Incorporation of α-Hemolysin in Different Tethered Bilayer Lipid

496

Langmuir 2008, 24, 496-502

Incorporation of r-Hemolysin in Different Tethered Bilayer Lipid Membrane Architectures Inga K. Vockenroth,†,‡ Petia P. Atanasova,† A. Toby A. Jenkins,‡ and Ingo Ko¨per*,† Max Planck Institue for Polymer Research, 55128 Mainz, Germany, and UniVersity of Bath, Bath BA27AY, UK ReceiVed October 1, 2007. In Final Form: October 30, 2007 Tethered bilayer lipid membranes are stable solid supported model membrane systems. They can be used to investigate the incorporation and function of membrane proteins. In order to study ion translocation mediated via incorporated proteins, insulating membranes are necessary. The architecture of the membrane can have an important effect on both the electrical properties of the lipid bilayer as well as on the possibility to functionally host proteins. R-Hemolysin pores have been functionally incorporated into a tethered bilayer lipid membrane coupled to a gold electrode. The protein incorporation has been monitored optically and electrically and the influence of the molecular structure of the anchor lipids on the insertion properties has been investigated.

Introduction In the pharmaceutical industry, as well as in environmental diagnostics, there exists a growing demand for fast and reliable detection methods at the molecular level. Crucial parameters for the development of a novel device are high detection sensitivity and selectivity. Biological compounds such as receptors in cell membranes are highly selective recognition units at the single molecule level. Binding of a single analyte to a receptor site situated on an ion channel can change the conductivity of the channel and thus lead to a detection event. Therefore, the use of biological entities is a promising approach toward novel biosensing devices. Unfortunately, most biological receptors and channels only work in their natural environment, which is a cell membrane.1 However, these structures are mostly too complex to be used in practical applications, and thus special biomimetic architectures are needed. For measurements of transmembrane currents, black lipid membranes (BLM)2 have been shown to be very useful. However, the mechanical stability of these free-standing membranes is rather poor and the preparation is quite tedious. There have been several attempts to increase the stability of BLMs, but in most cases they are restricted to use in the laboratory setup.3,4 In order to increase the mechanical stability, the deposition of artificial membranes on solid supports has been widely attempted.5-7 These solid-supported BLMs (sBLM) show a higher mechanical stability than BLMs, and the incorporation of several model channels and pores has been reported.1,7,8 However, the restricted fluidity and the space limitation between membrane and substrate do not allow for the functional incorporation of more complex membrane * Corresponding author. E-mail: [email protected]. † Max Planck for Polymer Research. ‡ University of Bath. (1) Purrucker, O.; Hillebrand, H.; Adlkofer, K.; Tanaka, M. Electrochim. Acta 2001, 47, 791-798. (2) Mu¨ller, P.; Rudin, D. O. Nature 1967, 213 (5076), 603-604. (3) Kang, X.; Cheley, S.; Rice-Ficht, A. C.; Bayley, H. J. Am. Chem. Soc. 2007, 129 (15), 4701-4705. (4) Schuster, B.; Sleytr, U. B.; Diederich, A.; Ba¨hr, G.; Winterhalter, M. Eur. Biophys. J. 1999, 28, 583-590. (5) Sackmann, E. Science 1996, 271, 43-48. (6) Steinem, C.; Janshoff, A.; Ulrich, W.-P.; Sieber, M.; Galla, H.-J. Biochim. Biophys. Acta 1996, 1279, 169-180. (7) Vallejo, A. E.; Gervasi, C. A. Bioelectrochemistry 2002, 57, 1-7. (8) Favero, G.; Annibale, A. D.; Campanella, L.; Santucci, R.; Ferri, T. Anal. Chim. Acta 2002, 460, 23-34.

proteins.9 Their extramembrane domains will touch the support and thus lose their functionality due to denaturation, or the proteins will not incorporate at all. Tethered bilayer lipid membranes (tBLM), on the other hand, are lipid bilayers where the proximal layer is coupled to the substrate via a spacer group. Their tethers provide a reservoir beneath the membrane to allow for the incorporation of more complex proteins. Furthermore, they have extremely high electrochemical sealing properties, comparable to those of natural membranes. Therefore, they allow for the electrical monitoring of the incorporation and function of channel proteins within the membrane. The exotoxin R-hemolysin (R-HL) is secreted by Staphylococcus aureus as a water-soluble, monomeric, 293-residue polypeptide and forms heptameric pores in lipid bilayers.10-13 The overall shape of the pore resembles a mushroom, with the stem penetrating the membrane bilayer and the cap extending into the extracellular space.14 In nature, incorporation of R-HL into cell membranes leads to lysis of the cells. The monomers bind to the cell membrane, and pore assembly occurs upon subsequent collision during lateral diffusion in the bilayer.15-17 Due to the well-characterized structure of the heptamer14 and the possibility to self-assemble in the absence of cellular machinery, R-HL is a prototype for the study of the assembly of transmembrane proteins.13 The monomers are easy to manipulate. Attempts to engineer this pore and to use it as a detector for metal ions, organic compounds, proteins, and DNA have been published.18-20 The main aspect in this approach is to place a (9) Glazier, S. A.; Vanderah, D. J.; Plant, A. L.; Bayley, H.; Valincius, G.; Kasianowicz, J. J. Langmuir 2000, 16, 10428-10435. (10) Gray, G. S.; Kehoe, M. Infect. Immun. 1984, 46 (2), 615-618. (11) Bhakdi, S.; Tranum-Jensen, J. Microbiol. ReV. 1991, 55 (4), 733-751. (12) Gouaux, J. E.; Braha, O.; Hobaugh, M.; Song, L.; Cheley, S.; Shustak, C.; Bayley, H. Proc. Natl. Acad. Sci. U.S.A. 1994, 91 (2), 12828-12831. (13) Bhakdi, S.; Bayley, H.; Valeva, A.; Walev, I.; Walker, B.; Weller, U.; Kehoe, M.; Palmer, M. Arch. Microbiol. 1996, 165, 73-79. (14) Song, L.; Hobaugh, M.; Shustak, C.; Cheley, S.; Bayley, H.; Gouaux, J. E. Science 1996, 274, 1859-1866. (15) Reichwein, J.; Hugo, F.; Roth, M.; Sinner, A.; Bhakdi, S. Infect. Immun. 1987, 55 (12), 2940-2944. (16) Walker, B.; Krishnasastry, M.; Zorn, L.; Bayley, H. J. Biol. Chem. 1992, 267 (30), 21782-21786. (17) Walker, B.; Braha, O.; Cheley, S.; Bayley, H. Chem. Biol. 1995, 2, 99105. (18) Bayley, H. Sci. Am. 1997, 42-47. (19) Bayley, H.; Cremer, P. S. Nature 2001, 413, 226-230.

10.1021/la7030279 CCC: $40.75 © 2008 American Chemical Society Published on Web 12/18/2007

R-Hemolysin in Tethered Bilayer Architectures

binding site specific for a target analyte into or near the lumen of the pore. The conductivity of the pore will change upon binding of this analyte. The frequency of binding reveals the concentration, whereas the current signature (magnitude and duration of currentblock) encodes for the nature of the analyte. This approach, termed “stochastic sensing”, allows for highly sensitive detection of analytes at the molecular level.18 Present platforms use R-HL in BLMs, where incorporation and measurements of transmembrane currents of R-HL have been reported.18,21-24 The stability and lifetime of BLMs could be enhanced by supporting them with a crystalline S-layer4,25,26 or by using protective layers such as agarose.3 So far, only partial incorporation into sBLMs could be shown; complete reconstitution could not be observed in such thin films,9 mainly due to the small space between the membrane and the support. In a tBLM, the reservoir formed by the spacer group of the anchor lipids facilitates the incorporation. Furthermore, the coupling of the membrane to an electrode allows for electrical characterization. The properties of the spacer region are of extreme importance for the characteristics of the membrane.27,28 We have used two different tBLM architectures for the incorporation of R-HL.29-31 The structure of the spacer region is shown to have a major influence on the incorporation probability. The membrane assembly and pore incorporation could be monitored optically by surface plasmon resonance spectroscopy (SPR) and characterized electrically by electrical impedance spectroscopy (EIS). Materials and Methods Assembly of the System. A tBLM is typically assembled in a two-step procedure. First, lipid monolayers were formed via selfassembly by immersion of an ultraflat gold substrate into a diluted solution of an anchor lipid (0.2 mg/mL).32 The anchor lipids DPhyTL and DPhyHDL were synthesized as described previously.32,33 They consist of a diphytanyl chain that is coupled via a glycerol linker to an oligoethylene oxide spacer. The two lipids have different spacer lengths and anchor groups, as shown in Figure 1. DPhyTL has a spacer consisting of four ethylene oxide units, whereas DPhyHDL has six ethylene oxide units. The molecules were grafted to the gold surface by sulfur-gold bonds via lipoic acid for DPhyTL and via two lipoic acid groups for DPhyHDL. In the second assembly step, the monolayers were completed to bilayers by fusion with freshly prepared small unilamellar DPhyPC (1,2-diphytanoyl-sn-glycero3-phosphocholine, Avanti polar lipids, 2 mg/mL in ultrapure water) vesicles (50 nm by extrusion). In the final step, 20 µL of an R-HL monomer solution (Sigma, Steinheim, 0.5 mg/mL in NaCl/Mops (20) Braha, O.; Gu, L.-Q.; Zhou, L.; Lu, X.; Cheley, S.; Bayley, H. Nat. Biotechnol. 2000, 18, 1005-1007. (21) Korchev, Y. E.; Alder, G. M.; Bakhramov, A.; Bashford, C. L.; Joomun, B. S.; Sviderskaya, E. V.; Usherwood, P. N. R.; Pasternak, C. A. J. Membr. Biol. 1995, 143, 143-151. (22) Bayley, H.; Braha, O.; Gu, L.-Q. AdV. Mater. 2000, 12 (2), 139-142. (23) Belmonte, G.; Cescatti, L.; Ferrari, B.; Nicolussi, T.; Ropele, M.; Menestrina, G. Eur. Biophys. J. 1987, 14, 349-358. (24) Menestrina, G. J. Membr. Biol. 1986, 90, 177-190. (25) Schuster, B.; Sleytr, U. B. Bioelectrochemistry 2002, 55, 5-7. (26) Schuster, B.; Pum, D.; Braha, O.; Bayley, H.; Sleytr, U. B. Biochim. Biophys. Acta 1998, 1370, 208-288. (27) Raguse, B.; Braach-Maksvytis, V.; Cornell, B. A.; King, L. G.; Osman, P. D. J.; Pace, R. J.; Wieczorek, L. Langmuir 1998, 14, 648-659. (28) Krishna, G.; Schulte, J.; Cornell, B. A.; Pace, R. J.; Osman, P. D. Langmuir 2003, 19, 2294-2305. (29) Naumann, R.; Schiller, S. M.; Giess, F.; Grohe, B.; Hartman, K. B.; Ka¨rcher, I.; Ko¨per, I.; Lu¨bben, J.; Vasilev, K.; Knoll, W. Langmuir 2003, 19, 5435-5443. (30) Cornell, B. A.; Braach-Maksvytis, V. L. B.; King, L. G.; Osman, P. D. J.; Raguse, B.; Wieczorek, L.; Pace, R. J. Nature 1997, 387, 580-583. (31) Ko¨per, I. Mol. BioSyst. 2007, 3 (10), 651-657. (32) Schiller, S. M.; Naumann, R.; Lovejoy, K.; Kunz, H.; Knoll, W. Angew. Chem. 2003, 42 (2), 208-211. (33) Atanasov, V.; Atanasova, P. P.; Vockenroth, I. K.; Knorr, N.; Ko¨per, I. Bioconjugate Chem. 2006, 17, 631-637.

Langmuir, Vol. 24, No. 2, 2008 497 buffer) was added to give a final concentration of 180 nM in the measurement cell. The Mops/NaCl buffer consists of 0.1 M NaCl with 0.01 M Mops at pH 7 in ultrapure water (Millipore, Schwalbach, resistance >18.2 MΩcm). All salts were purchased from Acros organics at ACS purity. Surface Plasmon Resonance Spectroscopy (SPR). A custombuilt setup in the Kretschmann configuration with a He/Ne laser (λ ) 632 nm) was used. In the kinetic mode, reflectivity changes occurring at a fixed angle are monitored as a function of time. Angular SPR spectra (cf. Figure 3) were analyzed using a three-layer model including the prism, gold, and thiolipid monolayer. After vesicle fusion, a fourth layer corresponding to the outer leaflet of the bilayer was added to the model. The refractive indices of n ) 1.5 and 1.45 for the mono- and bilayer, respectively, were used.32 R-HL adsorption and incorporation were modeled as a very thin fifth layer with a refractive index of 1.41.34 Electrochemical Impedance Spectroscopy (EIS). Measurements were conducted using an Autolab spectrometer PGSTAT 12 (Eco Chemie, Utrecht, Netherlands). Spectra were recorded for frequencies between 2 mHz and 100 kHz at 0 V bias potential with an ac modulation amplitude of 10 mV. Raw data were analyzed using the ZVIEW software package (Version 2.90, Scribner Associates). Threeelectrode measurements were performed with the substrate as the working electrode, a coiled platinum wire as the counter electrode, and a DRIREF-2 reference electrode (World Precision Instruments, Berlin, Germany). The home-built Teflon cells had a buffer volume of 1 mL and an electrochemically active area on the substrates of about 0.28 cm2. The experimental data has been analyzed using a model equivalent circuit of resistors (R) and capacitors (C).35 The different components can be attributed to the individual parts of the membrane architecture. We used a R(RC)C-circuit consisting of a RC element describing the bilayer in series with a capacitor (CSC) and an electrolyte resistance (Relectrolyte) (cf. inset in Figure 4A). CSC represents the space charge capacitance due to the spacer region combined with the capacitive effects of the electrochemical double layer at the gold interface.29,33 The data can be visualized in Bode plots (cf. Figure 4A,B), where pure capacitances show up as slopes of -1 with high phase shifts of -90° and ideal resistances are represented as horizontal regions of low phase angles.35 An alternative representation is the admittance plot, where the frequency reduced real part of the admittance (Y ) 1/Z) is plotted versus the imaginary part. RC elements can be identified as semicircles. The intersection on the y-axis corresponds to the capacitance, while resistances give a line parallel to the y-axis.

Results and Discussion The assembly of the system consists of distinct steps: formation of the monolayer, completion to a bilayer, and incorporation of the protein. The three steps have been followed using EIS and SPR. Fusion of vesicles with the preformed monolayer leads to an increase in the optical thickness, an increase in resistance, and a decrease in capacitance. In Figure 2, a kinetic experiment using a DPhyHDL monolayer is shown. An increase in membrane resistance could be observed already in the first scan after addition of the vesicles. Simultaneously, the formation of the outer leaflet was monitored by SPR, where an exponential increase in thickness could be seen in the kinetic mode. By analyzing the angular scans in Figure 3, the thickness for the mono- and bilayer was calculated as 3 ( 0.1 and 3 ( 0.2 nm, respectively. The fusion process takes 2-3 h, as judged from the SPR kinetic. It has been observed in other studies that the impedance parameters still improve toward a better sealing of the membrane after reaching the final thickness.36 However, after rinsing, the bilayer is stable, (34) Vo¨ro¨s, J. Biophys. J. 2004, 87, 553-561. (35) Barsoukov, E.; Macdonald, J. R. Impedance SpectroscopysTheory, Experiment, and Applications, 2nd ed.; John Wiley & Sons: New York, 2005. (36) Vockenroth, I. K.; Atanasova, P. P.; Long, J. R.; Jenkins, A. T. A.; Knoll, W.; Ko¨per, I. Biochim. Biophys. Acta, Biomembr. 2007, 1768, 1114-1120.

498 Langmuir, Vol. 24, No. 2, 2008

Vockenroth et al.

Figure 1. Chemical structures of the anchor lipids DPhyTL and DPhyHDL. They consist of a lipid head group, a spacer unit, and an anchor that allows for assembly on gold surfaces. The molecules differ in their spacer length and anchor groups. Names are non-IUPAC.

as judged from the impedance as well as from the SPR data (Figure 2, 130-400 min). Similar experiments have been also performed with DPhyTL monolayers. The electrical parameters of the membranes have been analyzed by EIS (Figure 4) using a simple equivalent circuit. This model does not always describe all experimental details, as many minor processes contribute to the systems response that cannot be discriminated by the analysis. On the other hand, the characteristics of a more complex model, especially when composed from a large number of elements, are rarely unique. Consequently, a simple model that describes the essential characteristics of the spectra is preferred and allows for the systematic investigation of the system. In Table 1, the values for the electrochemical parameters of the monolayers and the final bilayers are given. The experimental data for bilayers built on both monolayer systems show an almost ideal shape, e.g., in the mid-frequency range, the phase shift reaches almost -90°, as for an ideal capacitor. In that region, the impedance shows a slope of -1. The changes in capacitance are more clearly visible in the admittance plots (Figure 4C,D). For DPhyTL, the resistance increased from 9.9 to 19.3 MΩ cm2, while the capacitance decreased slightly from an already low value of 0.74 to 0.69

µF/cm2. The low monolayer capacitance did not allow for a larger decrease in capacitance upon bilayer formation. The monolayer capacitance was already very low and thus highly sealing. These sealing properties could not be much further increased upon formation of the outer leaflet. In the case of DPhyHDL, the scenario looks different. The monolayer had a very low resistance of 0.33 kΩ cm2 that increased during vesicle fusion to 1.1 MΩ cm2. At the same time, the rather large initial capacitance of 3.5 µF/cm2 decreased to a final capacitance of the bilayer of 0.77 µF/cm2. This extremely high decrease in capacitance can be attributed to a very poor sealing of the monolayer that was increased upon formation of an outer layer and filling of the defects in the monolayer. However, the final quality of the bilayer, as judged from the electrical properties, is not as high as for DPhyTL. In the next step, R-HL was introduced to the membranes. R-HL is a toxin that incorporates spontaneously into membranes by the formation of heptameric pores.12,14 Due to the large pore diameter of 1-2 nm, the penetration into the bilayer leads to a drastic decrease in membrane resistance. The adsorption and incorporation of the pore was monitored via SPR (Figure 2), where a slight increase in thickness (or mass on the surface)

R-Hemolysin in Tethered Bilayer Architectures

Langmuir, Vol. 24, No. 2, 2008 499

Figure 2. Kinetic experiment for the formation of a functionalized tBLM. The left axis shows the membrane thickness as measured by SPR illustrating the formation of the outer leaflet of the bilayer on top of a DPhyHDL monolayer (solid line). After 10 min, liposomes were added and the thickness increased to a final value of 3 nm after 2 h. After rinsing the bilayer was stable and addition of R-hemolysin led to a further increase in thickness. On the right axis, the membrane resistance as calculated from subsequent impedance measurements is plotted (squares, with a line to guide the eye). An increase in resistance upon bilayer formation and a decrease upon R-hemolysin addition can be monitored.

Figure 3. SPR angular reflectivity scans of the monolayer, the bilayer, and the bilayer after R-hemolysin incorporation. Fits are shown as solid lines for the first and last scan.

leads to a moderate shift in the reflectivity curve. It was proposed that the pore formation occurs in several steps, i.e., adsorption of monomers to the surface and pore formation upon lateral collision of seven monomers. In the SPR kinetics, an increase in thickness during several hours can be monitored. Upon rinsing, some material is lost from the surface, probably due to loosely adsorbed monomers or aggregates. A quantitative analysis of the SPR angular scans in Figure 3 is difficult. The SPR signal is proportional to the thickness and refractive index of a layer. The incorporated proteins, however, do not form a homogeneous layer, but they are dispersed, waterfilled pores in the membrane. The height of the extramembrane mushroom cap, as estimated from the crystal structure, is about 5 nm.14 However, when simulating the shift of the SPR curve by an additional layer, a height of 0.5 nm is obtained. This indicates a rather high distribution of pores and monomers in or on the membrane, which is in good agreement with AFM results.37 (37) Fang, Y.; Cheley, S.; Bayley, H.; Yang, J. Biochemistry 1997, 36, 95189522.

EIS is a more sensitive method to detect the incorporation of R-HL. Here only incorporated proteins contribute to the measured decrease in membrane resistance. When the protein monomers were added to a preformed tBLM, the resistance decreased almost instantaneously (Figure 2). The membrane resistance further decreased as a function of time. After rinsing, an additional decrease in membrane resistance could be observed, probably due to some rearrangements within the membrane, which might allow nonincorporated monomers or nonconductive aggregates to form functional heptamers. However, similar structures have not been observed in previous experiments.9,26 After this step, the resistance remained constant at a very low level, compared to the initial values. We investigated the influence of the monolayer architecture on the incorporation of the R-HL pores. DPhyTL and DPhyHDL should form different architectures, i.e., the lateral lipid density should be lower for a DPhyHDL layer and thus facilitate the protein incorporation. This hypothesis is confirmed by the EIS data. The decrease in resistance upon incorporation of R-HL is more pronounced in a DPhyHDL-based membrane than in a DPhyTL-based membrane (Figure 4A,B). Once the protein is functionally incorporated, the resistance decreases and an increase in capacitance can be observed. However, the phase shift does not reach -90° anymore, due to a deviation from an ideal capacitor. It has to be assumed that the large number of incorporated pores acts as a distribution of RC elements and therefore is seen as a deviation from an ideal capacitor. For DPhyHDL the changes in the electrochemical parameters are more pronounced than for DPhyTL. The membrane resistance decreases upon incorporation of R-HL to a value of only 840 Ω cm2. Also, the capacitance is affected by the incorporation in the case of DPhyHDL in contrast to DPhyTL (Figure 4C,D). After incorporation of the pores, the capacitance increases significantly, as the sealing properties of the membrane are corrupted. However, the capacitance of the last scan is still lower than that of the initial monolayer, the bilayer properties themselves stay untouched, and no complete breakdown of the membrane has to be assumed. In general, the bilayer capacitance in a DPhyHDL-based membrane was much less stable than in a

500 Langmuir, Vol. 24, No. 2, 2008

Vockenroth et al.

Figure 4. (A) Bode plot of a DPhyTL monolayer, bilayer, and a bilayer with incorporated R-hemolysin pores in NaCl/Mops buffer (c ) 0.1M) at 0 V potential; closed symbols show the impedance, open symbols the phase shift; fits are shown as solid lines. An increase in membrane resistance after bilayer formation and a decrease by a factor of 13 upon pore incorporation can be seen. The inset shows the equivalent circuit used for fitting the impedance data, a resistance in series with a RC element, and a space charge capacitance. The RC element describes the mono- and bilayer. (B) The same as in part A is shown for a DPhyHDL monolayer. An increase in membrane resistance by 3000 after bilayer formation and a decrease by a factor of 1300 upon pore incorporation can be seen, in contrast to the rather small changes with a DPhyTL monolayer. (C) Frequency reduced admittance plot of the same datasets as in part A. At the intersection of the semicircle with the y-axis, the change in capacitance between the scans can be seen. During membrane formation, the capacitance decreased and reincreased upon pore insertion. (D) The same as in part C is shown for a DPhyHDL monolayer. The decrease in capacitance during membrane formation and reincrease upon pore insertion are more pronounced in the case of DPhyHDL (note that C and D have the same data scale). Table 1. Membrane Resistance and Capacitance for the Two Different tBLM Systems Shown in Figure 4a

DPhyTL DPhyHDL

monolayer bilayer +R-HL monolayer bilayer +R-HL

Rmono/bi

Cmono/bi, µF cm-2

9.9 ( 0.34 MΩ cm2 19.3 ( 1.0 MΩ cm2 1.5 ( 0.32 MΩ cm2 0.33 ( 0.09 kΩ cm2 1.1 ( 0.11 MΩ cm2 0.84 ( 0.16 kΩ cm2

0.74 ( 0.011 0.69 ( 0.015 0.95 ( 0.06 3.50 ( 0.53 0.77 ( 0.02 1.53 ( 0.18

a The values were obtained by fitting the spectra to the equivalent circuit shown in Figure 4A. All values are normalized to 1 cm2. The error gives the deviation from the fit.

DPhyTL-based membrane, where the capacitance changed just slightly (Table 1). The differences in the membrane architecture have been systematically analyzed in a set of six experiments for each monolayer system (Figure 5). The differences in the two systems can be seen by the increase of the membrane resistance and the decrease in membrane capacitance upon vesicle fusion. This can be expressed by the coefficients Rbi/Rmono and Cbi/Cmono, respectively. Similarly, Rbi/RR-HL and Cbi/CR-HL show the effect of the protein incorporation on the bilayer resistance and capacitance, respectively.

The Rbi/Rmono coefficients show that the increase in resistance during membrane formation is much higher for DPhyHDL (mean value of 1769) than for DPhyTL (mean value of 4.6, cf. Figure 5A). However, the DPhyTL system shows higher absolute values and is more reproducible, whereas the scattering of the experimental data for DPhyHDL is quite high. The main difference between the two systems is the chemical structure of the anchor lipids. DPhyTL is known to form very densely packed membranes,38 as can be seen by the already high monolayer resistance. DPhyHDL monolayers have a very low resistance, which is probably due to a loosely packed monolayer. However, defects in the layer can be filled up by lipids from the vesicles during fusion. The DPhyHDL layers might therefore be more sensitive to defects during the membrane assembly. The quality of a membrane depends on factors such as surface roughness (below 0.5 nm by AFM) and already small impurities or irregularities lead to different electrochemical properties.36 The Rbi/RR-HL coefficients in Figure 5B show a similar trend. The changes in resistance upon protein incorporation are very high for DPhyHDL, with an average of 2774, whereas DPhyTL shows a much lower change by only an average factor of 44. The decrease in resistance due to ionic currents through incorporated (38) Kunze, J.; Leitch, J.; Schwan, A. L.; Faragher, R. J.; Naumann, R.; Schiller, S. M.; Knoll, W.; Dutcher, J. R.; Lipkowski, J. Langmuir 2006, 22, 5509-5519.

R-Hemolysin in Tethered Bilayer Architectures

Langmuir, Vol. 24, No. 2, 2008 501

Figure 5. The coefficients Rbi/Rmono (A), Rbi/RR-HL (B), Cbi/Cmono (C), and Cbi/CR-HL (D) from six independent experiments with DPhyTL and DPhyHDL monolayers are plotted as columns. Averages for both systems are given on the right of the corresponding columns. The symbols correspond to the absolute values and refer to the y-axis on the right. The higher increase in resistance and decrease in capacitance for DPhyHDL can be seen, as well as the more prominent changes upon pore insertion.

pores is therefore much higher in DPhyHDL-based membranes than in DPhyTL membranes. Similar details can be seen in the membrane capacitance, which gives a measure of the charge separation across the membrane. A monolayer has a higher capacitance than a bilayer, and upon membrane formation the capacitance decreases according to an ideal capacitor

C)

A0 d

for an area A, with plates separated by a distance d, a dielectric between the plates having a dielectric permittivity , and 0 being the permittivity of free space ()8.84 × 10-14 F cm-1).35,39 Upon formation of a bilayer, the distance d increases and leads to a decrease in capacitance. The Cbi/Cmono coefficients (Figure 5C) show smaller values for a DPhyHDL-based membrane (mean 0.25) than for a DPhyTL-based one (mean 0.86). However, the absolute values of Cmono of DPhyHDL are by far higher than the capacitances of DPhyTL monolayers. Though the differences in the monolayer capacitance are very pronounced, the bilayer capacitances reach about the same value with an average of 0.78 and 0.88 for DPhyTL and DPhyHDL, respectively. This shows that insulating membranes can be obtained with both systems, independent of the monolayer capacitance. Also, the pore incorporation shows a significant difference for the two architectures (Figure 5D). The distinct increase in (39) Bard, A. J.; Faulkner, L. R. Electrochemical MethodssFundamentals and Applications. 2nd ed.; John Wiley & Sons: New York, 2001.

capacitance upon pore incorporation (Cbi/CR-HL) is in good agreement with the change in capacitance upon membrane formation (Cbi/Cmono) for the two architectures. After protein incorporation, the Cbi/CR-HL coefficients for the DPhyHDL architecture are again smaller than for DPhyTL, which is due to the significantly higher capacitances of DPhyHDL membranes with incorporated pores. In all cases, the capacitance stays below 2.4 µF/cm2, which illustrates the integrity of the membrane. Even in the DPhyHDL-based system, the protein-doped membrane still showed a clearly defined capacitance, and therefore no loss of the sealing properties of the membrane could be observed. The different capacitance values of the two systems support the hypothesis of a more loosely packed DPhyHDL architecture. DPhyTL has a shorter spacer unit and is anchored via one lipoic acid moiety, whereas DPhyHDL has a longer spacer and is coupled to the support via two lipoic acid groups. These differences lead to variations in the structure of the two membranes. DPhyTL forms much denser membranes, while the bulky anchor group of DPhyHDL results in a more diluted layer. However, the bilayers of both systems show similar values, and therefore self-healing and back-filling of the loosely packed monolayer in the case of DPhyHDL are realistic. In this system, the addition of R-HL causes a higher decrease in R and a higher increase in C, which can signify a higher fluid and less densely packed membrane. Furthermore, the longer spacer group might lead to a larger ion reservoir under the membrane. This can influence the current flowing through the incorporated R-HL pores in a low-frequency ac measurement.

502 Langmuir, Vol. 24, No. 2, 2008

Conclusion Stochastic sensing approaches using ion channels offer in principle a highly sensitive and selective detection method. R-HL has been proposed as a model pore to be used in such approaches, however, in more fragile membrane architectures. Tethered bilayer lipid membranes are stable and versatile model platforms, which allow for the incorporation and functional study of membrane proteins. Controlling the membrane assembly for an optimal incorporation of a desired protein can give useful information about the pathway a protein inserts into a membrane. This facilitates the incorporation of complex proteins that so far could not be reconstituted in other biomimetic systems. The functional incorporation of R-HL into tethered bilayer systems was reported. Modification of the membrane architecture on a molecular scale had a significant effect on the protein incorporation.

Vockenroth et al.

This platform opens the way for novel sensing devices. For stochastic sensing, recent experiments using a similar architecture have shown the feasibility of single channel measurements in solid supported bilayers40,41 that however need to be extended and integrated into a read-out concept. However, the high sealing properties of the tethered membranes succeed in providing the necessary membrane requirements. Acknowledgment. This work was partially sponsored by DARPA through the MOLDICE program. LA7030279 (40) Andersson, M.; Keizer, H. M.; Zhu, C. Y.; Fine, D.; Dodabalapur, A.; Duran, R. S. Langmuir 2007, 23 (6), 2924-2927. (41) Keizer, H. M.; Dorvel, B. R.; Andersson, M.; Fine, D.; Price, R. B.; Long, J. R.; Dodobalapur, A.; Ko¨per, I.; Knoll, W.; Anderson, P. A. V.; Duran, R. S. ChemBioChemsCommun. 2007, 8 (11), 1246-1250.