**3. Conclusions**

In this paper, we presented a simple and new microfluidic approach to investigate the electrophysiological properties of the recombinantly produced transmembrane protein Arch-3 inserted in a free-standing DOPC/DPhPC bilayer. By applying a voltage across such an Arch-3-containing DOPC/DPhPC bilayer, the light-induced opening of individual Arch-3 ion channels could be observed. The corresponding pore radius of the Arch-3 ion channel was determined to be (0.31 ± 0.02) nm, which is in excellent agreemen<sup>t</sup> with values found for similar protein pores.

The in vitro system described here presents the advantage of quick testing and prototyping of modified Arch-3, since only the DNA needs to be adapted. Moreover, even non-canonical amino acids can be incorporated [28,44,45]. Because G protein-coupled receptors involved in many signaling cascades exhibit a similar structure, we expect our work to be helpful for in vitro studies focusing on this kind of protein. This may pave the way for the creation of artificial signaling cascades.

#### **4. Materials and Methods**

## *4.1. Gene Expression*

For gene expression, a commercially available cell-free expression system (*E. coli* T7 S30 Extract System for Circular DNA; Promega, Madison, Wisconsin, USA) was used. The plasmid in our experiments was VV020: WT Arch-3-EGFP in pET28b, a gift from Adam Cohen (Addgene plasmid # 58488; http://n2t.net/addgene:58488 (last accessed on 4 September 2021); RRID: Addgene\_58488; Addgene, Watertown, Massachusetts, USA) [16]. The gene expression was performed in the presence of SUV (small unilamellar vesicles). For this, first a solution containing vesicles was prepared as follows: 400 μL of S30 Premix without Amino Acids, 50 μL of Amino Acid Mixture Minus Cysteine at 1 mM, 50 μL of Amino Acid Mixture Minus Leucine at 1 mM and 500 μL of ultrapure water were mixed. To this solution, 1 mg of DPhPC was added and sonicated three times, applying the continuous cycle of 4 s pulse, 4 s break for 4 min, and 2 min pause between each cycle. After that, the solution was put into the fridge to rest for 24 h.

Gene expression reactions were performed as follows: The components of the cell-free expression system were combined to obtain a "mastermix" containing 40 μL of S30 Premix without Amino Acids, 5 μL of Amino Acid Mixture Minus Cysteine at 1 mM, 5 μL of Amino Acid Mixture Minus Leucine at 1 mM and 30 μL of T7 S30 Extract System for Circular DNA. To this mastermix, 10 nM of plasmid DNA was added and filled up with vesicles solution to achieve a total reaction volume of 100 μL. The expression reaction was then incubated at 37 ◦C for 48 h. The synthesized Arch-3-EGFP was incorporated into the vesicles. Each reaction solution was directly injected into the microfluidic device after expression and could no longer be used after 24 h.

#### *4.2. Lipid Preparation, and Device and Bilayer Fabrication*

DOPC (1,2-dioleoyl-sn-glycero-3-phosphocholine), and DPhPC (1,2-diphytanoyl-snglycero-3-phosphocholine) were used for bilayer preparation. The lipids were from Avanti Polar Lipids (Avanti Polar Lipids, Birmingham, Alabama, USA). To prepare the oil–lipid solution, 5 mg of lipids (1:1 DOPC/DPhPC) were dissolved in 1 mL of pure squalene oil (Sigma-Aldrich, St. Louis, Missouri, USA) at 45 ◦C while undergoing continuous stirring for 3 h.

The microfluidic chip was produced by standard soft lithographic protocols and consisted of Sylgard 184 bonded to a glass slide, see e.g., ref. [24] for fabrication details. The chip was designed with two side-to-side rectangular channels with a width of 500 μm and height of 100 μm, forming an X geometry (see Figure 4).

**Figure 4.** Design and structure of the microfluidic setup including the hydrostatic pressure system.

The membrane was formed across an orifice with a width of about 150 μm that connected the two parallel channels, as sketched in Figure 4. A hydrostatic pressure system was used to control the flow of the aqueous solution in the microfluidic chip [24]. The two inlets were connected to two syringes, which were fixed on a motorized stage, and the two outlets were left open. By adjusting the height of the motorized stage, positive or negative pressures could be applied to the channels causing the aqueous solution to move forward or backward.

For bilayer formation, the whole chip was first filled with the squalene oil containing dissolved lipids. Subsequently, the cell-free expression reaction solution containing synthesized Arch-3-EGFP proteins that were fused to vesicles was injected gently into both microfluidic channels, displacing the oil but leaving behind an oil inclusion at the orifice connecting the microfluidic channels. During this process the two oil–water interfaces were being decorated with a monolayer of lipids and Arch-3-EGFP. Due to the drainage of oil into the PDMS, the two lipid monolayers came into contact with each other, leading to the formation of a bilayer. While bilayer formation, the Arch-3-EGFP at the interface of the two monolayers fuses into the bilayer (sketched in Figure 1).

#### *4.3. Microscope Setup and Electrical Measurements*

An inverted epifluorescence microscope (Axio Observer Z1; Zeiss, Oberkochen, Germany) with 473 nm (blue) and 532 nm (green) laser illumination was used. As Arch-3 was tagged with enhanced green fluorescent protein (EGFP), we used the wavelength of 473 nm to excite the EGFP and monitor Arch-3 production (see Figure 2b). The electrical properties of the Arch-3-EGFP-containing bilayer were analyzed by electrophysiological measurements using a patch-clamp amplifier, EPC 10 USB (Heka Electronics, Reutlingen, Germany). For that purpose, Ag/AgCl electrodes were prepared by inserting a 5 cm-long silver wire into a borosilicate glass pipet containing 150 mM of NaCl electrolyte solution while applying 5 V for 30 min. The prepared electrodes were inserted into the inlets of the microfluidic device. The current passing through the bilayer was measured over time using an excitation signal with an amplitude of 20 mV and a time resolution of 100 ms.

**Author Contributions:** Conceptualization, N.K., M.F., R.S., A.O. and J.-B.F.; methodology, N.K., M.F. and J.-B.F.; software, N.K. and M.F.; validation, N.K. and M.F.; formal analysis, N.K. and M.F.; investigation, N.K. and M.F.; resources, R.S. and A.O.; data curation, N.K. and M.F.; writing—original draft preparation, N.K. and M.F. with help from A.O.; writing—review and editing, N.K., M.F., R.S., A.O. and J.-B.F.; visualization, N.K. and M.F.; supervision, R.S., A.O. and J.-B.F.; project administration, R.S., A.O. and J.-B.F.; funding acquisition, R.S., A.O. and J.-B.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the German Research Foundation (Projects B4, and C1 of CRC 1027) and, in part, by the Human Frontier Science Program (HFSP, RGP0037/2015).

**Conflicts of Interest:** The authors declare no conflict of interest.
