Micropatterning of functional lipid bilayer assays for quantitative bioanalysis

The plasma membrane is integral to the behavior and interactions of cells with their external environment. The cell membrane is decorated with a wide range of molecules that can impart many important functions, such as regulating material transport across the cell and directing intracellular signaling.¹ Thus, it has been of great interest to recapitulate the dynamic feature of the cell membrane in order to better understand membrane-related cellular functions. However, the cell membrane is known to be difficult to directly study with, primarily due to the varied membrane compositions across cells. Moreover, transient yet complex dynamic interactions that occur at different spatial and temporal scales between molecules further increase the dynamic composition and property of the membrane.² As a result, much effort has been placed on producing artificial membranes that can deliver high throughput analysis of molecular interactions with simplified model membranes. For this purpose, lipid bilayer membranes have been versatile as platforms for mimicking specific cellular conditions and are broadly used in the field of quantitative bioanalysis. Compared to other model membrane systems, supported lipid bilayers (SLBs) have a planar geometry that is formed on a solid substrate. The lipid composition is typically based on the biological system or cell type to be modeled and can be tailored accordingly.

Beyond the study of membrane properties, the versatility and capability of SLBs as a quantitative bioanalytical platform is further established by the flexibility to incorporate other biomolecules, including proteins, peptides, DNAs, etc., by various bioconjugation methods. One of the practical properties of SLBs is their non-fouling nature, which allows for the study of ligand–receptor interactions on membranes with minimal occurrence of non-specific binding while enhancing specific binding.³ There is also much interest in studying biomolecules, especially membrane proteins, which comprise over 70% of known drug targets.⁴ However, proteins and lipids behave differently in solution and in a cell membrane. At the molecular level, a membrane surface imposes translational and rotational constraints on membrane proteins and can induce conformational changes to these molecules. On average, this alters the protein–ligand collision rates and, hence, the reaction kinetics. Additionally, lipid and protein interactions with different ligands may induce changes, such as protein compartmentalization or clustering; thus, it is necessary to study membrane proteins in its native 2D environment.⁵ For similar reasons, cell signaling events can be better understood on an SLB. For typical signaling events, the overall duration and strength of the signaling responses are determined by the reaction kinetics of each elementary step during membrane localization, including membrane association and dissociation, and the lateral mobility of the recruited proteins. In this case, physical properties of membranes often have a net result of enhancing the collision rate. While the reduction of dimensionality accelerates reactions by increasing the local molecular concentration, the viscosity drag reduces the diffusional mobility and, hence, slows down reactions.^6,7

In this article, we give a brief overview of the current methods that have been demonstrated for micropatterning lipid membrane assays for quantitative bioanalysis. Advances in patterning methods have led to highly multiplexed systems that can deliver statistical amounts of data in a shorter time and with an efficient use of materials. These patterned SLBs have also been used in creative ways to address some challenges in analytical biology. Specifically, its utilization as an engineered bio-interface to study the dynamic cell signaling events and the interaction between lipid and biomolecules (proteins, small molecules, DNA, etc.) will be presented. It is noted that despite the advances in the lipid bilayer patterning, there is still potential to extend its range of applications. Likewise, the materials and equipment necessary for the sophisticated SLB patterning may present a barrier to entry into more widespread use. In this regard, we further discuss the future perspectives of the micropatterning of functional SLBs as high throughput and cost-effective platforms for modeling cellular interactions.

Lipid bilayer patterning generally occurs in a two-step process. First, a patterning element is deposited onto a substrate, such as glass. These patterning elements can function as building blocks to segregate SLBs, increase their affinity to the substrate, or facilitate their controlled removal. The SLBs are then deposited onto the substrate by a number of possible methods. However, the most used method is the substrate adsorption, rupture, and fusion of small unilamellar vesicles (SUVs). Vesicle fusion occurs as a result of the polar interactions between the lipid and glass. When lipid vesicles contact a hydrophilic surface, they adsorb onto the surface. Under sufficiently high density, adjacent vesicles spontaneously fuse together upon contact and rupture.⁸ The ruptured vesicles then spread across the substrate surface and self-assemble into a continuous lipid bilayer.^9,10

Several methods have been demonstrated to pattern SLBs. One of the most common strategies for confining SLBs into permanent arrays is by nano- or micro-fabrication of solid microstructures onto the underlying substrate. Often referred to as “corrals,” these solid microstructures act as diffusion barriers that prevent lipid diffusion and mixing across these corrals [Fig. 1(a)]. First proposed by Groves et al., photolithography was used to prepare micrometer-sized grid patterns of photoresist on silicon substrates. Followed by vesicle fusion, SLBs can be partitioned into the corrals. Smaller features of lipid bilayer arrays can be achieved by further processing of the patterned photoresists as negative photoresists for electron-beam (e-beam) lithography.^11,12 Additionally, multi-photon photolithography has also been implemented with photolithography to prepare patterns as small as 100 nm wide.¹³ Although both e-beam and multi-photon lithography are advantageous because of the higher spatial resolution, their wider utility is often limited due to lesser accessibility. Photolithographic patterning has been extensively used to spatially resolve juxtacrine signaling in T cells and cancer cells^14,15 as well as for single-molecule studies of protein–DNA interactions.¹⁶ Alternatively, light sources can be used for the photothermal ablation of lipid molecules from the substrate. For example, femtosecond laser pulses have been used to pattern patches of lipid bilayers by selectively removing lipids from a lipid bilayer coated surface.¹⁷ Photolithography can be similarly used to etch SLBs from a substrate with the application of a photomask to selectively expose lipids to the UV radiation [Fig. 1(b)].¹⁸ The exposed areas can be backfilled with different lipid vesicles or analytes to realize multiplexed patterning. This technique has been used to produce 1–4 μm diameter circles for cell receptor ligand immobilization to study the dynamics of ephrinA1:EphA2 signaling.^19,20

Soon after photolithography had been demonstrated for SLB patterning, the same technology was used to develop micropatterned molds to prepare elastomeric stamps for microcontact printing. The microscale features of the molds were replicated onto the stamps, typically made of PDMS, to directly transfer or strip molecules from a substrate. Three methods of SLB patterning using microcontact printing have been proposed. In the first case, called “protein printing,” the PDMS stamp is dipped into a protein solution and then stamped onto a substrate. The proteins are immobile and act as diffusion barriers for the mobile lipids. The SLB is then deposited onto the substrate by the rupture and fusion of SUVs on the exposed surface. In the second case, called “protein caulking,” the SLB is first deposited onto the substrate without any diffusion barriers. The stamp is pressed onto the bilayer to selectively remove lipids. Proteins are then deposited to fill in the exposed area.²¹ Hovis and Boxer further developed the technique to directly transfer an SLB by first depositing the SLB onto the stamp [Fig. 1(c)].²² The resolution of the stamps used for microcontact printing depends on the quality of the mold. Because of the facile SLB patterning offered by microcontact printing, it has become a widely used method with applications in the study of sphingomyelinase−lipid nucleation kinetics,²³ lipid raft localization,²⁴ and lipid raft kinetics.²⁵ 

Through techniques like microcontact printing, it became possible to pattern SLBs alongside other biomolecules, such as proteins or other types of lipids. Such platforms provided researchers with an improved control to recapitulate the biological environment to perform cellular studies and allowed the simultaneous investigation of cellular interactions with lipids and proteins. The crucial step in preparing these platforms is the selective deposition or removal of lipids from the SLB. To this end, a strategy called a polymer lift-off technique was developed.²⁶ Polymer lift-off techniques pattern SLBs by the deposition of SLBs onto a polymer film that can be physically peeled off to selectively remove lipids [Fig. 1(d)]. In this method, di-para-xylene (parylene) and a photoresist are first successively deposited onto the substrate. Using a combination of photolithography and reactive ion etching, a patterned parylene film is left on the substrate. The SLB is then deposited by vesicle fusion onto the whole substrate, including the parylene layers. The parylene layers are then peeled off to selectively remove lipids. Like microcontact printing, the resolution depends on the photolithography step. Polymer lift-off has been demonstrated for the preparation of patterned SLBs for the study of mast cell activation,²⁷ bacterial toxin binding,²⁸ etc.

Recently, a different lift-off technique was developed in combination with microcontact printing for the selective deposition of SLBs onto a substrate. Microcontact printing has been extensively used to transfer different materials onto various substrates.²⁹ Of these materials, self-assembled monolayers (SAMs) are commonly used.³⁰ SAMs are notable for their stability and well-controlled surface characteristics. Belling et al. demonstrated that bicelles can be ruptured onto the SAMs patterned on a gold substrate by microcontact printing in a process called “chemical lift-off lithography” (CLL) [Fig. 1(e)].³¹ In this method, the SAM with hydrophilic group is first deposited onto a gold substrate. A pattern is produced by the lift-off of SAM with a PDMS stamp. The exposed substrate is then backfilled with a SAM bearing a hydrophobic end group. After incubation with bicelles, the SLB is spontaneously formed onto the hydrophilic SAMs because of polar interactions with the lipid head group. Importantly, the results also demonstrated that the hydrophobic SAM does not permit the diffusion of lipids across it.

Regarding lipid membrane arrays, there has also been interest in patterning them with nanopatterning methods. Colloidal lithography takes advantage of the sub-micrometer spaces between self-assembled monolayers of colloidal spheres as masks for the selective thermal deposition of metals, such as gold [Fig. 1(f)].³² After removal of colloids, an SLB can be deposited by vesicle fusion.³³ This patterning strategy is of interest because the gold nanostructures have plasmonic properties that allow for their use as optical nanoantennas for surface-enhanced Raman spectroscopy (SERS). This allows for label-free imaging and fluorescence enhancement of low quantum efficiency dyes. Additionally, colloidal lithography enables the facile preparation of nanostructures without the need for the photolithographic or e-beam etching equipment. For example, Lohmüller et al. demonstrated that with an additional plasma-etching step, the colloidal particles could be melted together such that they shrink while forming threads at their points of contact.³⁴ When used as a mask for gold deposition, triangular nanostructures could be prepared with gaps as small as 10 nm, significantly enhancing the Raman signals by 10⁴ times and the fluorescence intensities by 2.5–6.8 times. The signal enhancements offered by colloidal lithography have also been demonstrated to probe the organization of lipid bilayers under the influence of cholesterol³⁵ and glycans.³⁶ 

Atomic force microscopy (AFM) has also been used to directly “write” lipids onto a substrate in a method termed “dip-pen nanolithography” (DPN). Under controlled humidity, the tip of the cantilever can be coated with lipids and transferred onto a substrate by contact with the tip [Fig. 1(g)]. At >90% humidity, the deposited lipids spread and assemble into a single bilayer featuring a linewidth of 250 nm.³⁷ Alternatively, strict humidity control can be avoided by using AFM nanoshaving. Here, proteins, such as bovine serum albumin (BSA), are first deposited onto the substrate followed by selective etching with the AFM tip. The exposed area can then be backfilled with vesicles to deposit SLBs without consideration of the humidity, thereby simplifying the process and removing the limitations on lipids that can be used. Compared to direct writing, nanoshaving of proteins resulted in a minimum linewidth of 55 nm.³⁸ Other modifications to the DPN technique include using multiple cantilevers to parallelize the writing of multiple lipid patterns. This could increase the throughput of patterning to realize multiplex SLB-binding studies.³⁹ The primary advantage of DPN is the capability to design arrays of sub-micrometer patterns, allowing observation of multiple SLB patches in a single field of view and the minimization of the amount of biomolecules needed for experiments. For example, DPN has been used to prepare SLB arrays to study glucocorticoid receptor localization in the study of mast cell signaling.⁴⁰ 

Developed simultaneously to microcontact printing, microfluidic lithography uses microfluidic channels to deposit lipid bilayers onto a substrate. A typical microfluidic patterning chip features parallel microchannels with widths of a few micrometers and independent fluid inlets and outlets. Lipid vesicle solutions are injected into the channels, where they can rupture and undergo vesicle fusion to form the SLBs [Fig. 1(h)].⁴¹ Each microchannel is independently addressable, giving flexibility in the type of analyses that can be performed. For example, ligand binding studies can be performed by depositing SLBs with similar composition into the microchannels. Solutions with different ligand concentrations can then be introduced into each channel. Various optical microscopes, such as epifluorescence or total internal reflection fluorescence microscopy (TIRFM), can then be used to produce a lipid-ligand binding curve from a single experiment.^42,43

Compared to other methods, microfluidic patterning strategies are advantageous in that the channel design can provide direct manipulation of fluids. Furthermore, microfluidic channels can be designed with higher complexity. For example, microfluidic chips can produce solute gradients using designs, such as a Y-channel mixer⁴⁴ or a “Christmas tree” gradient generator.⁴⁵ Furthermore, multi-layer chips can be designed with hydraulic valves for precise control of multiple fluid flow.⁴⁶ Maerkl and Quake leveraged this valve strategy to produce arrays of button membranes above microfluidic channels to mechanically entrap transcription factor complexes.⁴⁷ The button membranes are large circular chambers separated from the microfluidic chamber by a thin membrane [Fig. 1(i)]. Upon hydraulic or pneumatic pressurization, the membrane expands and pushes down on the floor of the chamber. This strategy has also been used to prepare circular patterns of proteins.⁴⁸ When the button is fully pressurized, molecules cannot be deposited onto the surface. Slight reduction of the pressure decreases the expansion of the button, exposing a small ring-shaped area of the substrate, which can be backfilled with new molecules. This process can be repeated until the button is completely deflated. This results in concentric rings of the patterned material. Recently, we have applied this strategy combined with the Christmas tree gradient generator to prepare multiplexed SLBs to study NF-κB activation in fibroblast cells.⁴⁹ 

Patterned SLBs have been demonstrated as robust platforms for studying dynamic cell signaling events.⁵⁰ Particularly, juxtacrine signaling complexes formed by the intercellular binding of ligands and membrane receptors may be recapitulated on a planar SLB surface. This approach has provided valuable insights into cell signaling systems, such as the immunological synapse formed by T cells,⁵¹ integrin adhesion of cancer cells,⁵² small GTPase (Ras) activation by guanine-nucleotide exchange factor (SOS),^53–55 and neurite initiation in neurons.⁵⁶ It is common for signal transduction across the plasma membrane to be accompanied by the coalescence of membrane components into signaling microclusters (MCs). Cell signaling studies, thus, require a fluid environment where the ligands can freely diffuse across the membrane. During T cell responses, MCs rearrange to form an immunological synapse (IS), a dynamic protein array that modulates T cell activation. Hsu and co-workers investigated the influence of membrane fluidity on the activation of T cells by modulating the fluidity of SLB-bound ligands. They observed the slow MC and IS formation with low mobility ligands, resulting in delayed T cell activation. This study highlights the importance of ligand mobility on cell signaling.⁵⁷ More recently, Chen et al. used a micropatterned SLB for the spatially resolved study of the crosstalk between EphA2:ephrinA1 and integrin:RGD adhesion. Quantitatively studying these two receptor signaling systems is inherently difficult because the EphA2:ephrinA1 system is localized to the lateral side of the cell (cell–cell), while the integrin:RGD system is localized to the basal side (cell–tissue). By micropatterning arrays of mobile SLB-ephrinA1 surrounded by immobile RGD on a planar substrate, cellular behavior can be observed as the cell spreads across the surface [Fig. 2(a)]. Coupled with single-molecule tracking, Src kinase, which is involved in the regulation of cell migration, was found to translocate from EphA3:ephrinA1 to integrin:RGD. The crosstalk between the two receptors triggered the simultaneous assembly and disassembly of focal adhesions around the integrin clusters, thus stimulating cell migration.¹⁹ In addition, spatially patterned SLBs coupled with single-molecule tracking have been demonstrated as powerful tools for studying cell signaling. These have also been used to study other signaling events, such as the T cell feedback mechanism that modulates T cell receptor–pMHC binding⁵⁸ and the EphA2:ephrinA1 recruitment of ADAM10 and clathrin to mediated ephrinA1 trans-endocytosis.¹⁵ 

On the other hand, patterned SLBs have also been used as analytical platforms to investigate how lipid bilayers interact with different molecules, such as proteins,⁵⁹ charged analytes,⁶⁰ small molecules,⁶¹ and metals.⁶² This is particularly useful for modeling diseases that affect cellular membranes. For example, neurodegenerative diseases have been linked to abnormally high levels of oxidized lipids. In vitro models for lipid oxidation are vital in improving our understanding of these diseases. Poyton et al. used a microfluidic patterned SLB to study lipid oxidation by reactive oxygen species produced by Fenton-like reactions involving Cu²⁺ and H₂O₂. These patterned SLBs produce parallel channels of independent SLBs, which allow different lipid compositions to be formed. Fluorescence monitoring of the lipids indicated that Cu²⁺ binds bivalently to phosphatidylethanolamine (PE), which is the second most abundant lipid in cells [Fig. 2(b)]. Furthermore, the rate of oxidation was linearly related to the concentration of PE, reaching a maximum of 8.2 times the rate of PE-free bilayers.⁶³ Although such models may not completely simulate cellular conditions, the results provide insights into a high level of lipid oxidation associated with neurodegenerative diseases. Alternatively, patterned SLBs may serve as tools for evaluating the insertion mechanism of therapeutic compounds into cellular membranes. For example, Sun et al. elucidated the mechanism of ibuprofen insertion into lipid membranes using microfluidic patterned SLBs. The chip produced four parallel channels in which SLBs could be treated with varying concentrations of ibuprofen. Using vibrational sum frequency generation spectroscopy, changes in the water (–OH) and alkyl (–CH₂) content of the membranes determined that ibuprofen insertion occurs in three steps: (1) electrostatic adsorption of ibuprofen to the lipid headgroup at <10 μM, (2) hydrophobic insertion at 10–300 μM, and (3) solubilization of the membrane at >300 μM. These were complemented with fluorescence recovery after photobleaching (FRAP) measurements, which observed the increase in membrane fluidity until its disruption.⁶⁴ 

Another interesting demonstration of patterned SLBs for analytical applications is the DNA curtain.⁶⁵ The DNA curtains feature lipids functionalized with a single double-stranded DNA molecule. The DNA-lipids are incorporated onto an SLB, which provides the DNA with a biologically functionalized and passivated environment. Under hydrodynamic force, the DNA-lipids are pushed toward solid diffusion barriers. The diffusion barriers are designed with a height of a few nanometers such that the DNA strands extend over them under flow conditions. This produces arrays of DNA strands that extend toward the flow direction, allowing each strand to be visualized linearly by high resolution imaging techniques [Fig. 2(c)]. By incorporating different fluorophores on binding proteins, their movement along a strand can be visualized. Notably, DNA curtains were used to visualize the mechanism by which the mismatch repair complex Msh2–Msh6 inspects undamaged DNA. Using TIFRM, the movement of Msh2–Msh6 could be tracked as it traveled along the DNA strands. Since this occurred despite an opposing hydrodynamic force, it was deduced that Msh2–Msh6 moved along the DNA strands using a “sliding” mechanism wherein it was in constant contact with the DNA during the inspection.⁶⁶ Others have leveraged DNA curtains for studying DNA–protein interactions.⁶⁷ Here, the bound base pair of a known DNA strand can then be determined from the position of the protein along the curtain.

As simple models for cellular membranes, SLBs with synthetic lipids are useful for studying specific cellular phenomena across different biomolecules and lipids. However, the findings from simple synthetic models must be complemented and verified with observations acquired from membranes that more closely resemble living cells. While directly studying cell membranes remains a formidable task, an “intermediate” system named giant plasma membrane vesicle (GPMVs) has been used to bridge the gap between SLBs and cell membranes.⁶⁸ GPMVs are lipid vesicles budded directly from live cells and, thus, enriched with a wide variety of native membrane proteins and phospholipids,⁶⁸ which allows the reconstitution of some lipid–protein and even protein–protein interactions that are lacking in the synthetic membrane. Compared to giant unilamellar vesicles (GUVs), which are synthetic membranes broadly used to study membrane biophysics, GPMVs offer more natural and cell membrane-resembling properties.⁶⁹ Notably, studies have shown that GUVs and GPMVs have different properties. For example, the localization patterns of glycoproteins like influenza virus hemagglutinin (HA) in lipid phases were different between GUVs and GPMVs,⁷⁰ as it has been observed that the contrast among lipid phases in GUVs was more distinct compared to GPMVs, which show some degree of mixing.⁷¹ Because they are derived from cells, GPMVs are enriched with transmembrane proteins that influence the membrane properties.^72–74 Thus, patterning lipid bilayers using GPMVs could be advantageous in accelerating the progress of understanding bio-membrane functions.

An intrinsic challenge in fabricating planar membranes of GPMVs is how to rupture them without physically or chemically influencing their integrity. Recent developments, nevertheless, have presented opportunities to use microfluidic patterning to achieve the fabrication of planar GPMVs. For example, Chiang et al. successfully ruptured GPMVs onto a substrate by compression with an air–water interface that passed above the GPMVs.⁷⁵ Sezgin et al., on the other hand, integrated a piezoelectric microfluidic element that applied acoustic pressure to rupture GPMVs.⁷⁶ Most recently, Teiwes et al. also demonstrated that GPMVs can be ruptured and spread onto an oxygen plasma-activated silicon substrate.⁷⁷ These methods have showed that the mobility of transmembrane proteins and lipid domains was retained on the planar GPMVs. Note that despite these efforts, currently little, if any, effort focuses on making planar GPMVs with complex patterns and high throughput. In this regard, effort into developing microfluidic strategies for patterning supported GPMV membranes will be interesting [Fig. 3(a)]. Instead of SLBs, GPMVs could be used to study protein–ligand or lipid–ligand interactions at a more native environment. Compared to other methods, such microfluidic GPMV membranes may be an ideal system for high throughput studies of bio-membranes derived from mass-limited cell samples, such as circulating tumor cells, rare cells, or clinical biospecimen and samples.

In addition to GPMVs, free-standing lipid bilayers (FSLBs) offer opportunities to further bridge the gap between SLBs and cell membranes. FSLBs are lipid bilayers that are not anchored to a solid substrate. There is much interest in FSLBs because they can be used to study transport proteins, which play a key role in pharmacology since they regulate the drug uptake into the cell.⁷⁸ More importantly, transport proteins are targeted in the design of drugs against multi-drug resistant cancer cells.⁷⁹ However, standard assays for trans-cellular drug transport require the cell monolayers.⁸⁰ Using model membranes with single transporters could be beneficial as tools to rapidly study drug transport through specific transport proteins.

Currently, arrays of FSLBs have been produced by patterning lipids by two ways: SLB formation on nanopore microchip arrays or lipid “painting” across microcavities.⁸¹ Nanopore arrays are silicon substrates with femtoliter cavities. The bottom opening is capped with a glass slide for microscopy, while the top surface is capped with a silicon nitride substrate patterned with an 80-nm hole array. The small hole size allows the bilayer to spread efficiently by vesicle fusion and remain stable for the experimental duration. The cavities, meanwhile, are used to observe the flux of materials across the transmembrane proteins embedded within the SLB. Fluorescence microscopy can be used to determine the amount of flux through the proteins across the nanopores.^82,83 The small pore size and a large substrate area allow the bilayer to be free-standing at the pores while retaining membrane fluidity. These can accommodate single transmembrane proteins for single transport recordings, and the vast number of arrays allows for high throughput studies.⁸⁴ On the other hand, lipid bilayers can be sequentially suspended across microcavities or microwells to produce FSLBs. For example, the microwells can be fabricated from a fluoropolymer called CYTOP and covered with a glass block with a hole for sample injection. Compared to the nanopore array wherein the lipid bilayer is formed by vesicle fusion, the FSLBs in microwells are formed by LB transfer, which is often called lipid “painting” because of the layer-by-layer deposition of the lipid membrane. The FSLBs are produced by sequentially flowing in an aqueous solution, a lipid solution in chloroform, and a second aqueous solution. The first aqueous solution fills the wells and comprises the external environment of one leaflet of the membrane. Subsequent lipid introduction in chloroform pushes out the aqueous solution above the wells while entrapping the fluid inside. This produces an interface for the lipids to “paint” a monolayer across the top orifice of the wells. As the second aqueous solution pushes out the chloroform solution, excess lipids assemble onto the first lipid monolayer to produce a bilayer.⁸⁵ Additionally, a second lipid solution can be injected after the first one to generate an asymmetric lipid bilayer.⁸⁶ This can be useful in cell-mimicking since cells naturally have asymmetric membranes.⁶⁹ Like nanopore arrays, fluorescence measurements of the wells can be used to examine the flux of materials through the transmembrane proteins. This platform has been demonstrated to study molecule transport through α-hemolysin,⁸⁷ F-type ATPase,⁸⁸ and the phospholipid scramblase TMEM16F.⁸⁹ 

Although the two methods of FSLB formation allow for high throughput analysis, they have yet to be used for more multiplexed experiments. For example, the cavity-based microarrays are all enclosed within a single channel wherein the purpose of the chamber is solely for injecting the solution. Microfluidic patterning has been used to pattern SLBs from different formulations of lipids into separate channels for multiplexed analysis. It would be interesting to use microfluidic patterning to produce FSLBs with different compositions on the cavity-based microarrays [Fig. 3(b)]. Additionally, FSLBs from GPMVs may also provide model cell membranes that more closely mimic live cells. For example, Kuo et al. have recently been able to produce a facile method to rupture GPMVs across arrays of microwells. Preparation of the microwells over a gold and titanium substrate allowed for the label-free detection of glucose across the pore-spanning GPMV membranes using surface plasmon resonance.⁹⁰ Free-standing membrane platforms could further benefit from pore-spanning GPMVs by taking advantage of their naturally asymmetric bilayers. Further work can also be done to improve the density and complexity of FSLBs.

As important as SLB patterning is to the development of cellular membranes, the analytical methods used for the readout are of equal importance. Progress in SLB patterning has been supported by its compatibility with established analytical tools, such as various optical microscopies. This has enabled many laboratories to adapt SLB patterning into the workflow.^91,92 SLB arrays have seen the most benefit from microscopy techniques because their planarity offers sufficient surface area for microscopic resolution, and that multiple arrays can be simultaneously investigated and compared in a single image.⁹³ 

Among possible readouts, lately, there has been interest in adapting smartphones as analytical tools because of their portability and wide availability. Each generation of smartphone has seen an increase in both processing speed and imaging capability. The rate of these improvements has made smartphone cameras competitive and compelling mode for assay readouts.⁹⁴ Thus, smartphones have potential to be utilized as analytical tools as they have both the processing and imaging quality required for optical analysis. Although the camera lenses were not designed for microscopy, simple external modifications can be done at a low cost.^95,96 An external mount with optical filters and LEDs can be constructed to make smartphone-based fluorescence microscopes.⁹⁷ Currently, smartphones have been shown to image fluorescent beads up to 0.8 μm in diameter, demonstrating their resolving power.⁹⁸ Smartphone microscopy has indeed been demonstrated for applications in cell counting^99,100 and analyte quantification.^101,102 An interesting development from such analytical assays is the use of analytes patterned as QR codes for rapid analysis.¹⁰³ As illustrated in this Perspective, many patterning techniques can produce square or circular SLB arrays; thus, a QR code-based readout may be easy to implement in the design. Aside from streamlining the analytical workflow, this could be useful in the clinical settings as a point-of-care system or when working in the environment with limited resources.

As discussed, SLB patterning has become a powerful tool for researchers to study quantitative bioanalysis. Yet, certain patterning strategies may not readily be available to users due to limited resources and inaccessibility to instrumentations. As such, to maximize the utility of patterned SLBs for biological sciences, an effort should be made to find simple and accessible methods of producing desired types of SLBs. Among possible options, 3D printing is of particular interest because its development and utility has seen exponential growth in the past few decades, ranging from niche hobby machines to highly commercialized devices that have found ubiquity in consumer, industrial, and research applications.¹⁰⁴ For instance, digital light projection (DLP) is a common 3D printing method, where a liquid crystal display (LCD) projects the desired pattern onto a vat of photo-crosslinkable polymer resin.¹⁰⁵ A DLP 3D printer is, thus, essentially a miniaturized photolithography system that has potential to serve as an economic alternative for producing molds for microcontact printing or microfluidic devices.

Furthermore, 3D-printed microfluidic chips have an intrinsic property, which allows high flexibility to create multi-layer devices and, thus, spares the time and effort to make molds of individual layers used in conventional multi-layer construction. Such flexibility enables the construction of 3D microfluidic chips with perpendicular channels or unconventional channel shapes.^106–108 For example, Heo et al. have recently leveraged this feature to fabricate a two-layer microfluidic chip with perpendicular channels to form asymmetric FSLBs.^109–111 Compared to FSLBs suspended across microcavities, this FSLB is suspended across a larger area, and both leaflets can be selectively exposed to different solutions via their respective channels. Note the current design is still limited to a single interface per chip, while the nanopore and microcavity-based chips can have arrays of tens-of-thousands of wells.

Directly 3D printing microfluidic chips for lipid bilayer patterning may also be a means to further simplify the tool production for SLB patterning [Fig. 3(c)]. With respect to lipid bilayers, current studies have been demonstrated for liposome generation¹¹² and lipid droplet interfaces.¹¹³ However, the biocompatibility of 3D printed chips has been demonstrated for several biological applications, such as cell culture,¹¹⁴ cell encapsulation,¹¹⁵ biosensing,¹¹⁶ and organ-on-a-chip.¹¹⁷ Additionally, chips can be printed with built-in hand-driven pumps, namely, a push button design, to obviate the use of bulky external pumps.^118,119 Although there is still much room to improve the throughput and multiplexity of 3D printing techniques, these can be practical economic alternatives to other lipid bilayer patterning methods or platforms for the rapid prototyping of new strategies.

In this article, the current methods of SLB micropatterning were briefly reviewed and compared to discuss their development, working principle, use case, and some advantages and limitations. The accelerated development of microfabrication technologies over the past three decades has trickled down to analytical biology, offering techniques to fabricate micro-scale platforms for studying single biomolecule dynamics. This has resulted in several micropatterning strategies for SLBs that could be used to study a dynamic cell phenomenon. However, there is still much to be improved, such as the cell-mimicry of the model membrane, integration of new tools into established systems, reduce the complexity of some micropatterning techniques, and address the lack of accessibility to these techniques. Improvements in the integration with emerging rapid and economic fabrication technologies could result in the development of the next-generation platform for SLB micropatterning. Together with the development of supported cell-derived membranes, micropatterned planar lipid bilayers can lead to advances in our understanding of dynamic cell processes.

This study was supported by the Ministry of Science and Technology in Taiwan (MOST 110-2113-M-001-019-MY2).

The authors have no conflicts to disclose.

Reynaldo Carlos K. Montalbo: Data curation (lead); Formal analysis (lead); Investigation (equal); Visualization (lead); Writing – original draft (equal); Writing – review & editing (equal). Hsiung-Lin Tu: Conceptualization (lead); Funding acquisition (lead); Investigation (equal); Supervision (lead); Writing – original draft (equal); Writing – review & editing (lead).

Data sharing is not applicable to this article as no new data were created or analyzed in this study.