Direct encapsulation of biomolecules in semi-permeable microcapsules produced with double-emulsions

Abstract

Compartmentalization can serve different purposes such as the protection of biological active substances from the environment, or the creation of a unique combination of biomolecules for diagnostic, therapeutic, or other bioengineering applications. We present a method for direct encapsulation of molecules in biocompatible and semi-permeable microcapsules made from low-molecular weight poly(ethylene glycol) diacrylate (PEG-DA 258). Microcapsules are produced using a non-planar PDMS microfluidic chip allowing for one-step production of water-in-PEG-DA 258-in-water double-emulsions, which are polymerized with UV light into a poly-PEG-DA 258 shell. Semi-permeable microcapsules are obtained by adding an inert solvent to the PEG-DA 258. Due to the favorable hydrophilicity of poly-PEG-DA 258, proteins do not adsorb to the capsule shell, and we demonstrate the direct encapsulation of enzymes, which can also be dried in the capsules to preserve activity. Finally, we leverage capsule permeability for the implementation of a two-layer communication cascade using compartmentalized DNA strand displacement reactions. This work presents the direct encapsulation of active biomolecules in semi-permeable microcapsules, and we expect our platform to facilitate the development of artificial cells and generating encapsulated diagnostics or therapeutics.

Figure 1

Production of semipermeable microcapsules in a PDMS device with 3D geometry. (A) Schematic representation of the PDMS device with 3D geometry. W/O/W double-emulsions are generated with a PEG-DA 258 middle phase encapsulating an inner aqueous core. (B) Micrographs of the PDMS device operation. (C) Brightfield image of microcapsules obtained after UV polymerization of the collected double-emulsion. (D) Size distribution of a representative batch of polymerized microcapsules. (E) Schematic representation of PIPS upon UV illumination. 15% Butyl-acetate (porogen) was mixed with PEG-DA 258 to form semi-permeable microcapsules. (F) In the collected double-emulsion, both high molecular weight 500 kDa FITC-dextran and lower molecular weight 40 kDa RITC-dextran are retained in the inner aqueous phase. (G) After UV polymerization and PIPS, pores are formed in the capsule shell and capsules become semi-permeable.

Figure 2

PEG-DA 258 microcapsules are semi-permeable and the pore size can be adjusted by changing the porogen. Schematic representation of PEG-DA 258 microcapsules produced using (A) 15% butyl-acetate or (C) 15% 1-decanol as porogen. 15% butyl-acetate microcapsules selectively allowed the permeation of 10 kDa RITC-dextran while excluding 32.7 kDa EGFP. The larger pore size of microcapsules produced with 15% 1-decanol as porogen allowed the diffusion of both fluorescent molecules. Microcapsules produced using (B) 15% butyl-acetate or (D) 15% 1-decanol porogen were immersed in a solution containing 10 kDa RITC-dextran and 32.7 kDa EGFP. The evolution of the fluorescent signal in the Cy3 and FITC channels was observed after 5 min, 1 h and 24 h. While microcapsules produced using 15% 1-decanol were permeable to both fluorescent molecules, we clearly observed the selective permeability of microcapsules produced using 15% butyl-acetate as a porogen.

Figure 3

Direct encapsulation of proteins inside semi-permeable PEG-DA 258 capsules. EGFP was added to the aqueous inner phase for direct encapsulation. (A–C) Microscope images of double-emulsions showing a fluorescent signal in the FITC channel. (D–F) After polymerization, the fluorescent signal was still present in the interior of most capsules. Variability in the pore size or a size cutoff close to the 32.7 kDa EGFP resulted in some protein leakage. (G) Fluorescent intensity profile of the capsule indicated in panel E. The fluorescent profile suggests a homogeneous distribution of EGFP in the interior of the capsule without protein adsorption to the shell material. (H–J) Microscope images of polymerized capsules containing FITC-streptavidin after direct encapsulation. Fluorescent signal is present in all capsules, suggesting that the pore size is too small for FITC-streptavidin leakage. (K) Fluorescent profile across two capsules from panel I. The profile suggests a homogeneous distribution of FITC-streptavidin in the interior of the capsule without protein adsorption to the shell material.

Figure 4

Direct encapsulation of functional enzymes in semi-permeable PEG-DA 258 microcapsules. A luciferase-GFP fusion protein was directly encapsulated in semi-permeable capsules. The fluorescent fusion protein allows for the visualization of the enzyme (A) in the double-emulsion, and (B), in the polymerized capsules. (C) The encapsulated econoLuciferase shows a strong signal in a bioluminescent assay. Direct encapsulation of $\beta$-galactosidase. (D,E) Enzyme-containing capsules were dispersed in trehalose and air-dried at 37 ${}^{\circ }$C. (F) After rehydration with a solution containing CPRG, the substrate was hydrolyzed to chlorophenol red. (G) The solution was imaged with a color camera mounted on an inverted microscope with $\times$ 4 magnification. Capsules displayed a purple color in their interior, indicative of $\beta$-galactosidase enzymatic activity.

Figure 5

Immobilization of DNA strand displacement reaction network in semi-permeable microcapsules and implementation of a two-layer signalling cascade. (A) Schematic representation of the two-layer signalling cascade as developed by Joesaar et al. (B) Implementation of the two-layer signalling cascade in poly-PEG-DA 258 capsules. The two capsule populations were mixed together and imaged on a cell-counting slide immediately after addition of 50 nM input strand (A). An increase in Cy5 and Cy3 fluorescent signals was observed corresponding to the activation of the first and second populations, respectively. (C) Median intensity of detected particles. An increase in the Cy5 signal was observed corresponding to the activated first population of capsules. After release and diffusion of the signal strand ($Q_1$) to the second capsule population, an increase in Cy3 signal was observed corresponding to their subsequent activation. The larger symbols correspond to the median of all detected particles in a given fluorescent channel.