ANALYTICAL SCIENCES JULY 2012, VOL. 28 661

Introduction

The development of lipid bilayer-based sensors by using pore-forming compounds, such as gramicidin and α-hemolysin, has attracted increasing interest.1–6 Not only free-standing bilayers, but also supported lipid bilayers, have been exploited, in which peptide channels are embedded as molecular transducers of electric signals. The lipid bilayers containing pore-forming compounds have the potential of designing highly sensitive molecular sensing interfaces based on the principle of ion channels.7

Gramicidin A is a pentadecapeptide that forms highly selective cation channels. The gating is known to occur by the formation and dissociation of a transmembrane dimer. The total length of the gramicidin dimer is about 2.5 to 3.5 nm.8–10 The formation of the gramicidin dimer is associated with a local stress of the bilayer structure caused by the thickness of hydrocarbon region of a PC bilayer.11 The kinetics of dimer formation is modulated by the membrane thickness,10,12–19 the charges of the lipid head groups,20,21 the chemical structures of the head groups,22 the  structure of hydrophobic tails,15,23 and the cholesterol concentration.24 Such regulation of the gramicidin function by the bilayer physical properties is discussed extensively in a recent review.11 Although these studies clarified the fundamental properties of single gramicidin channel, most studies have focused on the properties of ion-channel function specific for monovalent cations, rather than the design of molecular sensing interfaces.

There are two potential approaches for the design of biosensors based on gramicidin channels in lipid bilayers. One approach is based on the direct, chemical modification of gramicidin channels with receptors.25–28 The binding of ions or molecules

(analytes) to receptor-tagged gramicidin modulates the kinetic of dimer formation in either free-standing or supported lipid bilayers. The modulation can be detected as changes in electric signals, such as admittance and ion currents. Another approach is that lipid bilayers are designed to cause a local stress of the bilayer structure upon the selective binding of an analyte to receptor sites of lipids rather than gramicidin. The analyte induces changes in the local stress of a lipid bilayer, leading to modulation of the kinetics of dimer formation.5,29,30 We have demonstrated that channel events with the short lifetime of gramicidin in a receptor-bound lipid bilayer increases with the concentration of an analyte.29,30 This phenomenon has been proposed as a useful one for the detection of biomolecules, such as avidin and antibody. The membrane-bound receptor approach has the advantage that versatile receptors can be used to design bilayer interfaces. In addition, the number of receptor sites is independent from that of gramicidin channels, allowing the recording of a sensor signal based on single gramicidin channel without the deteriorating the dynamic range and sensitivity.

In the present study, considering the importance of controlling the structures and properties of lipid bilayers for the design of molecular sensing interfaces based on the membrane bound receptor approach, we performed recordings of channel currents of gramicidin A in free-standing lipid bilayers with different alkyl chain lengths, head group charges and kinds of head groups. We analyzed the recorded single channel currents in terms of the conductance and distribution of the lifetime. Based on these analyses, a lipid bilayer sensor based on gramicidin A is described, in which an integrated current rather than the frequency of short-lifetime events is used as an analytical signal. The integrated current is a very easily accessible, amplified analytical signal.

2012 © The Japan Society for Analytical Chemistry

† To whom correspondence should be addressed.E-mail: sugawara@chs.nihon-u.ac.jp

Planar Lipid Bilayers Containing Gramicidin A as a Molecular Sensing System Based on an Integrated Current

Masato NISHIO, Atsushi SHOJI, and Masao SUGAWARA†

Department of Chemistry, College of Humanities and Sciences, Nihon University, Sakurajyosui, Setagaya, Tokyo 156–8550, Japan

The channel activity of gramicidin A in free-standing planar lipid bilayers with different charges of polar head groups and various lengths of hydrocarbon tails were analyzed in terms of the channel conductance, the lifetime of channel events and the magnitude of integrated currents. The channel activity of gramicidin A in lipid bilayers is tunable by adjusting the membrane composition. The in situ coupling of the anti-BSA antibody as a model protein to the amine moiety of phosphatidylethanolamine (PE) in a lipid bilayer by the amine coupling method allowed us to design an antigen (BSA)-sensitive interface, in which the integrated current, rather than the frequency of channel event, can be used as an analytical signal. The potential of the present system for highly sensitive and selective detection of BSA at 10–9 g/mL level is demonstrated.

(Received April 23, 2012; Accepted May 23, 2012; Published July 10, 2012)

662 ANALYTICAL SCIENCES JULY 2012, VOL. 28

Experimental

ReagentsGramicidin A was obtained from Sigma-Aldrich Chemical

Co. (St. Louis, MO). 1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC(C18:1), 50 mg/mL chloroform solution), 1-2-dipalmitoleoyl-sn-glycero-3-phosphocholine (DPPC(C16:1), 10 mg/mL chloroform solution), 1,2-dimyristoleoyl-sn-glycero-3-phosphocholine (DMPC(C14:1), 10 mg/mL chloroform solution), 1,2-dioleoyl-sn-glycero-3-(phospho-L-serine) (sodium salt) (DOPS(C18:1), 10 mg/mL chloroform solution), 1,2-dioleoyl-sn-glycero-3-ethylphosphocholine (chloride salt) (EPC(C18:1), 10 mg/mL chloroform solution), 1,2-dioleoyl-sn-glycero-3-phospho-ethanolamine (DOPE(C18:1), 10 mg/mL chloroform solution) and 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC(C18:0), 10 mg/mL chloroform solution) were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL). Cholesterol (Chol) was obtained from Wako Pure Chemicals Co. (Osaka, Japan) and recrystallized three times from methanol. 2-[4-(2-Hydroxyethyl)-1-piperazinyl]-ethanesulfonic acid (HEPES) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) were obtained from Dojindo Laboratories (Kumamoto, Japan). N-Hydroxysuccinimide (NHS), chloroform, n-hexane, and methanol were from Wako Chemicals Co. Sheep polyclonal anti-BSA antibody (anti-BSA) was obtained from Bethl Lab. (TX, USA). Albumin from bovine serum (BSA, >97%) was from Sigma-Aldrich Chemical Co. (St. Louis, MO). Other reagents used were all of analytical reagent grade. A Milli-Q water (Millipore Reagent Water System, Bedford, MA) was used throughout the experiments.

Formation of planar bilayer lipid membranesPlanar lipid bilayer membranes (BLMs) were formed at an

aperture (diameter 30 – 70 μm) of a Teflon film by a monolayer folding method.31,32 A lipid solution of phospholipid and Chol in a weight ratio of 4:1 in a mixture of chloroform and hexane (volume ration 1:1) was used for bilayer formation, unless otherwise noted. The chamber solution (1.4 mL) was a 0.15 M KCl solution containing 10 mM HEPES/KOH buffer (pH 7.4) (abbreviated as a KCl solution). For convenience, two aqueous solutions separated by a Teflon film are called cis (ground side) and trans (applied potential side) solutions, respectively.

Channel current recordingsGramicidin A was dissolved in methanol, and then diluted

with a KCl solution to give a 20 ng/mL solution. After a lipid bilayer was formed in a KCl solution, a small portion (5 – 10 μL) of a 20 ng/mL gramicidin A solution was added to both cis and trans solutions, unless otherwise noted. Both solutions were usually stirred for 1 – 2 min until the spontaneous fusion of gramicidin A, which can be known by the appearance of single-channel currents at +100 mV or –100 mV. The apparatus used for current recordings and data analyses were the same as described in our previous paper.29 All recordings were carried out at room temperature (20 – 25°C). The single-channel conductance was analyzed with a pClamp 8.1 software Fechan (Axon Instruments Inc., Burlingame, CA) and the lifetime was analyzed with Lab chart 7.2 Japanese (AD Instruments Inc.).

Procedure for in situ modification of receptorThe modification of a receptor at a bilayer interface was

performed as follows. First, gramicidin A was fused to a DOPE(C18:1)/DPPC(C16:1)/Chol bilayer from a trans solution. After the channel activity of gramicidin A was observed,

a  25-μL aliquot of a KCl solution containing anti-BSA (10–2 mg/mL) was add to the cis solution (final concentration 1 nM). Then, a mixture (50 μL) of 0.15 M NHS and 0.06 M EDC in a KCl solution was added to the cis solution in order to couple anti-BSA to the amine moiety of DOPE. After incubation for 15 min, the cis solution was exchanged for a KCl solution to remove excess receptor and coupling agents. The exchange was carried out by synchronized aspiration and injection through Teflon tubes set in the cis side of the chamber, the end of which was connected to a peristatic pump (Advantec AP-1600, Tokyo, Japan). Then, a 12.5-μL portion of anti-BSA (10–2 mg/mL) in a KCl solution was added again to the cis solution (final concentration 0.5 nM). It necessitated approximately 10 min for exchanging a cis solution (1.475 mL) for a KCl solution at a flow rate of 0.5 mL/min. Single channel currents were recorded at –100 mV or +100 mV (trans side) and at room temperature. The very magnitudes of the channel currents were not affected by the sign of the applied potential.

Impedance analysisElectrochemical impedance spectroscopy was performed in a

4-electrode system with a high-resolution impedance spectrometer (INPHAZE, Sydney, Austria). The cell used for impedance measurements is schematically shown in Fig. S1 (Supporting Information). The electrochemical cell contained four Ag-AgCl electrodes. A 10-mV amplitude sine-wave signal perturbation was applied in the 0.1 – 106 Hz frequency range, as shown in Fig. S2 (Supporting Information). All experiments were performed at room temperature. The low-frequency capacitance (1 Hz) is related to the thickness, d, of the bilayer by

dC

= ε ε0 1

m, (1)

where ε1 is the dielectric constant of the alkyl chains, which is assumed to be 2.1,33 ε0 is the permittivity of free space (8.85 × 10–12) and Cm the capacitance per unit area. The bilayer area was assumed to be equal to the aperture area of a Teflon film, because an aperture size of approximately 300 μm in diameter was used for the impedance measurements, and hence the contribution of the border of a lipid bilayer on the membrane area was minimized.

Results and Discussion

Effect of membrane thicknessThe bilayer thickness of the hydrophobic region is known to

vary linearly with the alkyl chain length.12 We have reported that the number of channel events with short lifetime increase linearly with an increase in the length of alkyl chains of phosphatidylcholine.32 Here, we assesses the lifetime of gramicidin A in lipid bilayers with different alkyl chain lengths. Figure 1 shows a single-channel recording for gramicidin A in lipid bilayers formed with phosphatidylcholine of varying alkyl chain lengths. The thickness of the hydrophobic region of the bilayer determined from the impedance analysis is given in Table 1. The lifetime of a gramicidin A dimer was markedly dependent on the length of alkyl chains, as given in Table 2. The lifetime varied from 180 ms in a DOPC(C18:1)/Chol (4:1) bilayer to 20 min in a DMPC(C14:1)/Chol (4:1) bilayer. The hydrophobic length of the gramicidin A dimer is known to be approximately 2.2 nm.10 Therefore, the membrane thicknesses (hydrophobic thickness) of a DOPC(C18:1)/Chol bilayer (2.7 nm)

ANALYTICAL SCIENCES JULY 2012, VOL. 28 663

and a DPPC(C16:1)/Chol (4:1) bilayer (2.3 nm) are longer than the dimer length. Such a local stress of the bilayer structure may lead to unstable dimer formation, leading to a short lifetime of the dimer. By contrast, in a DMPC(C14:1)/Chol bilayer, the lifetime of gramicidin A was very long, probably because the hydrophobic thickness was close to the length of the gramicidin A dimer.

In a DOPE(C18:1)/Chol (4:1) bilayer, the channel events with a short lifetime markedly increased (Table 2). A lifetime shorter than 100 ms occupied approximately 80% of the total opening events, while in a DOPC(C18:1)/Chol bilayer, the short lifetime events occupied only 36%. This difference can be explained by the thickness of a DOPE bilayer, which is larger than that of a

DOPC bilayer, owing to a much closer packing of PE molecules.34 In a DOPC(C18:1)/DSPC(C18:0)/Chol (2:2:1) bilayer, the fraction of channel events with short lifetime significantly increased as compared with the case of a DOPC(C18:1)/Chol bilayer (Fig. S3, Supporting Information), suggesting that the formation of gramicidin A dimer was affected by the presence of unsaturated acyl chains.

Effect of membrane surface chargesRecordings of channel activity of gramicidin A in lipid

bilayers with different charges of polar head groups are shown in Fig. 2. The channel conductance and lifetime are given in Table 2. The conductance is much larger in a negatively charged DOPS(C18:1)/Chol (4:1 in weight ratio) bilayer and much smaller in a positively charged EPC(C18:1)/Chol (4:1) bilayer as compared with that in a neutral DOPC(C18:1)/Chol (4:1) bilayer. A similar enhancement of the conductance for gramicidin channels has been reported in lipid membranes of diphytanoyl phophatidylserine (PS).20,21 On the other hand, the membrane surface charges did not affect the lifetime of the channel events, except for the DOPS bilayer, where short-lifetime events appeared to a lesser extent.

The suppression of conductance, and hence suppression of the entry of cations (K+) into the channel, is due to the fixed positive charges of an EPC(C18:1)/Chol bilayer, which leads to a higher charge density near the cation-permeable channel mouth than in the case of a planar surface.35 A similar explanation is adaptable

Fig. 1　Effect of the alkyl chain lengths of the lipid on the lifetime of a single gramicidin A channel in a lipid bilayer. A lipid bilayer made of lipid and cholesterol in a weight ratio of 4:1 was formed at an aperture (diameter 30 – 70 μm) of a Teflon film by the monolayer folding method. The chamber solution was a KCl solution. (a) DOPC(C18:1)/Chol bilayer, (b) DPPC(C16:1)/Chol bilayer, and (c) DMPC(C14:1)/Chol bilayer. The channel currents were recorded at +100 mV (trans side).

Table 1　Membrane hydrophobic thickness and specific capacitance of lipid bilayers obtained by electric impedance measurements

Lipid bilayerSpecific

capacitancea/μF cm–2

Hydrophobic thicknessb/

nm

Number of measurements

DOPC(C18:1)c

DPPC(C16:1)c

DMPC(C14:1)c

DPPC(C16:1)/DOPE(C18:1)e

0.70 ± 0.010.81 ± 0.04

n.m.d

0.86 ± 0.02

2.7 ± 0.22.3 ± 0.3

—2.2 ± 0.1

33

—3

a. The membrane area was assumed to be equal to the geometric area of an aperture (diameter 300 μm) of a Teflon film.b. Hydrophobic thickness of carbon chain regions calculated from specific capacitance at 1 Hz using dielectric constants of ε0 = 8.85 × 10–12 and ε1 = 2.1.c. Lipid bilayers were composed of each lipid and cholesterol in a weight ratio of 4:1.d. Not measured because of an unstable lipid bilayer.e. A mixed lipid bilayer of 3:1 in weight ratio.

Table 2　Effect of the chain lengths and membrane surface charges on the mean conductance, mean lifetime and fraction of lifetime

Lipid bilayera Conductance/pS Mean life timebFraction of lifetimec, %

Number of eventsd

≤ 100 ms ≥ 100 ms

DOPC(C18:1)DPPC(C16:1)DMPC(C14:1)DOPS(C18:1)EPC(C18:1)DOPE(C18:1)DOPC(C18:1)/DSPC(C18:0)f

151516402.4

1816

183 ± 145 ms6.0 ± 5.8 s

20 min455 ± 402 ms 157 ± 153 ms64 ± 54 ms61 ± 48 ms

36n.me

n.me

15488281

64n.me

n.me

85521819

13161

9216671

180

a. Lipid bilayers were composed of each lipid and Chol in a weight ratio of 4:1. b. The lifetime averaged for total open events. c. The fraction of lifetime over the total open events. d. Number of open events observed during 10 min recording. The recording time for DMPC was 40 min. e. Not measured. f. Lipid bilayer composed of DOPC(C18:1)/ DSPC(C18:0)/Chol of 2:2:1 in a weight ratio.

664 ANALYTICAL SCIENCES JULY 2012, VOL. 28

to larger conductance in a DOPS(C18:1)/Chol bilayer, where the negative surface charge plays a role in controlling the entry of K+ ions.

Effect of membrane compositionConsidering that an integrated current is used as an analytical

signal, the long lifetime of gramicidin A leads to the high sensitivity of a lipid bilayer sensor. However, when the lifetime of gramicidin A is very long, as observed in a DMPC(C14:1)/Chol bilayer, one needs to record its response for long time. In a DOPS(C18:1)/Chol bilayer, negative surface charges may induce non-specific interactions between the membrane surface and an analyte. Therefore, DPPC(C16:1)-based bilayers appeared to be suitable for a lipid bilayer sensor.

Comparisons of the channel activity of gramicidin A in lipid bilayers with different composition are shown in Fig. 3. In a DPPC(C16:1)/DOPE(C18:1)/Chol (3:1:1) bilayer, open events with lifetimes longer than 1 s occupied approximately 60% of the total events, while the events shorter than 1 s were approximately 40% (Fig. 3c). On the other hand, in a DPPC(C16:1)/DOPE(C18:1)/Chol (1:3:1) bilayer, events longer than 1 s rarely appeared (Fig. 3d). Thus, events having a long lifetime predominantly appeared in the DPPC-rich bilayer, and the events with a short lifetime were predominant in the DOPE-rich membrane, showing that the channel activity of gramicidin A reflected the fraction of DPPC(C16:1) or DOPE(C18:1) in the mixed bilayer.

The distribution of the conductance of gramicidin A dimer in a DPPC(C16:1)/DOPE(C18:1)/Chol (1:3:1) bilayer (Fig. 3f) shows that there exist two predominant conductance levels, i.e.,

17 – 18  and 14 – 15 pS. On the other hand, the conductance levels of 14 – 15 and 12 – 13 pS were observed in a DPPC(C16:1)/DOPE(C18:1)/Chol (3:1:1) bilayer (Fig. 3e). The appearance of the two conductance levels in the DOPE-rich bilayer implies that the bilayer is not homogenous. The presence of lipid rafts in vesicles of a DOPC-DPPC-Chol system has been reproted by using fluorometric36,37 and small-angle neutron scattering techniques.38 We suppose that there exist lipid rafts in the DPPC-DOPE bilayer. The conductance level of 17 – 18 pS may correspond to the activity of gramicidin A dimer, which stays in a DOPE raft, while that of 14 – 15 pS corresponds to that of the dimer in a DPPC raft.

The conductance level of 17 – 18 pS was not observed in a DPPC(C16:1)/DOPE(C18:1)/Chol (3:1:1) bilayer. This indicates that gramicidin A dimer stays in a DPPC raft rather than staying evenly in both rafts, because the gramicidin A dimer is more stable in a DPPC bilayer than in a DOPE bilayer (vide supra).

Coupling of receptor to a lipid bilayerThe results obtained above show that the channel activity of

gramicidin A is tunable by controlling the membrane composition. The introduction of receptor sites to a lipid bilayer, together with the moderately long lifetime of channel

Fig. 2　Effect of membrane surface charges on the activity of single gramicidin A in a lipid bilayer. A lipid bilayer made of lipid and Chol  in a weight ratio of 4:1 was formed at an aperture (diameter 30 – 70 μm) of a Teflon film by the monolayer folding method. The chamber solution was a KCl solution. (a) EPC(C18:1)/Chol bilayer, (b) DOPC(C18:1)/Chol bilayer, and (c) DOPS(C18:1)/Chol bilayer. The channel currents were recorded at +100 mV (trans side).

Fig. 3　Effect of the membrane composition on the lifetime and conductance of single gramicidin A channel in lipid bilayers. (a), (c), (e): The lipid bilayers made of DPPC(C16:1), DOPE(C18:1) and Chol in a weight ratio of 3:1:1. (b), (d), (f): The lipid bilayers made of DPPC(C16:1), DOPE(C18:1) and Chol in weight ratio of 1:3:1. The bilayers were formed at an aperture (diameter 30 – 70 μm) of a Teflon film by the monolayer folding method. The chamber solution was a KCl solution. The channel currents were recorded at –100 mV (trans side). (a) and (b), channel current recordings; (c) and (d), histograms of the lifetime distribution; (e) and (f), histograms of channel conductance distribution.

ANALYTICAL SCIENCES JULY 2012, VOL. 28 665

activity, is necesary for designing a selective sensing interface based on an integrated current. For this purpose, lipid bilayers of DPPC(C16:1), DOPE(C18:1) and Chol in a weight ratio of 3:1:1 were used. Since the thickness of a DOPE(C18:1) raft is expected to be larger than that of a DPPC(C16:1) raft, the receptor (anti-BSA) appears to be more easily accessible to the amine moiety of DOPE(C18:1). In addition, the channel events of gramicidin A in this bilayer exhibit a moderately long lifetime (vide supra).

The chemical modification of anti-BSA (MW 150 kDa, pI = 8.4)39 as a receptor protein to the amine moiety of DOPE(C18:1) in a mixed lipid bilayer was performed by the amine coupling method. First, gramicidin A as a signal transduction element was incorporated with the bilayer from a trans side. Then, anti-BSA was coupled to DOPE(C18:1) from a cis side, followed by solution exchange for the removal of excess anti-BSA and coupling agents. The introduction of anti-BSA was monitored by measuring an integrated current in a unit time (C/s), i.e., an integrated channel current divided by the recording time. The integrated current increased with an increase in the anti-BSA concentration and reached the maximum at a concentration above 0.5 nM (Fig. 4a). For the response accompanied with an increase in the conductance (Fig. 4b), however, the lifetime was not much affected by the anti-BSA coupling (Fig. 4c). Therefore, the increase in the integrated current by the coupling of anti-BSA was due to the appearance of channel events with larger conductance. The coupling of anti-BSA may result in the

modulation of lipid rafts in the DPPC(C16:1)-DOPE(C18:1) bilayer, thereby the fraction of gramicidin A dimmer that stays in a DOPE(C18:1) raft increases and/or the distribution of potassium ions at the membrane surface is modulated because of the bulkiness of the receptor, leading to an increase in the conductance.

Concentration dependenceThe response of the anti-BSA-modified bilayer containing

gramicidin A to BSA (pI = 4.8)40 was investigated. In the presence of 0.5 nM anti-BSA in the cis solution, a small aliquot (10 μL) of a BSA solution at various concentrations was added. Figure 5a shows the integrated current vs. BSA concentration plot at –100 mV. The magnitude of the integrated current was normalized to that in the absence of BSA, because the very magnitude of the integrated current in the absence of BSA varied slightly. The normalized integrated current increased with an increase in the BSA concentration, ranging from 1.0 × 10–9 to 1.0 × 10–7 g/mL (curve 1). Such an increase in the response was not observed with a receptor-free lipid bilayer (curve 2), and also for avidin with a BSA-modified lipid bilayer (curve 3). The BSA-dependent increase in the response with an anti-BSA-modified lipid bilayer is ascribable to an increase in the channel conductance caused by the binding of BSA to the anti-BSA sites at the bilayer interface (Fig. S4, Supporting Information). It is noted that in the absence of anti-BSA in the cis solution, the response of the bilayer to BSA shifted to its

Fig. 4　Effect of chemical modification of anti-BSA on the channel activity of gramicidin A dimer. (a) Effect of the anti-BSA concentration on the integrated current, (b) the distribution of conductance and (c) the distribution of lifetime. The lipid bilayers made of DPPC(C16:1), DOPE(C18:1) and Chol in a weight ratio of 3:1:1were formed at an aperture (diameter 30 – 70 μm) of a Teflon film by the monolayer folding method. The chamber solution was a KCl solution. The channel currents were recorded at –100 mV (trans side).

666 ANALYTICAL SCIENCES JULY 2012, VOL. 28

higher concentration range, i.e., 10–7 to 10–5 g/mL (Fig. 5b). This suggests that the formation of a polymeric complex between BSA and anti-BSA is responsible for the response of the present system. The increase in the channel conductance is seemingly because of the modulation of the raft structure and the distribution of K+ ions at the membrane surface (vide supra).

Since the number of channel events that are observed during current recordings increase with the length of the recording period, the magnitude of the integrated current depends on the recording time. However, when the integrated current was divided by its recording time, the slopes of the plots were independent on the recording time (Fig. 5b). This suggests that the channel events appeared almost evenly during the recording time up to 20 min. Actually, the number of channel events observed with an anti-BSA-modified DPPC(C16:1)/DOPE(C18:1)/Chol (3:1:1) bilayer in the presence of 1.0 × 10–5 g/mL BSA was 32 for 5 min recording and 33 for next 5 min recording, followed by 74 events for the successive 10 min recording. Consequently, the integrated current divided by its recording

time was suitable as an analytical signal for the quantification of BSA.

Conclusion

The results obtained above demonstrate that the response of a lipid bilayer containing a gramicidin A channel is tunable by adjusting the membrane composition. An appropriate choice of the membrane composition allowed us to couple the receptor (antibody) to a lipid bilayer, and to use an integrated current as an analytical signal. Since the design of a molecular-sensitive interface is often hindered by the difficulty of fusing the receptor with lipid bilayers, the present in situ coupling approach has the advantage that the spontaneous fusion of the receptor is not necessary. As compared with our previous study, where the frequency of channel events was an analytical signal, the present system based on an integrated current is more sensitive, because of the amplified feature of an ion-flux through the channel. The combined use of the in situ receptor coupling system and modernized bilayer formation techniques may lead to the development of a practical bilayer-based sensor.

Acknowledgements

The authors greatly acknowledge Miss S. Endo for her help in electric impedance measurements. This work was financially supported by a Grant-in-Aid for Scientific and Research (B) (No. 21350045) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan. Authors also thank Bioresearch Center Co. (Tokyo, Japan) for lending us a high resolution impedance spectrometer.

Supporting Information

This material is available free of charge on the Web at http://www.jsac.or.jp/analsci/.

References

1. A. Liu, Q. Zhao, and X. Guan, Anal. Chim. Acta, 2010, 675, 106.

2. S. Majd, E. C. Yusko, Y. N. Billeh, M. X. Macrae, and J. Yang, M. Mayer, Curr. Opin. Biotechnol., 2010, 21, 439.

3. L.-Q. Gu and J. W. Shim, Analyst, 2010, 135, 441. 4. L. Movileanu, Trends Biotechnol., 2009, 27, 333. 5. A. Hirano-Iwata, M. Niwano, and M. Sugawara, Trends

Anal. Chem., 2008, 27, 512. 6. E. T. Castellana and P. S. Cremer, Surf. Sci. Rep., 2006, 61,

429. 7. M. Sugawara, A. Hirano, P. Bühlmann, and Y. Umezawa,

Bull. Chem. Soc. Jpn., 2002, 75, 187. 8. D. W. Urry, Proc. Natl. Acad. Sci. U. S. A., 1971, 68, 672. 9. E. Bamberg and P. Läuger, J. Membr. Biol., 1973, 11, 177.10. J. R. Elliott, D. Needham, J. P. Dilger, and D. A. Haydon,

Biochim. Biophys. Acta, 1983, 735, 95.11. J. A. Lundbæk, S. A. Collingwood, H. I. Ingo’lfsson, R.

Kapoor, and O. S. Andersen, J. R. Soc. Interface, 2010, 7, 373.

12. B. A. Lewis and D. M. Engelman, J. Mol. Biol., 1983, 166, 211.

13. V. S. Rudnev, L. N. Ermishkin, L. A. Fonina, and Yu. G.

Fig. 5　Concentration dependence for BSA with an anti-BSA-modified lipid bilayer containing gramicidin A, made of DPPC(C16:1), DOPE(C18:1) and Chol bilayer in a weight ratio of 3:1:1. The applied potential was –100 mV (trans side). The cis side solution contained 0.5 nM anti-BSA. (a) An integrated current vs. BSA concentration plot: (1) BSA, with an anti-BSA-modified BLM (n = 3 except for 10–7 g/mL where n = 2), (2) BSA, with anti-BSA-unmodified BLM (n = 2), (3) avidin, with an anti-BSA-modified BLM (n = 3). (b) Relationship between time-resolved integrated currents and BSA concentration. Recording period: (1) 0 – 5, (2) 0 – 10, and (3) 0 – 20 min.

ANALYTICAL SCIENCES JULY 2012, VOL. 28 667

Rovin, Biochim. Biophys. Acta, 1981, 642, 196.14. H.-A. Kolb and E. Bamberg, Biochim. Biophys. Acta, 1977,

464, 127.15. J. A. Lundbæk, P. Birn, J. Girshman, A. J. Hansen, and O.

S. Andersen, Biochemistry, 1996, 35, 3825.16. N. Mobashery, C. Nielsen, and O. S. Andersen, FEBS Lett.,

1997, 412, 15.17. T. A. Harroun, W. T. Heller, T. M. Weiss, L. Yang, and H.

W. Huang, Biophys. J., 1999, 76, 937.18. W. Rawicz, K. C. Olbrich, T. Mclntosh, D. Needham, and

E. Evans, Biophys. J., 2000, 79, 328.19. B. Martinac and O. P. Hamill, Proc. Natl. Acad. Sci., 2002,

99, 4308.20. H.-J. Apell, E. Bamberg, and P. Läuger, Biochim. Biophys.

Acta, 1979, 552, 369.21. T. K. Rostovtseva, V. M. Aguilella, I. Vodyanoy, S. M.

Berzrukov, and V. A. Parsegian, Biophys. J., 1998, 75, 1783.

22. E. Neher and H. Eibl, Biochim. Biophys. Acta, 1977, 464, 37.

23. J. Girshman, D. V. Greathouse, R. E. Koepper, II, and O. S. Andersen, Biophys. J., 1997, 73, 1310.

24. M. Weinrich, T. K. Rostovtseva, and S. M. Bezrukov, Biochemstry, 2009, 48, 5501.

25. B. A. Cornell, V. L. B. Braach-Maksvytis, L. G. King, P. D. J. Osman, B. Raguse, L. Wieczorek, and R. J. Pace, Nature, 1997, 387, 580.

26. S. W. Lucas and M. M. Harding, Anal. Biochem., 2000, 282, 70.

27. S. Futaki, Y. Zhang, T. Kiwada, I. Nakase, T. Yagami, S. Oiki, and Y. Sugiura, Bioorg. Med. Chem., 2004, 12, 1343.

28. M. X. Macrae, S. Blake, X. Jiang, R. Capone, D. J. Estes, M. Mayer, and J. Yang, ACS Nano, 2009, 3, 3567.

29. A. Hirano, M. Wakabayashi, Y. Matsuno, and M. Sugawara, Biosens. Bioelectron., 2003, 18, 973.

30. Y. Matsuno, C. Osono, A. Hirano, and M. Sugawara, Anal. Sci., 2004, 20, 1217.

31. M. Montal and P. Müller, Proc. Natl. Acad. Sci. U. S. A., 1972, 69, 3561.

32. M. Sugawara and A. Hirano, “Advances in Planar Lipid Bilayers and Liposomes”, ed. H. T. Tien and A. Ottova, 2005, Vol. 1, Chap. 8, Elsevier, Amsterdam.

33. P. Läuger, W. Lesslauer, E. Marti, and J. Richter, Biochim. Biophys. Acta, 1967, 135, 20.

34. T. K. Rostovtseva, H. I. Petrache, N. Kazemi, E. Hassanzadeh, and S. M. Bezrukov, Biophys. J., 2007, L23.

35. S. B. Hladky and D. A. Haydon, Biochim. Biophys. Acta, 1972, 274, 294.

36. A. C. Brown, K. B. Towles, and S. P. Wrenn, Langmuir, 2007, 23, 11180.

37. A. C. Brown, K. B. Towles, and S. P. Wrenn, Langmuir, 2007, 23, 11188.

38. K. Vogtt, C. Jeworrek, V. M. Garamus, and R. Winter, J. Phys. Chem. B, 2010, 114, 5643.

39. B.-S. Lee, S. Gupta, S. Krisnanchettier, and S. S. Lateef, Anal. Biochem., 2004, 334, 106.

40. G. B. Sigal, M. Mrksich, and G. M. Whitesides, J. Am. Chem. Soc., 1998, 120, 3464.