Periodically Self‐Pulsating Microcapsule as Programmed Microseparator via ATP‐Regulated Energy Dissipation

Abstract Living systems can experience time‐dependent dynamic self‐assembly for periodic, adaptive behavior via energy dissipation pathway. Creating in vitro mimics is a daunting mission. Here a “living” giant vesicle system that can perform a periodic pulsating motion using adenosine‐5'‐triphosphate (ATP)‐fuelled dissipative self‐assembly is described. This dynamic system is built on transient supramolecular interactions between the polymer and cellular energy currency ATP. The vesicles capturing ATPs will deviate away from equilibrium, leading to an energy ascent that drives a continuous vesicular expansion, until a competitive ATP hydrolysis predominates to break the ATP–polymer interactions and deplete the energy stored in the vesicles, leading to an opposing vesicular contraction. The input of ATP energy can sustain that these vesicles run periodically along this reciprocating expansile–contractile process, resembling a “pulsating” behavior. ATP level can orchestrate the rhythm, amplitude, and lifetime of this biomimetic pulsation. By pre‐programming the ATP stimulation protocol, this kind of adaptive microcapsules can function as high‐performance microseparators to perform size‐selective sieving of different nanoparticles through ATP‐mediated transmembrane traffic. This man‐made system offers a primitive model of time‐dependent dynamic self‐assembly and may offer new ways to build life‐like materials with biomimetic functions.

reactive solution was filtrated and then removed the organic solvent by reduced distillation.
The crude product was re-dissolved in methanol/hydrochloric acid solution (pH = 5.0, 1/3, v/v) and then excluded the insoluble precipitation by filtration. The filtrate was dialyzed in semipermeable tube (MWCO = 3.0 kDa) to completely remove the organic phase and unreactive ATP-receptor unit, then obtained the final copolymer (denoted as PHM) by Vesicles loaded with apolar-sensitive DPH probe. DPH was initially dissolved in THF to obtain a 1.0 mM stock solution and then 2.0 μL aliquot of this stock solution was added to a 1.0 mL cuvette containing copolymer solution in THF. The mixture was dropwise into deionized water to form polymer assemblies and removed the unloaded fluorescent dyes by dialysis three times, finally obtaining a DPH-loaded polymer micellar solution.

Co-encapsulation of NP PEIx into vesicles.
For the encapsulation with dye-modified NP PEIx (x = 5, 15 and 25) together into the vesicle lumen, equal molar amount of NP PEI-5 (0.1 mg), NP PEI-15 (0.3 mg) and NP  in N,N-dimethylformamide solution (0.5 mL) were added into the PHM in aqueous solution (0.40 mg/mL). The unloaded NP PEI-x was removed by dialysis tube (MWCO = 50 kDa). From UV-Vis spectra to monitor the unloaded NP PEI-x we can know that 0.052 mg of NP PEI-5 , 0.137 mg of NP PEI-15 , and 0.191 mg of NP PEI-25 were enclosed in the vesicles for subsequent experiments. Dye-modified silica NPs with different sizes or protein blends were encapsulated by a similar way.
Methods. TEM images were measured on a FEI Tecnai G2-F20 S-TWIN instrument at 80 kV accelerating voltage. The specimens were prepared by drop-casting polymer aggregate solutions (10 μL) onto carbon coated copper-grid and freeze-drying before observation.
Optical and fluorescent photographs were monitored by Nikon-C2 + laser scanning confocal microscope. 1 H, 13 C and 31 P NMR spectra of functional molecules and polymer sample were

Supporting Results and Characterization.
Figure S1. Optical image and particle-size auto-statistics of the PHM vesicles before (left panel) and after (right panel) ATP treatment (3.0 mM). Before the ATP stimulus, the initial vesicular size is averagely 0.72 μm. After addition of ATP, the vesicular size has a 4-fold increase up to averagely 2.80 μm.     a The ligand are ATP and its hydrolytic products Pi and AMP, whose concentration is at 1.0 The receptor is the model ATP-receptor unit (50 μM), which is comprised of βcyclodextrin with a biguanidine tail (as following blue structure). c The association constants      When the vesicles expanded to maximum size (30 min), the individual ATP signals were strongly depressed whereas three groups of down-shift splitting signals at δ = -7.5 (d, 3 J β,γ = 18.4 Hz), -8.6 (d, 3 J α,β = 10.5 Hz), and -19.3 (dd) ppm were enhanced, indicating ATP bound to the polymer micelles. After the micelles shrank autonomously to the minimum size (after 90 min), these phosphorus signals ascribed to ATP/polymer complex completely vanished, but a new signal ascribed to Pi (δ = +1.2 ppm) and AMP (δ = -6.3 ppm) species appeared. It indicates that the ATP was decomposed into AMP and Pi by the enzyme.    Programmed vesicular pulsation process with three distinct stages (Stage I: blue curve, 0.5 mM ATP fuel leading to 2.0-fold size jump; Stage II: green curve, 1.5 mM ATP fuel leading to 3.0-fold size jump; Stage III: red curve, 4.0 mM ATP fuel leading to 4.8-fold size jump).
Each stage contains three pulsating cycles. (c) Separation efficiency with respect to the different sizes of silica nanoparticles in the three stages: Stage I for selectively sieving silica-5 with 99.7% sieving fraction (c), Stage II for selectively sieving silica-8 with 97.8% sieving fraction (d), and Stage III selectively sieving silica-12 with 86.5% sieving fraction (e).