazide into a high-reactive nucleophilic benzylamine intermediate, and the latter is capable of attacking intramolecularly on the adjacent benzoyl to induce cascade self-elimination reaction, leading to a site-specific chemical scission.

Critical aggregate concentration (CAC) of the PEO-b-PAGMA in aqueous solution
A 5 mL solution of 2.0 g L -1 PEO-b-PAGMA in THF was added into 5 mL of deionized water under sonication. Then the solution followed by dialysis against deionized water to obtain an aggregate solution at a concentration of 1.0 g L -1 for further experiments. The critical aggregate concentration (CAC) of PEO-b-PAGMA was tested by pyrene fluorescent probe method. A 10 μL of 5×10 -5 g L -1 pyrene solution in acetone was mixed with the polymer to obtain a series of PEO-b-PAGMA/pyrene aqueous solutions with different concentrations (1×10 -3 , 2×10 -3 , 5×10 -3 , 0.01, 0.02, 0.05, 0.1, 0.2, 0.5 and 0.75 g L -1 ) and the solutions were sonication for 10 min before fluorescent measurements. The results exhibited that the CAC is about 0.02 g L -1 (The CAC was chosen as the concentration when pyrene probe showed an apparent decrease in the I 1 /I 3 ratio with an increasing concentration of the copolymer, indicating that the aggregation occurred, Fig. S7). 3 Fig. S7. The CAC of the copolymer PEO-b-PAGMA by using the fluorescent method and the CAC is determined to be ~0.02 g/L in aqueous media.

The hydrodynamic radius of the PEO-b-PAGMA vesicles in aqueous solution
When the PEO-b-PAGMA copolymer dissolved in aqueous media, in the absence of trigger, they can self-assemble into aggregated nanostructures. Dynamic light scattering (DLS) analysis showed that the hydrodynamic radius (R h ) of these polymer aggregates is about 34.6 nm (Fig. S8).

The copolymer counterpart
From all the experimental results, we found that the PEO-b-PAGMA vesicles could be completely disassembled by H 2 S gas stimulation. Besides the mechanism of H 2 Sinduced polymer cleavage, mechanical explosion effect driven by gas bubbles might be another possible explanation for this vesicular disruption phenomenon. To eliminate this possibility, we synthesized a copolymer counterpart, PEO-bpoly(benzoyl glycerol methacrylate) (PEO-b-PBGMA). PEO-b-PBGMA has a similar chain structure to PEO-b-PAGMA but is lack of the azido group (Fig. S9a). In aqueous solution, they can also self-assemble into analogous vesicular morphology with the average diameter of 74 nm. However, when we exerted H 2 S gas (100 μM) to the PEO-b-PBGMA aggregate solution, these assemblies had no obvious changes either in size or morphology (Fig. S9b─c), which indicates that H 2 S is unable to dissociate PEO-b-PBGMA vesicles. Thereby, this result eliminates the possibility of mechanical explosion effect and further supports that the disassembly mechanism of our polymer vesicles arises from H 2 S-sensitive polymer structural alteration.

H 2 S-Responsive Specificity of PEO-b-PAGMA Copolymer
We have demonstrated that our PEO-b-PAGMA polymersomes possess H 2 S-triggered disassembly ability. Because we expected that these responsive nanocapsules could be applied in biological cells, thus these polymersomes should possess high-specificity to intracellular H 2 S neurotransmitter. However, besides H 2 S signaling molecule there are other sulphur-containing bioactivators in cell such as cysteine (Cys), methionine (Met) and glutathione (GSH). 4 To detect whether they have similar responsiveness, we used UV-Vis spectroscopy to monitor their reactivity to our PEO-b-PAGMA copolymer. As shown in Fig. S10a, we found that H 2 S causes the decomposition of PEO-b-PAGMA and the reactive byproduct, indolin-1-one, led to a remarkable change in UV-Vis spectra. In a similar way, when we added other biological stimulants into our polymer system (stimulant concentration is 45 μM and the treatment time keep in 60 min), respectively, their solution UV-Vis spectra had negligible changes (Fig. S10b,  Cys; Fig. S10c, Met; Fig. S10d, GSH). These confirmed that these sulphur-containing bioactivators have no ability to induce the polymer vesicle disassembly. These results demonstrate that our polymersomes are of high-specific H 2 S-responsiveness.

CSE Enzyme Co-assembled with the Polymer Vesicles
To further extend the responsive scope of our polymer vesicles, we attempted to anchor CSE enzyme onto the vesicular membrane. First, we adopted a thin filmrehydration method to co-assemble the CSE enzyme with our polymer vesicles: 0.2 mg of the PEO-b-PAGMA copolymer was dissolved in 5 mL of THF placed into a 25 mL round-bottom flask, and then the organic solvent was removed to form a dry film by rotary evaporation. After purging with N 2 for 15 min, 20 nM of CSE enzyme solution (PBS buffer, 10 mL) was added into the flask using a syringe. The mixing solution was stirred overnight to ensure the self-assembly process. The CSE-anchored vesicles were separated out as the precipitant by ultracentrifugation at 11000 rpm for 20 min. Then the precipitant was re-dispersed in PBS buffer (10 mL) to form hybrid assemblies. After the vesicle formed, there should be a part of CSE residue dispersed in the solution. To remove these residues, the aggregate solution was centrifuged for 5 min at 11000 rpm at certain intervals (2 h), and 2 mL of the supernatant was withdrawn and replaced by fresh medium. The CSE residue was existed in the supernatant and was assayed via absorbance at typical 469 nm in UV-Vis spectrum. Repeating this centrifugation process ensures that the supernatant had no obvious absorption, which indicates that the CSE residue was removed from the vesicle solution. According to the established calibration curve, there were ~35% CSE enzyme associated with the PEO-b-PAGMA vesicles. Finally, to further confirm that CSE enzyme was anchored onto the vesicle membrane surface, we employed zeta-potentiometer to monitor the surface charge of the polymer aggregate before and after the addition of CSE enzyme. As shown in Fig. S11, the pure PEO-b-PAGMA vesicles exhibited a low positive potential (+4 mV, black square). However, since the CSE protein is a kind of enzyme with negative surface charges, the surface potential of the CSE-polymer hybrid vesicles changed from +4 mV to -22 mV (hollow black square), which indicates that the CSE enzyme can be anchored onto the polymer vesicular membrane.