Cellular and Molecular Bioengineering
Controlled Assembly of Synthetic Cells using Megamolecules
Atiriya Iyer (she/her/hers)
Undergraduate Student
Northwestern University
Plano, Texas, United States
Timothy Vu
Graduate Student
Northwestern University, United States
Neha Kamat
Professor
Northwestern University, United States
Developing artificial tissues has been a long-standing goal of bioengineers. Other researchers have engineered cells with synthetic adhesion molecules (SAMs) to create cellular structures. SAMs are nonnative membrane proteins that mediate cell-cell interactions. Recently, Chao et al. utilized SAMs to induce the adhesion of e. coli cells with higher binding specificity than naturally occurring adhesion molecules [1]. One limitation of these systems is that they rely on living cells, which divide and are difficult to control. A solution would be to design fully synthetic tissues of artificial cells. The ability to program synthetic cell communication would further study in tissue repair and therapeutic cell design [2]. However, a method to assemble synthetic cells in predictable ways is still needed.
Many cellular functions, including membrane structure and communication, can be mimicked and studied using lipid vesicles. Giant unilamellar vesicles (GUVs) are micron-scale vesicles with a lipid bilayer resembling the natural cell membrane. GUVs are easily visualized by microscopy, making them a useful model for studying the interactions of membranes with proteins and adhesion [3]. These vesicles can be embedded with proteins to develop artificial cell membranes and loaded with molecules for targeted delivery [4]. In the Kamat and Mrksich labs, we developed specific chemistries for enzyme-mediated molecular assembly known as megamolecules. However, the ability of these megamolecules to serve as SAMs is yet to be explored.
To bridge this gap, we explored the use of the megamolecules: Snaptag (Snap), Cutinase (Cut), and Crabtag (Crab) as SAMs enabling programmed GUV adhesion [5].
Snap, Cut, and Crab proteins selectively bind benzyl guanine (BG), p-nitrophenyl phosphonate (pNPP), and retinoic acid (RA), respectively. BG, pNPP, and RA were purchased from Avanti Polar Lipids conjugated to 1,2-Distearoyl-sn-glycero-3-phosphoethanolamine-Poly(ethylene glycol) (DSPE-PEG-2000) allowing for membrane integration.
5 mM lipid stocks were made with 98.9% 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 0.3% binding lipid, 0.7% 18:0 PEG-2000, and 0.1% of lipid dye in chloroform (Table 1). 10 µL of the lipid stock was dried on a glass slide coated with indium-tin-oxide. The lipid film was rehydrated in 300 mOsm sucrose. The Vesicle Prep Pro was used to make GUVs by electroformation.
To form programmed GUV assemblies, 10 µL of each GUV population was combined with one of the following fusion proteins: Crab-Cut, Crab-Snap, or Snap-Cut. These proteins can bind two of the three binding lipids allowing controlled linkage of two GUV populations while excluding the third. Protein concentrations of 0.5, 1, 2, 5, 10, 20, and 60 times the binding lipid concentration (75 nM) were tested. The GUVs were incubated with the linker protein in a final volume of 100 µL PBS for 2 hours at 25 ºC. 1-5 µL of the reaction was imaged by confocal microscopy in 100 µL of 300 mOsm glucose on a 96-well clear bottomed plate that was blocked with 1% Bovine Serum Albumin for 20 minutes.
Visual observation by microscopy demonstrated that the Crab-Cut, Crab-Snap, or Snap-Cut specifically linked RA and pNPP, RA and BG, or BG and pNPP GUVs, respectively (Fig. 1a). Pearson R coefficients supported these observations.
At a Crab-Cut concentration of 20x, the Pearson’s coefficient for RA-pNPP colocalization was 0.251 ± 0.154, while the RA-BG and BG-pNPP values were both near zero. This suggests that the linker protein bound RA and pNPP GUVs while excluding BG as the RA-pNPP colocalization coefficient indicated positive correlation of the RA and pNPP GUVs, while the RA-BG and BG-pNPP values indicated no correlation (Fig. 1b). Similarly, at a Crab-Snap concentration of 20x, the Pearson’s coefficient for RA-BG colocalization was 0.604 ± 0.066 and greater than 0, while the RA-pNPP and BG-pNPP R values were approximately zero (Fig. 1c). This demonstrates that the linker protein successfully bound RA and BG GUVs while excluding pNPP. Finally, at a Snap-Cut concentration of 20x, the Pearson’s coefficient for BG-pNPP was 0.370 ± 0.151, while the values for RA-pNPP and RA-BG were both near zero demonstrating that the Snap-Cut linker assembled BG and pNPP GUVs while RA GUVs were unbound (Fig. 1d).
The binding affinity of each protein/linker pair was unique, thus the minimal protein concentration for successful assembly differed. At protein concentrations lower than 2x, the Snap-Cut linker was unable to link BG and pNPP GUVs, demonstrated by near zero Pearson R values. At concentrations lower than 5x, the Crab-Snap linker was unable to link RA and BG GUVs. Finally, at concentrations lower than 10x, the Crab-Cut linker was unable to link RA and pNPP GUVs. This suggests that the Crab-Cut linker had the lowest binding as binding begun at much higher concentrations than Snap-Cut and Crab-Snap Linkers Additionally, when protein concentration was 60x, binding resulted in tensions sufficient to rupture the vesicles. These results demonstrated that megamolecule-mediated assembly can be used for programmable, synthetic cell binding at intermediate protein concentrations. This method for GUV assembly can be used to further study synthetic cell communication and progress the development of fully synthetic tissues with programmed assembly.
[1] G. Chao et al., “helixCAM: A platform for programmable cellular assembly in bacteria and human cells,” Cell, vol. 185, no. 19, pp. 3551-3567.e39, Sep. 2022, doi: 10.1016/j.cell.2022.08.012.
[2]A. J. Stevens et al., “Programming multicellular assembly with synthetic cell adhesion molecules,” Nature, pp. 1–9, Dec. 2022, doi: 10.1038/s41586-022-05622-z.
[3] S. Aden, T. Snoj, and G. Anderluh, “The use of giant unilamellar vesicles to study functional properties of pore-forming toxins,” Methods Enzymol, vol. 649, pp. 219–251, 2021, doi: 10.1016/bs.mie.2021.01.016.
[4] K. Göpfrich et al., “One-Pot Assembly of Complex Giant Unilamellar Vesicle-Based Synthetic Cells,” ACS Synth Biol, vol. 8, no. 5, pp. 937–947, May 2019, doi: 10.1021/acssynbio.9b00034.
[5] B. R. Kimmel and M. Mrksich, “Development of an Enzyme‐Inhibitor Reaction Using Cellular Retinoic Acid Binding Protein II for One‐Pot Megamolecule Assembly,” Chemistry A European J, vol. 27, no. 71, pp. 17843–17848, Dec. 2021, doi: 10.1002/chem.202103059.