Anal Bioanal Chem (2009) 393:1601–1605 DOI 10.1007/s00216-008-2588-5
TECHNICAL NOTE
Bilayer lipid membranes from falling droplets Michele Zagnoni & Mairi E. Sandison & Phedra Marius & Hywel Morgan
Received: 23 October 2008 / Revised: 20 November 2008 / Accepted: 17 December 2008 / Published online: 19 January 2009 # Springer-Verlag 2009
Abstract We describe a system that provides a rapid and simple way of forming suspended lipid bilayers within a microfluidic platform from an aqueous droplet. Bilayer lipid membranes are created in a polymeric device by contacting monolayers formed at a two-phase liquid–liquid interface. Microdroplets, containing membrane proteins, are injected onto an electrode positioned above an aperture machined through a conical cavity that is filled with a lipid–alkane solution. The formation of the BLM depends solely on the device geometry and leads to spontaneous formation of lipid bilayers simply by dispensing droplets of buffer. When an aqueous droplet containing transmembrane proteins or proteoliposomes is injected, straightforward electrophysiology measurements are possible. This method is suitable for incorporation into lab-on-a-chip devices and allows for buffer exchange and electrical measurements. Keywords Biosensors . Self-assembled monolayers . Electrophysiology
M. Zagnoni (*) : M. E. Sandison : P. Marius : H. Morgan School of Electronics and Computer Science, University of Southampton, Highfield, So17 1BJ, Southampton, UK e-mail:
[email protected] Present address: M. Zagnoni Department of Electronics and Electrical Engineering, Faculty of Biomedical and Life Sciences, University of Glasgow, G12 8QQ Glasgow, UK Present address: M. E. Sandison Integrative and Systems Biology, Faculty of Biomedical and Life Sciences, University of Glasgow, G12 8QQ Glasgow, UK
Introduction Lipid bilayers are the fundamental element of all biological membranes, hosting transmembrane proteins that regulate the majority of cellular functions. Artificial planar lipid membranes have been used since 1960s for bioanalytical and electrochemistry laboratory experiments [1–2]. However, in more recent years, there has been growing interest in developing novel methods for forming artificial bilayer lipid membranes (BLMs) within microsystems. In the past 10 years, BLM-based systems have found widespread applications in many fields of biology [3] and in the development of ultrasensitive biosensors [4–5]. They also have uses in proteomics, where new high throughput analysis platforms are required to test biomarkers for drug screening [6–7]. Artificial BLM-based biosensors can be used as single molecule sensors by exploiting transmembrane protein ligand gating as a detection mechanism. A number of different approaches have been proposed for labon-a-chip (LOC) BLM systems [8–15], but reliability and reproducibility are still a major concern. In particular, the BLM formation procedure often requires manual expertise, thus hampering the rapid development of lipid membrane based biosensors on a large scale, where automation and ease of use is required. Recently, a new approach to forming artificial lipid bilayer has been described [16–17]. These methods differ from traditional approaches, having a lipid–alkane solution as the bulk phase. At the interface between an aqueous droplet and the organic solution, a lipid monolayer forms between the two phases. When two droplets (or two buffer streams) are brought into contact, a BLM forms at the droplet–droplet (or buffer–buffer) interface (from the two contacting monolayers). This approach has been used to form single BLMs and BLM networks. Manipulation of the aqueous phase is achieved either manually (using micromanipulators) or by driving buffer through microfluidic
1602
channels using syringe pumps, thus necessitating the use of external components. This approach, using alkane and lipids as a bulk phase, has also been used to create a BLM at the interface between a water droplet and an agarose gel layer [18]. This setup creates gel-supported BLMs, and so, buffer perfusion around the BLM is not possible. In addition, transmembrane proteins were incorporated in the gel prior to droplet delivery (a procedure that is not suitable for all ion channels). Very recently, robotic dispensing of liquid droplets has also been employed to form multiple BLMs in an array format [19]. However, this architecture requires the positioning of electrodes into pre-dispensed droplets, and the mixing or exchange of buffer was not demonstrated. Here, we describe a system that provides a rapid and simple way of spontaneously forming suspended lipid bilayers within a microfluidic platform with a near 100% yield without using external instrumentation. The system requires only the dispensing of a water droplet onto a wire electrode, positioned at the center of a conical-shaped reservoir as shown in Fig. 1. The reservoir contains a lipid–decane solution and is connected via an aperture to a lower microfluidic channel containing buffer. Owing to the density difference between the aqueous and organic solutions, the droplet sinks to the bottom of the conical reservoir and forms a BLM at the interface between the droplet surface and upper surface of the channel buffer. The system is simple and enables electrical measurements without the need for micromanipulators. The device geometry also allows several droplets to be merged with Fig. 1 a The polymeric device used to form BLMs. b–d Procedure for forming the BLM by dispensing a droplet onto the wire electrode. e Schematic of a BLM formed between the water droplet and lower buffer
M. Zagnoni et al.
one another, allowing for the alteration of buffer composition. If the droplet contains proteins, ion channel behavior can be monitored, thus rendering the platform suitable for the development of ultrasensitive biosensors.
Materials and methods Materials Asolectin (Sigma) was dissolved in decane to a concentration of 20 mg/mL. Ion channel measurements were made with 100 μL of gramicidin (1 ng/mL in ethanol) in 1 mL of 100 mM KCl, pH 7 or 100 μL of alamethicin (40 μg/mL in ethanol) in 1 mL of 100 mM KCl, pH 7. KcsA reconstitution and delivery method KcsA potassium channels were reconstituted into lipid vesicles. KcsA with an N-terminal hexahistidine tag was over-expressed and purified as previously described [20]. A combination of POPE/POPG/ergosterol, at molar ratios of 5.5:3:1.5, was mixed in chloroform, dried, and re-suspended in buffer (10 mM HEPES, 450 mM KCl, pH 7.4) containing 40 mM b-D-octyl glucoside to give a total lipid concentration of 10 mg mL−1. The sample (1–2 mL) was then sonicated to optical clarity in a bath sonicator. KcsA (40 mg) was mixed with the lipid sample to give a 22,000/1 molar ratio of lipid/KcsA tetramer. Detergent was removed using SM2 Bio-Beads, which produced 150–350 nm
Bilayer lipid membranes from falling droplets
proteoliposomes (diameter estimated from light scattering) that were kept on ice until use or stored at 4 °C for up to 2 weeks. Droplets (2 μL) of proteoliposome solution were dispensed when the buffer in the lower channel was 150 mM KCl, pH 4 [11]. Device structure and droplet dispensing procedure The device structure is illustrated in Fig. 1a. It consists of a bottom channel made from a sheet of polymer on top of which a small channel is defined using double sided adhesive tape. The top half is made from polymethylmethacrylate and has a conical cavity created by mechanical milling. At the bottom of the conical well, an aperture (400–800 μm in diameter) was machined to connect the top reservoir and the lower channel. Figure 1b–d illustrates the principle of BLM formation. First, the lower channel is filled with electrolyte (100 mM KCl) and the top reservoir with lipid–alkane solution. A single Ag/AgCl electrode is inserted into the lower channel
Fig. 2 a Capacitance trace of typical BLM formation event showing that a stable lipid bilayer is formed in 10–40 s. An arrow indicates the time at which a droplet is dispensed. b A combined capacitance and current trace showing bilayer formation (capacitance), followed by a current check, a second capacitance measurement, and lastly moni-
1603
inlet. A second Ag/AgCl electrode (500 μm diameter) with a sharpened tip (approximately 50 μm) covered with agarose gel (to ensure good electrical connection with the droplet) is positioned in the center of the cone, 1 mm above the aperture. After filling the top half of device with the lipid–decane solution, a droplet of buffer (3 μL of 100 mM KCl) is dispensed onto the agarose-coated Ag/AgCl electrode. The drop adheres to the wire and quickly descends to the bottom of the reservoir (<1 s). The conical geometry ensures that the droplet comes to rest over the aperture. A BLM spontaneously forms, as illustrated in Fig. 1e.
Results and discussion Bilayer formation A typical capacitance trace obtained when forming BLMs is shown in Fig. 2a, where after droplet injection, the
toring of the current across the bilayer showing the reconstitution of gramicidin channels (inset shows close-up of subsequent gramicidin activity, recorded at 1 kHz and low-pass-filtered at 100 Hz). c Alamethicin and d KcsA current traces, recorded at 10 kHz and lowpass-filtered at 2 kHz
1604
capacitance rises quickly to a stable value. The transmembrane current across the BLM was approximately 1 pA with 100 mV applied. The typical lifetime of bilayer without any ion channels was approximately 2 h, while when proteins were present in the bilayer, the lifetime was reduced to around 30–60 min (long enough to acquire electrophysiological data). Capacitance measurements provide information on the thickness and the area of the dielectric (the decane and lipids) between the droplet and the lower channel. Over the course of 20 experiments, using six different devices, BLMs were nearly always obtained (95% yield), with capacitance values in the range of 30–300 pF, for aperture sizes of 400–800 μm. The specific membrane capacitance for lipid bilayer membranes is in the order of 0.5 μF cm−2 [5], and this value can be used to estimate the BLM diameter from the measured capacitance. Estimated diameters were in the range of 85 to 270 μm, sizes considerably smaller than the diameter of the aperture. The results showed that the smaller the aperture, the lower the capacitance, although there was no clear relationship between BLM area and aperture size. Further work is required to understand the relationship between the bilayer area, the droplet volume, the aperture size, and the angle formed between the lower channel and the conical well. A BLM always formed, provided that the aperture region was smooth and tapered to a point. In other situations, where the aperture was deep or where debris or non-uniformities were present around its edge, a BLM did not form. This could be due to a thick layer of alkane solution becoming trapped between the droplet and the channel. These observations suggest that geometry is the key factor in ensuring reliable BLM formation. As discussed below, the presence of a bilayer was confirmed by delivering and recording activity from peptides and transmembrane proteins. Additional droplets could also be dispensed on top of the initial one. When a second droplet collided with the first, the two droplets merged (as a result of the force of impact). After the second and third droplet injections, the BLM capacitance changed slightly (data not shown), while after the fourth injection (on average), the BLM ruptured or became electrically leaky. This technique enables modification of droplet buffer concentration, while the solution in the lower channel could be exchanged using a pump. Electrophysiology In order to demonstrate ion channel recording, gramicidin and alamethicin were reconstituted into BLMs (Fig. 2b–c). Both gramicidin and alamethicin are peptides that spontaneously incorporate into bilayers from solution forming monovalent, cation-selective, and voltage-dependent ion channels, respectively [21]. Channels formed from gramicidin dimers are well-known for spanning only the
M. Zagnoni et al.
thickness of a single lipid bilayer; thus, the electrical observation of gramicidin ion channels confirms the presence of a BLM. The reconstitution of gramicidin and alamethicin was achieved in two different ways. In one method, a BLM was first formed using a buffer droplet. A second droplet containing ion channels was then allowed to run down the electrode. This droplet merged with the first one, leading to ion channel activity. Alternatively, a droplet containing peptides was dispensed in a single injection. As an example of the latter, a capacitance and current trace showing BLM formation and subsequent channel activity are shown in Fig. 2b. T1 represents the time before a stable capacitance value was obtained (40–80 s on average). During T2, a constant potential of 100 mV was applied and the current was measured, confirming a low transmembrane current. In T3, the capacitance was again measured, showing that an applied voltage changes the capacitance slightly (ΔC in Fig. 2b). Several experiments confirmed this trend, although the capacitance increase varied from 10 to 100 pF. During T4, the current was again measured, and gramicidin activity was observed (gramicidin molecules now having incorporated into the BLM from the solution). The BLMs obtained were stable up to an applied voltage of 300–400 mV, and the current noise was 1.2 pA (1 standard deviation) at 100 mV, sampling at 1 kHz. Figure 2c shows alamethicin activity obtained using the first delivery method. This method demonstrates that the platform allows for buffer alteration during electrophysiology without any significant change in BLM stability or recording conditions. Ion channel activity was also obtained using non-watersoluble proteins that do not spontaneously insert into membranes. Bilayers were formed using droplets containing small unilamellar lipid vesicles loaded with the K+ channel KcsA (MW 67 kDa for the tetrameric form). To initiate vesicle fusion and KcsA channel opening, a salt and pH gradient were established between the droplet and the lower channel, as reported previously [11]. This was achieved by either forming a BLM using a single droplet that contained proteoliposomes and had a higher salt concentration than lower channel or using a three-step procedure. In this case, the first droplet formed the BLM (using the same salt concentration as in the lower channel), the second delivered proteoliposomes, and the third adjusted the salt concentration to produce the required gradient. A recording of KcsA activity is shown in Fig. 2d. Only vesicles that fuse completely with the BLM at the interface between the droplet and the channel lead to KcsA activity. The number of vesicles incorporated into the droplet (∼10+6) suggests that most are likely to remain in solution or to only adhere to the lipid surface (adhesion of vesicles to the surface of the bilayer is not sufficient to observe ion channel activity).
Bilayer lipid membranes from falling droplets
1605
Conclusions In summary, a simple and reliable BLM formation method is described. It is easy to implement, requiring only dispensing of a single droplet. Its simplicity means that it could be easily incorporated into a LOC multi-site analysis platform. The suitability of the system for performing a variety of biological assays has been demonstrated, including fine adjustment of buffer properties, delivery of proteins, and electrical measurements. Acknowledgments This work was supported by the 6th Framework Programme of the European Commission, under the contract NMP4CT-2005-017114 “RECEPTRONICS”, and the UK Interdisciplinary Research Centre in Bio-Nanotechnology (R45659/01). KcsA was a kind gift of Prof. A. Lee from the School of Biological Sciences, University of Southampton.
7. 8.
9. 10. 11. 12. 13. 14. 15.
References
16. 17.
1. Tien HT (1974) Bilayer lipid membranes (BLM). Theory and practice. Dekker, New York 2. Montal M, Mueller P (1972) Proc Nat Acad Sci USA 69:3561– 3566 3. Cooper MA (2004) J Mol Recognit 17:286–315 4. Castellana ET, Cremer PS (2006) Surf Sci Rep 61:429–444 5. Tien HT, Ottova AL (1998) Electrochim Acta 43:3587–3610 6. Mischak H, Apweiler R, Banks RE, Conaway M, Coon J, Dominiczak A, Ehrich JH, Fliser D, Girolami M, Hermjakob H, Hochstrasser D, Jankowski J, Julian BA, Kolch W, Massy ZA,
18. 19.
20. 21.
Neusuess C, Novak J, Peter K, Rossing K, Schanstra J, Semmes OJ, Theodorescu D, Thongboonkerd V, Weissinger EM, Van Eyk JE, Yamamoto T (2007) Proteomics Clin Appl 1:148–156 Vitzthum F, Behrens F, Anderson NL, Shaw JH (2005) J Prot Res 4:1086–1097 Andersson M, Okeyo G, Wilson D, Keizer H, Moe P, Blount P, Fine P, Dodabalapur A, Duran RS (2008) Biosens Bioelectron 23:919–923 Le Pioufle B, Suzuki H, Tabata KV, Noji H, Takeuchi S (2008) Anal Chem 80:328–332 Maurer JA, White VE, Dougherty DA, Nadeau JL (2007) Biosens Bioelectron 22:2577–2584 Zagnoni M, Sandison M, Marius P, Lee AG, Morgan H (2007) Lab Chip 7:1176–1183 Malmstadt N, Nash MA, Purnell RF, Schmidt JJ (2006) Nano Letters 6:1961–1965 Schmitt EK, Vrouenraets M, Steinem C (2006) Biophys J 91:2163–2171 White RJ, Ervin EN, Yang T, Chen X, Daniel S, Cremer PS, White HS (2007) J Am Chem Soc 129:11766–11775 Sandison ME, Zagnoni M, Morgan H (2007) Langmuir 23:8277– 8284 Holden MA, Needham D, Bayley H (2007) J Am Chem Soc 129:8650–8655 Funakoshi K, Suzuki H, Takeuchi S (2006) Anal Chem 78:8169– 8174 Heron AJ, Thompson JR, Mason AE, Wallace MI (2007) J Am Chem Soc 129:16042–16047 Poulos JL, Tae-Joon Jeon, Damoiseaux R, Gillespie EJ, Bradley KA, Schmidt JJ (2008) Biosens Bioelectron. doi:10.1016/j. bios.2008.08.041 Marius P, Alvis SJ, East JM, Lee AG (2005) Biophys J 89:4081– 4089 Andersen OS, Koeppe RE, Roux B (2005) IEEE Trans Nanobiosci 4:10–20