Characterizing Stability Properties of Supported Bilayer Membranes

Jul 8, 2008 - Jong Ho Han, Joseph D. Taylor, K. Scott Phillips, Xiqing Wang, Pingyun Feng and Quan Cheng*. Department of Chemistry, University of ...
0 downloads 0 Views 500KB Size
Langmuir 2008, 24, 8127-8133

8127

Characterizing Stability Properties of Supported Bilayer Membranes on Nanoglassified Substrates Using Surface Plasmon Resonance Jong Ho Han,† Joseph D. Taylor,† K. Scott Phillips, Xiqing Wang, Pingyun Feng, and Quan Cheng* Department of Chemistry, UniVersity of California, RiVerside, California 92521 ReceiVed February 14, 2008. ReVised Manuscript ReceiVed May 6, 2008 Supported bilayer membranes (SBMs) formed on solid substrates, in particular glass, provide an ideal cell mimicking model system that has been found to be highly useful for biosensing applications. Although the stability of the membrane structures is known to determine the applicability, the subject has not been extensively investigated, largely because of the lack of convenient methods to monitor changes of membrane properties on glass in real time. This work reports the evaluation of the stability properties of a series of SBMs against chemical and air damage by use of surface plasmon resonance spectroscopy and nanoglassified gold substrates. Seven SBMs composed of phosphatidylcholine and DOPC+, including single-component, mixed, protein-reinforced SBMs (rSBMs) and proteintethered bilayer membranes (ptBLMs), are studied. The stability properties under various conditions, especially the effects of surfactants, organic solvents, and dehydration damage on the bilayers, are compared. PC membranes are found to be easily removed from the glassy surfaces using relatively low concentrations of the surfactants, while DOPC+ is markedly more stable toward nonionic surfactant. DOPC+ membranes also demonstrated remarkable air stability while PC films exhibited considerable damage from dehydration. Doping of cholesterol does not improve PC’s stability against SDS and Triton but changes the lipid membrane packing enough to protect against dehydration damage. Although rSBMs and ptBLMs improve air stability to a certain degree, they are still quite susceptible to significant damage/removal from ionic and nonionic surfactants at lower concentrations. Overall, DOPC+ has noted higher stability on glass, likely due to the favorable electrostatic interaction between the silicate surface and the lipid headgroup, making it a good candidate for application. Nanoglassy SPR proves to be an attractive platform capable of rapidly screening film stability in real-time, providing critical information for future work using supported membranes for sensing applications.

1. Introduction Over the last two decades, there has been considerable interest in attempting to harness the physicochemical properties of the cell membrane for sensing applications. The pivotal function of the cell membrane as the “gatekeeper” provides both access to and from the cell, while displaying key receptors for site-specific binding for a multitude of interactions, including cell-cell communication,1,2 receptor-ligand binding,3,4 and cell-surface enzymatic reactions.5,6 However, the membrane is difficult to extract from cells and reconstitute in situ. Tailoring its use to specific applications is also complicated by various imbedded constituents. Instead, the use of biomimetic membranes with controlled composition has proved to be simple and convenient. One of the key mimicking systems is the supported bilayer membrane (SBM) generated on solid substrates, with applications varying from transducer platforms in highly sensitive microchip immunoassays,7 to capillary surface coatings for eliminating nonspecific adsorption in electrophoretic separations.8 The fluid * Corresponding author. Tel: (951) 827-2702; e-mail: [email protected]. † Contributed equally as the first author. (1) Hafeman, D. G.; von Tscharner, V.; McConnell, H. M. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 4552–4556. (2) Groves, J. T.; Mahal, L. K.; Bertozzi, C. R. Langmuir 2001, 17, 5129– 5133. (3) Zemanova´, L.; Schenk, A.; Hunt, N.; Nienhaus, U.; Heilker, R. Biochemistry 2004, 43, 9021–9028. (4) Yang, T. L.; Baryshnikova, O. K.; Mao, H. B.; Holden, M. A.; Cremer, P. S. J. Am. Chem. Soc. 2003, 125, 4779–4784. (5) Bordier, C.; Etges, R. J.; Ward, J.; Turner, M. J.; Cardoso de Almeida, M. L. Proc. Natl. Acad. Sci. 1986, 83, 5988–5991. (6) Wacklin, H. P.; Tiberg, F.; Fragneto, G.; Thomas, R. K. Biochim. Biophys. Acta-Biomembr. 2007, 1768, 1036–1049. (7) Phillips, K. S.; Cheng, Q. Anal. Chem. 2005, 77, 327–334. (8) Cunliffe, J. M.; Baryla, N. E.; Lucy, C. A. Anal. Chem. 2002, 74, 776–783.

nature of the bilayer membrane also offers marked advantages over conventional methods for minimizing nonspecific protein adsorption,9 while simultaneously offering the ability to incorporate various receptors into the structural framework for increased ligand affinity.10 Although the number of diverse applications that exploit lipid bilayer membranes is mounting, the membrane stability and important factors that influence the stability have not been extensively studied. A few reports have recently demonstrated that damage may occur to SBMs under stresses such as dehydration,11,12 detergent washing,13 lipase interaction,14 aprotic solvent action,15 graft copolymer disruption,16 acidic and basic interferents,17 heat,18 and small organics.19 This information is important for biosensing applications, where the devices are used in complex systems with a myriad of chemicals that may impinge upon the usability of the lipid bilayer. Methods have been sought to stabilize the membrane in order to promote a more robust sensing platform. These have included the use of synthetic lipids presenting slight modifications to their natural counterparts,12 the addition of (9) Phillips, K. S.; Han, J. H.; Cheng, Q. Anal. Chem. 2007, 79, 899–907. (10) Doyle, E. L.; Hunter, C. A.; Phillips, H. C.; Webb, S. J.; Williams, N. H. J. Am. Chem. Soc. 2003, 125, 4593–4599. (11) Chiantia, S.; Kahya, N.; Schwille, P. Langmuir 2005, 21, 6317–6323. (12) Phillips, K. S.; Dong, Y.; Carter, D.; Cheng, Q. Anal. Chem. 2005, 77, 2960–2965. (13) Namita, D.; Somasundaran, P. Langmuir 2003, 19, 2007–2012. (14) Snabe, T.; Neves-Petersen, M. T.; Petersen, S. B. Chem. Phys. Lipids 2005, 133, 37–49. (15) Notman, R.; Noro, M.; O’Malley, B.; Anwar, J. J. Am. Chem. Soc. 2006, 128, 13982–13983. (16) Rossetti, F. F.; Reviakine, I.; Csu´cs, G.; Assi, F.; Vo¨ro¨s, J.; Textor, M. Biophys. J. 2004, 87, 1711–1721. (17) Artyukhin, A. B.; Stroeve, P. Ind. Eng. Chem. Res. 2003, 42, 2156–2162. (18) Lecuyer, S.; Charitat, T. Europhys. Lett. 2006, 75, 652–658. (19) Twardowski, M.; Nuzzo, R. G. Langmuir 2004, 20, 175–180.

10.1021/la800484k CCC: $40.75  2008 American Chemical Society Published on Web 07/08/2008

8128 Langmuir, Vol. 24, No. 15, 2008

stabilizing substances such as disaccharides,11 incorporating a protein “blanket” in the case of protein-reinforced supported bilayer membranes (rSBMs) that maintain lateral fluidity,20 and several recent tethered bilayer lipid membranes (tBLMs) involving either anchor molecules or polymer cushions that suspend the lipid sheet above the solid support.21–24 Each method has exhibited certain advantages over traditional SBMs; yet an efficient, convenient method for parallel evaluation of lipid systems has been previously unattainable. It is well-known that membrane systems require a hydrated environment in order to function properly. Many factors may interfere with the hydration conditions, including the alkalinity and ionic strength of the working buffers,16 matrix effects from nonpurified samples, and the actual contents of the bilayer’s lipid framework, such as cholesterol content.25–27 Very little work has been done to compare the effects of these factors sideby-side on the stability properties of membrane systems. Lack of this information impairs the development of biosensor technology that employs lipid membranes. Without a direct comparison of one system to another, it is difficult to know which lipid film functions most advantageously with the least deleterious effects resulting from chemical or environmental perturbants. To address this issue, we report here the evaluation of biomimetic “glass-based” lipid membrane systems by using realtime surface plasmon resonance (SPR) spectroscopy in combination with nanoglassified gold substrates we previously developed.28 These substrates allow SPR to be employed to monitor biological interactions confined to glass substrates,23 a very common surface for generating supported membranes. Bilayer formation on glass via vesicle fusion was conveniently monitored, and subsequent injections of different interferents were employed to assess and compare bilayer stability and durability. Both natural and synthetic lipid systems were used in this work, and different ratios of mixtures of these lipids were tested to unveil any deviations in stability. Previously we have focused on improving bilayer stability with protein-reinforced supported bilayer membranes,20 where biotinylated lipid was incorporated into PC vesicles for fusion onto a PDMS substrate, followed by injection of streptavidin to provide a protective layer that allowed the lipid bilayer to remain laterally fluid. We recently demonstrated that protein-tethered bilayer lipid membranes (ptBLMs) can be used to form patterned bilayer arrays using microfluidics,23 which have been shown to remain intact, fluid, and stable for several days in solution. A thorough comparison of these membrane systems on glass vs conventional supported membranes on the same substrate for their stability properties should provide key information for future work using supported membranes for sensing applications. The combination of SPR and nanoglassy substrates offers a fast, efficient screening method for glassbased membrane systems and may prove to be an invaluable tool for pharmaceutical analysis as the technique could be easily expanded to screen membrane systems consisting of transmem(20) Dong, Y.; Phillips, K. S.; Cheng, Q. Lab Chip 2006, 6, 675–681. (21) Terrettaz, S.; Mayer, M.; Vogel, H. Langmuir 2003, 19, 5567–5569. (22) Keizer, H. M.; Dorvel, B. R.; Andersson, M.; Fine, D.; Price, R. B.; Long, J. R.; Dodabalapur, A.; Koper, I.; Knoll, W.; Anderson, P. A. V.; Duran, R. S. ChemBioChem 2007, 8, 1246–1250. (23) Taylor, J. D.; Phillips, K. S.; Cheng, Q. Lab Chip 2007, 7, 927–930. (24) Purrucker, O.; Goennenwein, S.; Foertig, A.; Jordan, R.; Rusp, M.; Baermann, M.; Moroder, L.; Sackmann, E. Soft Matter 2007, 3, 333–336. (25) Raffy, S.; Teissie´, J. Biophys. J. 1999, 2072–2080. (26) Huang, J.; Buboltz, J. T.; Feigenson, G. W. Biochim. Biophys. Acta 1999, 1417, 89–100. (27) Pata, V.; Dan, N. Biophys. J. 2005, 88, 916–924. (28) Phillips, K. S.; Han, J. H.; Martinez, M.; Wang, Z.; Carter, D.; Cheng, Q. Anal. Chem. 2006, 78, 596–603.

Han et al.

brane proteins that must be kept in a pseudonatural environment for proper function evaluation.

2. Experimental Section 2.1. Materials and Instrumentation. 3-Mercaptopropionic acid (3-MPA), poly(allylamine hydrochloride) (PAH), avidin, Triton X-100, cardiolipin (CA) sodium salt from bovine heart, and cholesterol (5-cholesten-3β-ol) (CHO) were obtained from SigmaAldrich (St. Louis, MO). Biotinylated bovine serum albumin (biotBSA) was from Pierce (Rockford, IL). L-R-Phosphatidylcholine (PC), 1-palmitoyl-2-oleoyl-sn-glycero-3-ethylphosphocholine (chloride salt) (DOPC+), 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamineN-(cap biotinyl) (sodium salt) (biot-PE), and L-R-phosphatidylethanolamine, transphosphatidylated (PE) were from Avanti (Alabaster, AL). Sodium silicate was from Fisher, and sodium dodecyl sulfate (SDS) was from BIO-RAD (Hercules, CA). Gold substrates were fabricated with a 46-nm thick gold layer deposited by an e-beam evaporator onto cleaned glass slides pretreated with mercaptoalkylsilane. A Biosuplar II instrument from Analytical µ-Systems (Regensburg, Germany) was used for SPR experiments. Fluorescence recovery after photobleaching (FRAP) experiments for membrane characterization were performed on a Meridian Insight confocal laser scanning microscope (CLSM) with 488 nm argon laser excitation, SPOT Pursuit CCD, and a fluorescein emission filter used in conjunction with a 40×/0.75na Achroplan dipping objective. 2.2. Preparation of Silicate Chips and Vesicle Solutions. Preparation of calcinated silicate layers on gold substrates has been detailed in a previous report.28 In brief, cleaned gold substrates were immersed in a 10 mM 3-MPA ethanol solution overnight, followed by extensive rinsing with ethanol and pure DI water (>18 MΩ). Using spray bottles, PAH (1 mg/mL in DI water, adjusted to pH 8.0) and sodium silicate solution (22 g/L, adjusted to pH 9.5) were alternately sprayed for 30 s each and were rinsed/dried with DI water and nitrogen, respectively, between each spray. This process was repeated to build up to a total of 15 layers while monitoring with SPR for quality control of each layer. Finally, the layered chips were calcinated in a furnace by heating to 450 °C at a rate of 17 °C per min and brought to room temperature after 4 h. Vesicle solutions were prepared from stock solutions in chloroform. Generally, the appropriate mole percent of each lipid was mixed together in small vials, gently purged with a nitrogen stream to form a dry lipid film, and then resuspended in solution with Tris buffer (10 mM Tris; 150 mM NaCl; pH 7.4) to a lipid concentration of 1 mg/mL. After vigorous vortexing, the suspended lipids were probe sonicated with a sonifier for 20 min. Vesicles used for tethered membrane studies were made in phosphate-buffered saline (20 mM PBS; 150 mM NaCl; pH 7.3) to avoid free amine interaction Tris would have with the aldehyde-derivatized substrate. The resuspended lipids were similarly sonicated and then centrifuged at 8000 rpm for 5 min to remove any titanium particles released from the probe tip during sonication. The supernatant was extruded by passing the solution through a polycarbonate membrane of 100 nm pore size to generate small unilamellar vesicles (SUVs), followed by incubation at 4 °C for 1 h. 2.3. SPR Experimental Procedure. A variety of vesicle solutions were assembled on the calcinated chips by injection into the SPR flow cell via a syringe pump and a low pressure injector. The vesicle solutions were incubated by stopping flow for 1 h. Otherwise, flow was maintained at a rate of 8.3 mL/h. After a 10 min rinse to ensure complete removal of residual vesicles, different concentrations of removal agents were injected through the sample loop at a slower flow rate. For dehydration studies, air was injected through the sample loop and allowed to reach the center of the flow cell where it was stopped for 30 min to allow the membranes to dry. All experiments were carried out at room temperature. For the protein-reinforced membranes, vesicles composed of 95% PC and 5% biot-PE were incubated and allowed to fuse to form a bilayer. A 0.5 mg/mL solution of avidin was then injected through the sample loop and incubated for 10 min in a manner similar to previously published results.20 For the tethered membranes, the

Supported Bilayer Membranes on Nanoglassified Substrates procedure used here is similar to that reported in a previous publication.23 In short, an ethanol solution containing triethyloxysilane aldehyde was reacted with the calcinated gold substrates to render an aldehyde-presenting surface for attachment of biot-BSA (0.5 mg/mL in PBS) via formation of a Schiff’s base. After rinsing, a 0.2 mg/mL solution of avidin was injected through the sample loop and allowed to incubate for 1 h, rinsed, then exposed to vesicles composed of 46% PC, 35% PE, 18% CA, and 1% biot-PE. These vesicles were then loaded onto the avidin sublayer for 1 h, followed by triggered fusion by injecting PEG-8000 (30% w/v in PBS).

3. Results and Discussion 3.1. Fluorescence Characterization of Membranes On Nanoglassy Gold Substrates. Glass has long been the standard substrate in fluorescence-based biointeraction analysis where a wide assortment of established attachment schemes has been exploited. The previously reported nanometer-scale silicate coatings on the gold surface using a layer-by-layer (LbL) deposition and calcination process allow a glass layer to be formed for SPR analysis.28,29 On these substrates, lipid bilayer formation and membrane properties can be studied directly in real-time. Previously such studies would typically be performed with fluorescence microscopy or AFM, requiring tedious processes and time-consuming image analysis. Nevertheless, such methods are well-established techniques that aid in the verification of laterally fluid bilayer lipid membranes. To confirm the attributes of a single lipid bilayer, fluorescence recovery after photobleaching (FRAP) experiments were carried out using a confocal fluorescence microscope with DOPC+ vesicles doped with a small amount of fluorescent lipid. It was possible on the glassy film on gold because the distance between the fluorophore-doped membrane and the gold surface is adequately separated to suppress fluorescence quenching that may occur from the proximity of the metal.30 This is a distinct advantage of nanoglassy SPR substrates as compared to other chemistries available on gold, allowing FRAP measurement to be performed directly on the same chip. FRAP allows the study of membrane mobility and the determination of diffusion coefficients, D. Typical D values for DOPC+ and similar synthetic lipids obtained on glass surfaces lie between 2 and 5 µm2/s.12,31–33 Figure 1 shows FRAP images and the recovery profile for the DOPC+ membrane. Data were fit to previously established protocols,7 yielding a diffusion coefficient of 4.4 µm2/s, corresponding well to previous results. The results clearly show that the lipid structure on the calcinated film is bilayer and the lipid membrane is mobile. It should be noted that fluorescently labeled lipids need be incorporated into lipid membranes to characterize membrane properties such as lateral mobility. As one of the phospholipid acyl chains is modified with a fluorescent tag, it has been shown to hamper the actual lipid packing and arrangement of the model membrane under close scrutiny.34 Without having to rely on a label to transduce a quantifiable signal, SPR is more advantageous than traditional methods to study some of the membrane properties, albeit not including mobility, in a much more quasinatural environment. When used in conjunction with nanoglassy gold substrates, it is able to make parallel comparisons for a number of systems established on glass surfaces. (29) Phillips, K. S.; Wilkop, T.; Wu, J. J.; Al-Kaysi, R. O.; Cheng, Q. J. Am. Chem. Soc. 2006, 128, 9590–9591. (30) Lakowicz, J. R. Anal. Biochem. 2001, 298, 1–24. (31) Albertorio, F.; Diaz, A. J.; Yang, T.; Chapa, V. A.; Kataoka, S.; Castellana, E. T.; Cremer, P. S. Langmuir 2005, 21, 7476–7482. (32) Kalb, E.; Frey, S.; Tamm, L. K. Biochim. Biophys. Acta 1992, 1103, 307–316. (33) Tamm, L. K.; McConnell, H. M. Biophys. J. 1985, 47, 105–113. (34) Kuerschner, L.; Ejsing, C. S.; Ekroos, K.; Shevchenko, A.; Anderson, K. I.; Thiele, C. Nat. Methods 2005, 2, 39–45.

Langmuir, Vol. 24, No. 15, 2008 8129

3.2. Real-Time Monitoring of Interactions Using Nanoglassy Gold Substrates. We monitored membrane interactions including vesicle fusion on the glass and the effects of chemical or environmental perturbants on the assembled membrane structures with SPR. Stability experiments were carried out employing three removal methods: organic solvents, ionic and nonionic surfactants, and dehydration. PC and PE were chosen as representatives of natural membranes, as they are the two major constituents in mammalian cells. Data extracted from SPR sensorgrams for the various membranes used in this study are shown in Table 1. Data are presented as the absolute percentage of the membrane that had been removed or damaged, using (∆θ /∆θ°) × 100 where ∆θ° represents the change in minimum SPR angle from initial vesicle fusion and membrane assembly, and ∆θ represents the angular change (loss of materials) due to introduction of the perturbants. As expected from previous work, PC can be easily removed from the glassy surface using relatively low concentrations of the nonionic surfactant Triton X-100 (56% change with 0.1% Triton). Conversion of zwitterionic PC lipids to a cationic species through esterification of the phosphate moiety of the headgroup yields DOPC+, a sterically bulkier lipid due to both the added ethyl group and the greater electrostatic repulsion associated with the positively charged headgroup. DOPC+ has been shown to be an exceptionally stable synthetic lipid on PDMS in its ability to remain nearly fully functional, intact, and fluid despite being exposed to complete dehydration.12 Interestingly, SPR data shows a similar result for DOPC+ on glass, which is markedly more stable toward Triton X-100 than its natural counterpart PC. More interestingly, the membrane stability seems to be affected by the chemical properties of the disrupting agent. Figure 2 shows an SPR sensorgram displaying spontaneous fusion of DOPC+ vesicles and their removal by SDS on nanoglassy gold substrates. From Table 1, DOPC+ and PC exhibit similar removal effects from the same concentrations of the ionic surfactant SDS. High concentration of SDS (0.5%) removes both membranes effectively, demonstrating its suitability for complete removal of lipid bilayers from the surface to generate a fresh substrate. It should be mentioned that 0.01% Triton X-100 showed no effect on any of the membranes studied in this report while 5% Triton X-100 demonstrated complete membrane removal. For applications involving biomimetic lipid membranes, air stability is an issue that must be addressed, as well as the ability to resist common organic solvents that may be present. We tested each lipid system against 5% ethanol, approximately the same amount found in an average beer. In addition, dehydration of the SPR flow cell for 30 min was used to test the membranes’ ability to rehydrate and resist mechanical damage from the pressuredriven flow of the SPR cell. Results indicate that 5% ethanol has no influence upon any of the membrane systems studied in this report. However, PC membranes exhibited considerable damage from dehydration up to 17%, whereas DOPC+ was not affected at all. The results are very similar to the membrane dehydration test carried out in PDMS microchannels and characterized using fluorescence microscopy.12 In that case, results showed that 12% of the overall assembled PC membrane was subject to damage, while 11% was completely removed from the surface. On contrast, only 6% of DOPC+ membranes were damaged and only 1% removed.12 To understand if the membrane stability can be tuned through mixing, PC lipids were blended with synthetic lipid dopants to test for stability variations. Table 1 summarizes the results. PC mixed in a 1:1 molar ratio with DOPC+ exhibited complete protection against dehydration, while PC mixed with another

8130 Langmuir, Vol. 24, No. 15, 2008

Han et al.

Figure 1. (a) FRAP results for DOPC+ membranes on the nanoglassy gold SPR substrate. Image 1 shows a pristine membrane, while 2, 3, 4, 5, and 6 represent 5, 10, 15, 30, and 45 s after bleaching. (b) FRAP recovery curve for DOPC+ on nanoglassified gold. Table 1. The Percent Removal of Various Supported Membranes under Different Conditions on Nanoglassified SPR Substratesa 5% ethanol 0.1% Triton 0.5% Triton dehydration 0.1% SDS 0.5% SDS a

PC

DOPC+

PC:DOPC+ (50:50)

PC:PE (50:50)

PC:CHO (50:50)

PC:biot-PE (95:5)

ptBLM

0 56.2 76.1 17.1 20.5 83.9

0 24.8 53.9 0 18.9 98.5

0 23.9 55.1 0 56.1 91.4

0 34.5 81.8 4.7 29.2 94.8

0 64.1 79.7 0 92.2 s

0 39.8 64.6 0 26.3 89.8

0.2 77.4 100 s 100 s

0.01% Triton X-100 showed no effect on any of the membranes studied in this report while 5% Triton X-100 demonstrated complete membrane removal.

natural lipid PE in a 1:1 molar ratio showed about 5% damage to the membrane. The 1:1 PC:DOPC+ membrane exhibited similar removal effects as DOPC+ from 0.1% Triton yet could be substantially removed (>56%) by the lower concentration of 0.1% SDS. The mixed PC:PE membrane showed similar removal

effects from both ionic and nonionic surfactants at the lower concentration threshold. 3.3. Cholesterol-Containing Membranes. Cholesterol is a major component of cell membranes in body tissue, particularly in the plasma membranes of mammalian cells where it has been

Supported Bilayer Membranes on Nanoglassified Substrates

Figure 2. SPR sensorgram showing DOPC+ vesicle injection and spontaneous fusion on a nanoglassy gold SPR substrate, followed by stability test for partial and complete membrane removal by SDS.

found to range from about 25 to 50 mol%.35 Furthermore, inclusion of cholesterol at high concentrations in lipid membranes has shown to yield a much more stable bilayer, rendering an interface that reduces passive permeability.36 Feigenson and coworkers have demonstrated an upper solubility limit of 66 mol% cholesterol in PC bilayers using X-ray diffraction.26 The results for stability testing on 50 mol% CHO/50 mol% PC composition membranes are given in Table 1. Surprisingly, the PC:CHO membranes were quite removable from the surface, showing the largest % removal at the low concentrations of SDS and Triton. However, the cholesterol did alter the lipid membrane packing enough to completely protect against dehydration damage. This result is consistent with previous predictions using the Umbrella model, which suggests that the larger, polar headgroups of the phospholipids cover the nonpolar cholesterol in the membrane in order to elude the unfavorable free energy of water coming into contact with the largely hydrophobic cholesterol.37 In this case, the use of natural PC lipids containing unsaturated hydrocarbon chains allows less room beneath their large headgroups for the cholesterol because of the cis configuration of double bonds on the acyl chain. When the surfactant is introduced, the higher free energy associated with this configuration promotes the dissolution of the cholesterol-containing membrane, which is easily disengaged from the surface. The reason for enhanced air stability of cholesterol containing membranes is still unclear. To test the reproducibility between measurements, the silicate/ Au substrates were exposed to cholesterol-containing PC membranes, allowed to incubate, then exposed to 0.1% SDS. Triplicate measurements were performed, yielding an average removal of 92.2% with a relative standard deviation (RSD) of 7.4%. We extended the stability testing to DOPC+ membranes, running triplicate measurements on the same substrates with 0.5% SDS as the removal agent. The results show only slight variability between measurements, yielding an average removal of 98.5% with an RSD of 1.9%. These results clearly indicate the versatility of using SPR to monitor glass-based membrane interactions, and the reproducibility of the measurements is excellent. (35) Bloch, K. In New ComprehensiVe Biochemistry; Vance, D. E.; Vance, J. E., Eds.; Biochemistry of Lipids, Lipoproteins and Membranes; Elsevier: Amsterdam, 1991; Vol. 20, pp 363-381. (36) Bloom, M.; Evans, E.; Mouritsen, O. G. Q. ReV. Biophys. 1991, 24, 293–397. (37) Parker, A.; Miles, K.; Cheng, K. H.; Huang, J. Biophys. J. 2004, 86, 1532–1544.

Langmuir, Vol. 24, No. 15, 2008 8131

Figure 3. SPR sensorgram for vesicles composed of 95% PC and 5% biot-PE in the process of forming a reinforced supported bilayer membrane with avidin and its removal with different concentrations of SDS.

3.4. Stability of Protein-Reinforced Supported Bilayer Membranes. Cremer and co-workers recently demonstrated that incorporating a small percentage of a biotinylated lipid into PC membranes allows for the use of streptavidin proteins to provide a protective layer, enhancing the membrane stability at the air-water interface while retaining lateral fluidity.38 We have used a similar system in biosensor fabrication, taking advantage of the strong biotin-avidin interaction property and the simple fabrication procedure.20 Here we studied the stability of such membranes against different surfactants. Figure 3 shows the assembly process of the rSBMs on nanoglassy gold substrates followed by injection of different concentrations of SDS. Although this membrane does provide complete protection against dehydration damage, as has been previously demonstrated and reconfirmed here with SPR, it is still quite susceptible to significant damage/removal from ionic and nonionic surfactants at lower concentrations (Table 1). The coating of vesicles with avidin has been used in solute-membrane interactions in an electrophoresis study.39 The separation of phospholipids using micellar electrokinetic chromatography (MEKC) often uses 1-propanol as an organic modifier.40–44 These buffers typically consist of a surfactant and greater than 30% V/V short chain organic modifier. In order to determine the amount of 1-propanol that can be used in such a system without damaging the avidin-membrane coating, screening for the proper percentage of organic modifier to be used in the separation buffer is an important experimental step. SPR spectrometry in conjunction with nanoglassy gold substrates can be used for such a purpose. Figure 4 shows an SPR sensorgram for vesicle self-assembly, followed by injection and incubation of avidin. High volume percentages of 1-propanol were injected sequentially to find the threshold at which the membrane begins to decompose from the higher concentrations of organic solvent. (38) Holden, M. A.; Jung, S. Y.; Yang, T.; Castellana, E. T.; Cremer, P. S. J. Am. Chem. Soc. 2004, 126, 6512–6513. (39) Yang, Q.; Liu, X. Y.; Miyake, J. Supramol. Sci. 1998, 5, 769–772. (40) Ingvardsen, L.; Michaelsen, S.; Sørensen, H. J. Am. Oil Chem. Soc. 1994, 71, 183–188. (41) Szu¨cs, R.; Verleysen, K.; Duchateau, G. S. M. J. E.; Sandra, P.; Vandeginste, B. G. M. J. Chromatogr. A 1996, 738, 25–29. (42) Verleysen, K.; Sandra, P. J. High Resolut. Chromatogr. 1997, 20, 337– 339. (43) Lin, S.; Fischl, A. S.; Bi, X.; Parce, W. Anal. Biochem. 2003, 314, 97– 107. (44) Mwongela, S. M.; Lee, K.; Sims, C. E.; Allbritton, N. L. Electrophoresis 2007, 28, 1235–1242.

8132 Langmuir, Vol. 24, No. 15, 2008

Figure 4. Formation of reinforced supported bilayer membranes with avidin and vesicles composed of 95% PC and 5% biot-PE and their stability against 1-propanol.

At about 20% 1-propanol, a slight decrease in the minimum SPR angle was observed. To validate that the threshold has been reached, a slightly higher percentage was injected to see if the membrane would be further removed. Injection of 22% 1-propanol led to a major decrease in SPR signal, providing a real-time approach for fast prototyping and optimization of experimental conditions. 3.5. Protein-Tethered Bilayer Lipid Membranes. Several recent reports have studied the stability of tethered bilayer membranes.21,45–49 We have also demonstrated the fabrication of arrays of ptBLMs within PDMS microchannels. These arrays could be used for biosensing of cholera toxin and were found to be stable and remain fluid after days in buffer solution.23 Figure 5 shows the assembly process for the tethered bilayer on a calcinated surface. The initial tracking of the minimum SPR resonance angle is noticeably higher relative to other sensorgrams in this manuscript. This is due to functionalization of the nanoglassy substrates to render an aldehydic surface for protein sublayer attachment. This is yet another distinct advantage of the glassy gold surface in SPR, in that standard glass modification schemes can be easily employed for enhanced biointerface construction. The first segment of the sensorgram shows the interaction between biot-BSA and the aldehyde substrate, which forms Schiff’s bases. After the signal stabilizes, flow is resumed and the signal increases slightly, likely because of residual protein remaining in the flow loop. Avidin is then injected, incubated, rinsed, and subsequently exposed to biotin-containing vesicles. In this case, spontaneous vesicle fusion does not occur on such a surface, and therefore a concentrated solution of PEG-8000 (30% w/v) is used.50–52 The PEG solution is injected, allowed to incubate for ca. 8 min, and then adequately rinsed. Bilayer (45) Andersson, M.; Keizer, H. M.; Zhu, C.; Fine, D.; Dodabalapur, A.; Duran, R. S. Langmuir 2007, 23, 2924–2927. (46) Schiller, S. M.; Naumann, R.; Lovejoy, K.; Kunz, H.; Knoll, W. Angew. Chem., Int. Ed. 2003, 42, 208–211. (47) Krishna, G.; Schulte, J.; Cornell, B. A.; Pace, R.; Wieczorek, L.; Osman, P. D. Langmuir 2001, 17, 4858–4866. (48) Sinner, E. K.; Knoll, W. Curr. Opin. Chem. Biol. 2001, 5, 705–711. (49) 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. (50) Lentz, B. R. Eur. Biophys. J. 2007, 36, 315–326. (51) Elie-Caille, C.; Fliniaux, O.; Pantigny, J.; Mazie`re, J. C.; Bourdillon, C. Langmuir 2005, 21, 4661–4668. (52) Haque, M. E.; McIntosh, T. J.; Lentz, B. R. Biochemistry 2001, 40, 4340– 4348.

Han et al.

Figure 5. SPR sensorgram showing the formation of a protein-tethered bilayer lipid membrane on a calcinated gold substrate. The silicate layer was reacted with an aldehydic solution prior to the SPR measurement. Promoted fusion is accomplished using a high concentration (30% w/v) of PEG-8000.

membrane formation has previously been demonstrated at this point in a previous study by disassembling the gold chip from the SPR and carefully transferring it to the stage of a confocal fluorescence microscope for FRAP studies.23 Interestingly, the tethered bilayer membrane is relatively easily removed from the protein surface with a low concentration of Triton. After injection of 0.5% Triton X-100, the bilayer is completely removed and the signal returns to the avidin baseline. A 0.1% SDS solution also removes the lipid assembly entirely from the tethering layer. These rather unexpected results appear to indicate that tethered membranes have the least stability of all the forms tested herein. It is possible that the area of entrapped water molecules under the membrane increases the surface area of solvent access, thereby contributing to increased instability toward dissolved surfactants. In addition, loss of favorable interactions between the polar head groups of the lipid and the substrate may also be a factor. Nevertheless, the process can be repeated many times for membrane fusion and removal, indicating that the tethered membrane is an ideal platform for use in the fabrication of lipid arrays for high throughput measurement including SPR imaging analysis. With a little modification in the process, a fresh, regenerated patterned surface could be established after each experiment, drastically cutting down on materials, cost, and the intense labor of experimental assembly.

4. Conclusions In this work, we investigated the stability properties of a series of supported membranes on nanoglassy gold substrates with SPR. We chose two lipids that have been commonly used for biosensing applications, PC and DOPC+, and various structures including simple bilayer membranes, mixed membranes, protein-reinforced bilayer membranes, and protein-tethered bilayer membranes. The air stability and the stability against both low and high concentrations of ionic and nonionic surfactants are compared. The nanoglassy substrates prove to be an excellent platform for rapid prototyping of glass-based biointeractions when combined with surface plasmon resonance spectroscopy. Vesicle assembly can be easily monitored on these surfaces without extensive chemical modification of the substrate, and the flow cell setup allows introduction of various perturbants with ease. In addition, this substrate facilitates the implementation of a wide range of commercially available glass functionalization schemes for a

Supported Bilayer Membranes on Nanoglassified Substrates

variety of complex fabrication strategies to suit the needs of diverse applications. This was demonstrated by functionalizing the silicate to an aldehyde-presenting surface for the development of a tethered bilayer lipid membrane on glass, which previously was studied primarily by fluorescence methods. Using this technique, several supported bilayer membrane systems have been evaluated in parallel in a relatively short time frame. The stability of natural PC and PE lipids was evaluated against synthetic DOPC+ lipids, as well as protein-reinforced and tethered bilayer techniques. Comparing all of the membrane structures studied here, DOPC+ has noted stability on glass, which coincides with previous stability studies of this lipid in a microfluidic dehydration damage study. The positively charged lipid headgroup is an important aspect. The top layer of the calcinated substrate presents hydroxylated silanol moieties, yielding a net negative charge along the solid-liquid (buffer) interface. This fosters a favorable electrostatic interaction between the surface and the quaternary ammonium ion found on the lipid headgroup. The tightly packed membrane partially accounts for

Langmuir, Vol. 24, No. 15, 2008 8133

a considerably different refractive index relative to PC membranes as observed from the change in minimum SPR angle for vesicle fusion. This may also contribute to the marked stability toward surfactants and its resistance against dehydration damage. For the important characteristic of air stability, DOPC+ membranes demonstrated ideal behavior, as did protein-reinforced membranes and cholesterol-containing PC membranes. However, 100% DOPC+ exhibited substantially less surfactant damage relative to all other systems employed in this study, suggesting the high potential of this membrane for long-term biosensor applications. With the chip-to-chip variability of this approach baring excellent RSD as low as 1.9%, glass functionalized gold substrates and SPR provide a powerful tool for rapid prototyping of a variety of glass-based techniques in real-time. Acknowledgment. The authors acknowledge financial support from NSF grant CHE-0719224. LA800484K