Photopolymerization of Mixed Monolayers and Black Lipid

Dec 24, 2005 - 7 Wells AVenue, Newton, Massachusetts. ReceiVed August 25, 2005. In Final Form: October 25, 2005. We formed monolayers and black lipid ...
0 downloads 0 Views 213KB Size
Download PDF
Langmuir 2006, 22, 1215-1222

1215

Photopolymerization of Mixed Monolayers and Black Lipid Membranes Containing Gramicidin A and Diacetylenic Phospholipids Susan M. Daly,*,† Linda A. Heffernan,† William R. Barger,‡ and Devanand K. Shenoy†,§ Center for Bio/Molecular Science and Engineering, NaVal Research Laboratory, 4555 OVerlook AVenue SW, Washington, D.C. 20375, and Geo-Centers, Inc., 7 Wells AVenue, Newton, Massachusetts ReceiVed August 25, 2005. In Final Form: October 25, 2005 We formed monolayers and black lipid membranes (BLMs) of photopolymerizable lipids mixed with the channelforming protein gramicidin A to evaluate their miscibility and the potential for improved stability of the BLM scaffold through polymerization. Analyses of surface pressure vs area isotherms indicated that gramicidin A dispersed with three different synthetic, polymerizable, diacetylene-containing phospholipids, 1,2-di-10,12-tricosadiynoyl-sn-glycero3-phosphocholine (DTPC), 1,2-di-10,12-tricosadiynoyl-sn-glycero-3-phosphoethanolamine (DTPE), and 1-palmitoyl2,10,12-tricosadiynoyl-sn-glycero-3-phosphoethanolamine (PTPE) to form mixed monolayers at the air-water interface on a Langmuir-Blodgett (LB) trough. Conductance measurements across a diacetylenic lipid-containing BLM confirmed dispersion of the gramicidin channel with the lipid layer and demonstrated gramicidin ion-channel activity before and after UV exposure. Polymerization kinetics of the diacetylenic films were monitored by film pressure changes at constant LB trough area and by UV-vis absorption spectroscopy of polymerized monolayers deposited onto quartz. An initial increase in film pressure of both the pure diacetylene lipid monolayers and mixed films upon exposure to UV light indicated a change in the film structure. Over the time scale of the pressure increase, an absorbance peak indicative of polymerization evolved, suggesting that the structural change in the lipid monolayer was due to polymerization. Film pressure and absorbance kinetics also revealed degradation of the polymerized chains at long exposure times, indicating an optimum time of UV irradiation for maximized polymerization in the lipid layer. Accordingly, exposure of polymerizable lipid-containing black lipid membranes to short increments of UV light led to an increase in the bilayer lifetime.

Introduction Ion channels in cell membranes are ubiquitous and serve a wide variety of biological and physiological functions.1 These channels can be reconstituted into black lipid membranes to serve as sensors for the detection of molecules, ranging from small molecules and ions to biological macromolecules.2,3 Recently, interest in ion-channel-based sensors has been augmented by work suggesting their potential for sequencing of DNA.4-8 While this technology seems promising, the durability of lipid-bilayer-based stochastic sensors needs improvement.7 Furthermore, the fluidity of some phospholipid bilayers allows ion channels to diffuse laterally within the lipid layer. This fluidity can be a problem for sensor geometries that use unsupported bilayer membranes for channel reconstitution, e.g., painted or folded bilayers because the channel is not necessarily fixed over an aperture. Improvements of the sensor robustness and lifetime are necessary before the channel-based sensor can make the transition from laboratory to practical use.9,10 * Corresponding author. E-mail: [email protected]. † Center for Bio/Molecular Science and Engineering. ‡ Geo-Centers, Inc. § Currently on detail at DARPA. (1) Aidley, D. J.; Stanfeld, P. R., Eds. Ion Channels: Molecules in Action; Cambridge University Press: Cambridge, 1996. (2) Cheley, S.; Gu, L.; Bayley, H. Chem. Biol. 2002, 9, 829-838. (3) Kasianowicz, J. J.; Henrickson, S. E.; Weetall, H. H.; Robertson, B. Anal. Chem. 2001, 73, 2268-2272. (4) Muthukumar, M. Electrophoresis 2002, 23, 1417-1420. (5) Vercoutere, W.; Akeson, M. Curr. Opin. Chem. Biol. 2002, 6, 816-822. (6) Kasianowicz, J. J.; Brandin, E.; Branton, D.; Deamer, D. W. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 13770-13773. (7) Deamer, D. W.; Akeson, M. Trends Biotechnol. 2000, 18, 147-151. (8) Andersen, O. S. Biophys. J. 1999, 77, 2899-2901. (9) Bayley, H.; Cremer, P. S. Nature 2001, 413, 226-230. (10) Peterman, M. C.; Ziebarth, J. M.; Braha, O.; Bayley, H.; Fishman, H. A.; Bloom, D. M. Biomed. DeVices 2002, 4, 231-236.

It is well-known that lipid polymerization has a significant impact on the stability of bilayer vesicles.11,12 Polymerization behavior of lipids on solid substrates has also received considerable attention recently.13-18 Polymerized lipids on solid supports are being investigated for use as protein-resistant surfaces17 or biomimetic membranes.15,16,19 Polymerization of lipid bilayers causes decreased bilayer solubility in detergent solutions and organic solvents20 and a decreased lateral diffusion coefficient of the lipid.21 These properties suggest that polymerization of the lipid bilayer is one technique that could facilitate construction of more mechanically stable BLMs. Benz and co-workers were the first to use polymerized lipids to improve the lifetime of BLMs.22 Their results with dienoyl lipids showed enhanced stability and reduced carrier-mediated transport across the membrane. Their work sets the stage for an exploration of other (11) O’Brien, D. F.; Armitage, B.; Benedicto, A.; Bennett, D. E.; Lamparski, H. G.; Lee, Y. S.; Srisiri, W.; Sisson, T. M. Acc. Chem. Res. 1998, 31, 861-868. (12) Freeman, F. J.; Chapman, D. Liposomes as Drug Carriers; John Wiley & Sons: New York, 1988. (13) Britt, D. W.; Hofmann, U. G.; Mobius, D.; Hell, S. W. Langmuir 2001, 17, 3757-3765. (14) McCullough, D. H.; Regen, S. L. Chem. Commun. 2004, 2787-2791. (15) Morigaki, K.; Kiyosue, K.; Taguchi, T. Langmuir 2004, 20, 7729-7735. (16) Morigaki, K.; Baumgart, T.; Jonas, U.; Offenhaeusser, A.; Knoll, W. Langmuir 2002, 18, 4082-4089. (17) Ross, E. E.; Spratt, T.; Liu, S.; Rozanski, L. J.; O’Brien, D. F.; Saavedra, S. S. Langmuir 2003, 19, 1766-1774. (18) Chapman, D. Langmuir 1993, 9, 39-45. (19) Morigaki, K.; Schonherr, H.; Frank, C. W.; Knoll, W. Langmuir 2003, 19, 6994-7002. (20) Leaver, J.; Alonso, A.; Durrani, A. A.; Chapman, D. Biochim. Biophys. Acta 1983, 732, 210-218. (21) Kolchens, S.; Lamparski, H. G.; O’Brien, D. F. Macromolecules 1993, 26, 398-400. (22) Benz, R.; Elbert, R.; Prass, W.; Ringsdorf, H. Eur. Biophys. J. 1986, 14, 83-92.

10.1021/la052327p CCC: $33.50 © 2006 American Chemical Society Published on Web 12/24/2005

1216 Langmuir, Vol. 22, No. 3, 2006

lipid/protein combinations for immobilization of nanochannels in a polymerized scaffold for the purposes of stochastic sensing. Although a stable scaffold is desirable, there is no utility if the ion channel of interest does not disperse within the scaffold’s lipid matrix. Insertion of a channel into a lipid scaffold requires a delicate balance of lipid and channel properties. Subsequently, it is necessary to not only evaluate improved BLM mechanical properties upon polymerization, but also miscibility of ion channels with the lipid layer. Furthermore, because the lipid polymerization process is sensitive to the organization of the lipid monomers,13 it is not immediately obvious how channel insertion into the membrane will affect lipid polymerization. Hence, it is necessary to confirm that channel insertion does not prohibit polymerization and, subsequently, negate any mechanical improvement due to polymerization. One of the most extensively characterized membrane proteins23-26 is gramicidin A. Hladky and Haydon, in 1970, first reported the single-channel conductance and kinetics of gramicidin channel formation in a phospholipid bilayer membrane.23 Recently, gramicidin has been used for channel-based sensing applications.27 In a novel approach, Cornell and co-workers immobilized multiple gramicidin channels in a bilayer membrane that was tethered onto a solid surface whereby ions could flow into a reservoir beneath the lipid leaflet adjacent to the solid support. Gramicidin is known to form a stable monolayer at the air-water interface28 and to partition into nonpolymerizable lipid bilayers.25 We have recently shown that a different channelforming protein, R-hemolysin, forms ion-conducting channels in a polymerizable lipid scaffold.29 Unlike R-hemolysin, gramicidin forms a gated channel whereby two monomers must selfassemble inside the bilayer to form a pore. Hence, if gramicidin is miscible with polymerizable lipids, it can potentially offer more detailed insight into the effects of polymerization on channel assembly, function, and the lateral mobility of conducting channels. Commercially available polymerizable, diacetylenic lipids are a good starting point to assess miscibility of various lipid structures with ion channels and to evaluate the potential improvement in BLM stability through polymerization. We have previously shown that increasing lipid compressional modulus is directly related to decreasing ease of black lipid membrane formation29 by using commercially available phospholipids identified as follows: DTPC ) 1,2-di-10,12-tricosadiynoyl-sn-glycero-3-phosphocholine; DTPE ) 1,2-di-10,12-tricosadiynoyl-sn-glycero-3-phosphoethanolamine; PTPE ) 1-palmitoyl-2,10,12-tricosadiynoylsn-glycero-3-phosphoethanolamine; and DiPhyPC ) 1,2diphytanoyl-sn-glycero-3-phosphocholine. These lipids have either zero (DiPhyPC), one (PTPE), or two (DTPC, DTPE) polymerizable hydrocarbon tails and either a PC or PE headgroup. The altered compressional modulus for different lipid structures may also affect lipid polymerization kinetics or channel miscibility. Here, we perform a thorough investigation of the polymerization kinetics of the DTPC, DTPE, and PTPE diacetylenic lipids in monolayer arrangement. These lipids provide systematic (23) Hladky, S. B.; Haydon, D. A. Nature 1970, 225, 451-453. (24) Hladky, S. B.; Haydon, D. A. Biochim. Biophys. Acta 1972, 274, 294312. (25) Vogt, T. C. B.; Killian, J. A.; De Kruijff, B.; Andersen, O. S. Biochemistry 1992, 31, 7320-7324. (26) Kolb, H. A.; Bamberg, E. Biochim. Biophys. Acta 1977, 464, 127-141. (27) 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. (28) Ducharme, D.; Vaknin, D.; Paudler, M.; Salesse, C.; Riegler, H.; Mohwald, H. Thin Solid Films 1996, 284-285, 90-93. (29) Shenoy, D. K.; Barger, W. R.; Singh, A.; Panchal, R. G.; Misakian, M.; Stanford, V. M.; Kasianowicz, J. J. Nano Lett. 2005, 5, 1181-1185.

Daly et al.

Figure 1. Molecular structures of polymerizable phospholipid molecules. Note that DTPC and DTPE each have two polymerizable hydrophobic carbon chains while PTPE has a single polymerizable hydrophobic carbon chain.

variation of the number of diacetylene-containing hydrocarbon tails and the lipid headgroup size. These two lipid properties may affect organization of the lipid monolayer and, therefore, the polymerization behavior13 or the miscibility with ion channels. Furthermore, we evaluate the miscibility of the three diacetylenicbased polymerizable lipids with gramicidin A and the effects of gramicidin on lipid polymerization. By forming mixed monolayers of gramicidin and polymerizable lipids at the air-water interface on a Langmuir-Blodgett (LB) trough and analyzing pressure-area isotherms, as we compress the monolayer through its various thermodynamic phases, we demonstrate lipid/ gramicidin compatibility. To evaluate polymerization kinetics of the pure phospholipid monolayers and of those with gramicidin embedded in the lipid monolayer, we monitor the increase in film pressure on the water surface of the LB trough during UV light irradiation. Changes in film pressure at a constant trough area are indicative of a change in the structure of the monolayer. We correlate the kinetics of film pressure changes with the observed absorption peaks in the UV-vis spectra for monolayers of lipid deposited onto solid substrates after exposure to UV light for various lengths of time. In addition, by using a bilayer apparatus, we also measure ion conductance through gramicidin channels formed in a painted bilayer to confirm that gramicidin disperses with polymerizable lipid and to assess the functionality of the gramicidin channel embedded in the diacetylenic lipid support before and after scaffold polymerization. Finally, we demonstrate that polymerization of diacetylenic lipid-containing BLM can increase bilayer stability. Materials and Methods Materials. The synthetic, diacetylene-containing phospholipids, 1,2-di-10,12-tricosadiynoyl-sn-glycero-3-phosphocholine (DTPC, also called DC8,9PC in the literature), 1,2-di-10,12-tricosadiynoylsn-glycero-3-phosphoethanolamine (DTPE), 1-palmitoyl-2,10,12tricosadiynoyl-sn-glycero-3-phosphoethanolamine (PTPE), and the nonpolymerizable phospholipid 1,2-diphytanoyl-sn-glycero-3-phosphocholine (DiPhyPC) were purchased from Avanti Polar Lipids (Alabaster, AL). The structures of the polymerizable lipids are shown in Figure 1. Gramicidin A with greater than 95% purity was purchased from Calbiochem (San Diego, CA). Solvents and dichromate were obtained from Fisher Scientific (Hampton, NJ). All chemicals were used without further purification. Fused silica substrates for lipid deposition were purchased from G. M. Associates (Oakland, CA). Surface Pressure-Area Isotherms. Pressure-area isotherms were measured for monolayers of phospholipids and/or gramicidin

Photopolymerization of Mixed Monolayers and BLMs deposited at the air-water interface by using a NIMA Technology Model 611 MC Langmuir-Blodgett trough (Coventry, England) at 22 °C. The trough was fabricated from poly(tetrafluoroethylene), and the maximum area of the trough was 90 cm2. The subphase was deionized water, further purified by a Millipore triple ion-exchanged system with resistivity ∼18 MΩ cm. A predetermined volume of solution was added dropwise to the clean water surface. After allowing sufficient time for the solvent to evaporate from the air-water interface, pressure-area isotherms of gramicidin A, DTPC, DTPE, PTPE, and DiPhyPC, and of mixtures of each polymerizable lipid containing 0-100 mol % gramicidin, were collected by compressing the barriers at a rate of 10 cm2/min. Polymerization of Diacetylenic Lipid Monolayers. The polymerization behavior of each monolayer of lipid and/or gramicidin was studied at the air-water interface of the trough by exposing the monolayer to UV light from a UVP, Inc. UVS-28 mercury lamp (Upland, CA), whose output wavelength is quoted to be 254 nm. Before UV exposure, the spread films were compressed to a target pressure of approximately 30 mN/m, and the area between the trough barriers was then held constant so that pressure could vary with changes in the film structure. The target pressure was chosen to fall on the steepest part of the pressure-area isotherm (but well below the collapse pressure), so that the molecules were held in close proximity to each other. The volume of lipid solution spread at the interface was controlled so that the area between the barriers at the target pressure was approximately 30 cm2. Once the pressure and area values stabilized, the compressed monolayer was exposed to the UV light source described above. The distance from the lamp to the water surface was approximately 15 cm, and the lamp intensity was 1 mW/cm2. The surface pressure of the monolayer was continuously monitored during the time of UV exposure. To confirm that polymerization of the lipid monolayers was occurring over the time scales indicated in the surface pressure measurements, UV-vis spectroscopy was used. To clean the substrates used for UV analysis, we dipped each substrate into sulfuric acid/dichromate cleaning solution and thoroughly rinsed each substrate with water. Only substrates which exhibited complete wettability with the rinse water were accepted as clean hydrophilic surfaces. To deposit a lipid monolayer onto a solid substrate, a hydrophilic fused silica slide (1 cm × 2 cm) was initially submerged in the subphase before the film was spread, and the monolayer was transferred on the upstroke, after polymerization for a predetermined amount of time. The pressure was held constant during the depositions. The coated substrates were removed and analyzed with a Cary 5000 series UV-vis spectrometer from Varian (Palo Alto, CA). Absorption spectra of DTPC lipid monolayers deposited onto fused silica substrates were collected before any UV exposure and after various times of UV exposure. Gramicidin Insertion into Polymerizable Lipid Bilayers. A bilayer workstation from Warner Instruments (Hamden, CT) was used to monitor bilayer formation across an aperture via the wellknown painted bilayer method30 and subsequent gramicidin insertion into the bilayer. Electrical current data were filtered with a 100-Hz low-pass Bessel filter. To form lipid bilayers, a glass applicator was dipped into a 10 mg/mL solution of PTPE in decane and painted across a 200-µm aperture in a delrin cup. Prior to PTPE bilayer formation, the aperture was prepainted with DiPhyPC in decane, and the delrin cup and chamber were filled with 1 mL of 1 M KCl solution. The capacitance across the thinned bilayer was typically 80-90 pF. The membrane was highly insulating to ionic current, with a typical slope in a current-voltage plot of about 0.01 pA/mV. For channel formation, 1 mL of a gramicidin solution in ethanol was added to both the cis and trans chambers to a final gramicidin concentration of 5 × 10-7 M. A 100-mV dc holding potential was applied across the membrane to monitor gramicidin insertion into the membrane. After channel formation, the holding potential was removed, and the system was exposed to 2 min of UV irradiation. Following this irradiation, the 100-mV holding potential was reapplied to again monitor gramicidin channel conductance. (30) Mueller, P.; Rudin, D. O.; Tien, H. T.; Wescott, W. C. Nature 1962, 194, 979-980.

Langmuir, Vol. 22, No. 3, 2006 1217 Membrane stability was assessed by measuring the lifetime of the PTPE bilayer with 150-mV holding potential applied across the bilayer. Prior to application of the holding potential, membranes were either exposed to 6 min of UV light (for polymerized membranes) or allowed to incubate for 6 min in the dark (for nonpolymerized membranes).

Results and Discussion Gramicidin Dispersion with Polymerizable Lipids. The pressure-area (Π-A) isotherms for pure DTPC, DTPE, PTPE, and DiPhyPC are shown in Figure 2a. At a film pressure of 30 mN/m, the average area/molecule is 50.6, 52.2, and 58.5 Å2/ molecule for DTPC, PTPE, and DTPE, respectively. In contrast to the Π-A isotherms of the phospholipids, the isotherm for gramicidin A is expanded with a plateau around 20 mN/m (Figure 2b, rightmost curve). This result is consistent with pressurearea isotherms for gramicidin reported in the literature.31 The plateau in the gramicidin isotherms likely indicates a reorientation of the gramicidin molecule from the horizontal-to-vertical orientation upon compression of the monolayer film. The area/ molecule values just before the plateau, on the low-pressure side, and just before the film collapse, are consistent with the area occupied by a gramicidin molecule in the horizontal and vertical orientations.32 Analysis of the experimental Π-A isotherm for gramicidin A yields an area/molecule of 130.5 Å2/ molecule for gramicidin at 30 mN/m. The Π-A isotherms of the DTPC, DTPE, or PTPE-gramicidin A mixtures shown in Figure 2b, c, and d, respectively, contain features from both the phospholipid and the gramicidin A isotherm curves. Similar to the feature observed in the isotherm for pure gramicidin, the mixed isotherms all show evidence of a phase change between 14 and 21 mN/m. The measured area per molecule at 30 mN/m reported for each of the lipid/gramicidin mixtures is reported in Figure 3a. Either ideal miscibility or complete immiscibility can account for the linear variations of the molecular areas with the gramicidin mole fraction.33,34 According to Crisp’s rule,33,35 inflection points that occur at constant pressure over the range of the gramicidin mole fraction indicate complete immiscibility. On the other hand, continuous variation of pressure at the inflection point indicates ideal miscibility. Each isotherm shows two inflection points: one at a high pressure (film collapse) and one at a low pressure (gramicidin reconfiguration). Figure 3b shows that the pressure at the low inflection point continuously varies with composition, and therefore, on the basis of Crisp’s rule, DTPC, DTPE, and PTPE are all miscible with gramicidin A. Because this analysis was performed at a relatively high film pressure, it implies that gramicidin is not displaced out of the monolayer during compression. Conductance measurements show that gramicidin A does insert into a PTPE-containing bilayer to form channels (Figure 4). At an applied dc potential of 100 mV, the average duration for which single gramicidin channels remain open is about 0.9 min and the amplitude of the single channel is approximately 2.5 pA. These observations are in good agreement with the literature.25 We were unable to form stable PTPE-painted bilayers without the DiPhyPC prepaint, so although it is likely that the membrane is mostly PTPE, it is possible that some of the bilayer contains (31) Vila-Romeu, N.; Nieto-Suarez, M.; Dynarowicz-Latka, P.; Prieto, I. J. Phys. Chem. B 2002, 106, 9820-9824. (32) Lukes, P. J.; Petty, M. C.; Yarwood, J. Langmuir 1992, 8, 3043-3050. (33) Gaines, G. L. Insoluble Monolayers at Liquid-Gas Interfaces; John Wiley & Sons: New York, 1966. (34) Van Mau, N.; Trudelle, P.; Daumas, P.; Heitz, F. Biophys. J. 1988, 54, 563-567. (35) Crisp, D. J. In “Surface Chemistry”, Supplemental to Research; Butterworth: London, 1949; pp 17, 23.

1218 Langmuir, Vol. 22, No. 3, 2006

Daly et al.

Figure 2. Pressure vs area isotherms of (a) polymerizable phospholipids; DTPC (1), PTPE (2), DTPE (3), and the nonpolymerizable phospholipids DiPhyPC (4) and mixed monolayers of (b) PTPE, (c) DTPE, and (d) DTPC with gramicidin A at the air-water interface during compression on the Langmuir-Blodgett trough. The mole fraction of gramicidin ranges from 0 to 1, with the pure gramicidin curve at the rightmost of each figure and the pure lipid at the leftmost of each figure.

Figure 3. (a) Average area (Å2)/molecule at a surface pressure of 30 mN/m for PTPE (circles), DTPE (diamonds), and DTPC (squares) as a function of gramicidin molar fraction. (b) Surface pressure at the low-pressure inflection point of the isotherms for a mixed monolayer of gramicidin with PTPE, DTPE, and DTPC as a function of the gramicidin molar fraction.

this nonpolymerizable lipid. We cannot be certain that the gramicidin is dispersed within the PTPE lipid molecules of the bilayer rather than within the DiPhyPC molecules. At the least, these bilayer experiments confirm that gramicidin can disperse into a lipid bilayer containing polymerizable diacetylene moieties. Hence, it is feasible that the diacetylene moieties can be polymerized to increase the mechanical stability of the lipid bilayer.

Polymerization of Lipid Monolayers. Figure 5 shows the increase in film pressure at constant area for pure monolayers of DTPC, DTPE, PTPE, and DiPhyPC upon UV irradiation. The baseline surface pressure prior to UV irradiation (∼30 mN/m) is subtracted from each of the surface pressure profiles in Figure 5. The surface pressure, when DiPhyPC, the nonpolymerizable lipid, is spread at the air-water interface on the LB trough, remains constant during exposure to UV radiation, as expected.

Photopolymerization of Mixed Monolayers and BLMs

Figure 4. Ion transport through gramicidin channels inserted into mixed DiPhyPC/PTPE bilayers. The bilayer was painted over a 200-µm aperture in a delrin cup immersed in 1 M KCl solution.

Figure 5. Film pressure change due to UV irradiation for monolayers of DTPC (1), PTPE (2), DTPE (3), and DiPhyPC (4) at the airwater interface with fixed area. Prior to UV irradiation, the film pressure was held constant below the collapse pressure at approximately 30 mN/m.

On the contrary, the surface pressure, when the monolayers are formed by using the polymerizable lipids, exhibits a distinct increase in film pressure. It has been suggested in the literature that a pressure change observed for a monolayer of polymerizable lipid at the air-water interface is due to the alteration of the geometry of the molecules within the film.36 During the polymerization process, diacetylene triple bonds react to form conjugated double bonds within the polymeric linkages between the tails of the molecules. There is a transformation of the structure of the hydrocarbon tails from straight segments containing two acetylene groups that can be nested beside each other in a closepacked film to bent-chain bridges linking neighboring tails. This restricts the motion of the molecules and prevents dynamic alignment and close packing without twisting of some of the chain segments. The pressure variations seen in Figure 4 likely reflect the tilting of the lipid molecule upon the onset of polymerization. In effect, the average area occupied by each molecule in the horizontal plane varies during the polymerization process. Because the total area confining the film is fixed by the barriers of the LB trough, the film pressure changes during molecule tilt and layer reorganization. While the film pressure kinetics in Figure 5 for all of the polymerizable lipids show an increase in film pressure that reaches a maximum and subsequently declines back to the baseline, closer (36) Singh, A.; Gaber, B. P. In Applied BioactiVe Polymeric Materials; Gebelein, C. G., Carraher, C. E., Jr., Foster, V. R., Eds.; Plenum: New York, 1988.

Langmuir, Vol. 22, No. 3, 2006 1219

inspection reveals some differences between the kinetic profiles for the different lipids. The rate and magnitude of film pressure change differs for each polymerizable lipid. The polymerization kinetics for DTPC are the most rapid and produce the largest film pressure change. Polymerization is expected to be more extensive for the lipids with two polymerizable hydrocarbon tails (DTPE and DTPC) as opposed to the lipid with only one polymerizable tail (PTPE) because there are more possible sites for polymerization to occur. The difference in the polymerization kinetics for DTPE and DTPC, both double-diacetylenic-tailed lipids, may be due to the differences in layer packing upon compression of these lipids. The compression isotherms in Figure 2 show that area occupied by each DTPC molecule is smaller than for DTPE during the steepest part of the isotherm. Because polymerization is induced in this steep regime, the close proximity of one molecule to another could enhance the rate of DTPC polymerization and allow for even greater change in the average area per molecule during polymerization-induced restructuring. We also observe that the DTPE film pressure initially passes through a minimum before beginning to increase. The dip may correspond to rearrangements within the DTPE monolayer that improve the packing efficiency. It is well-known that lipid monomer organization or spacing has a significant effect on polymerization of diacetylenes.37 The most favorable monomer spacing involves neighboring monomers with aligned backbone kinks.37,38 Rearrangements that produce a more closely packed layer would decrease the area occupied by each molecule and result in decreased film pressure. We likely observe a film pressure minimum for only DTPE and not DTPC or PTPE because of the differences in monolayer packing during film compression for each of the lipids. One might expect layer packing to be less efficient for lipid layers with asymmetric hydrocarbon tails. Hence, the PTPE lipid might be expected to be less closely packed than either the DTPE or DTPC lipid monolayers. This expectation holds true only for the DTPC isotherm relative to the PTPE isotherm. Furthermore, the phosphoethanolamine headgroup is smaller than the phosphocholine headgroup, so it might be expected that the DTPE isotherm would be shifted toward even smaller area/molecule values than the DTPC isotherm. Instead, the area occupied by each DTPE molecule is larger than for DTPC and PTPE during the steepest part of the isotherm (see Figure 2a). This suggests that the DTPE molecules do not pack efficiently during compression and that rearrangements may be necessary before polymerization can take place. The onset of UV irradiation may trigger the molecular reorganization required for improved layer packing. The rearrangements that lead to a more ordered DTPE monolayer will decrease the area occupied by each DPTE molecule and, consequently, decrease the film surface pressure at a constant trough area. Because this dip does not occur for the DTPC and PTPE polymerization processes, it is possible that these molecules pack efficiently upon compression so UV irradiation causes an immediate increase in the area occupied by each molecule (or film pressure) as the molecules tilt during polymerization. To confirm the hypothesis that the film pressure increase corresponds to lipid polymerization, we collected absorbance spectra at different times of irradiation. Photopolymerization of diacetylene-containing lipids has been shown to produce both red and blue species of polymer with absorbance bands between 600 and 700 nm and 450-550 nm, respectively.16 Morigaki and co-workers report formation of a blue polymer upon photopo(37) Carpick, R. W.; Sasaki, D. Y.; Marcus, M. S.; Eriksson, M. A.; Burns, A. R. J. Phys.: Condens. Matter 2004, 16, R679-R697. (38) Mowery, M. D.; Menzel, H.; Cai, M.; Evans, C. E. Langmuir 1998, 14, 5594-5602.

1220 Langmuir, Vol. 22, No. 3, 2006

Daly et al.

Figure 6. Absorbance spectra for DTPC deposited on glass substrates after various times of UV exposure. DTPC monolayers were spread at the air-water interface and compressed to approximately 30 mN/m before UV exposure. Glass slides were immersed in the LB trough subphase during UV irradiation and the lipid was deposited by pulling the slide out of the solution. The time of UV exposure prior to deposition and UV analysis is indicated on the plot. The times at which UV spectra were recorded correspond to (a) increasing film pressure during polymerization and (b) decreasing film pressure after pressure passes through a maximum.

lymerization of lipid vesicles, followed by a red absorbance shift upon heating of the polymer.16 The blue-to-red transition is generally attributed to a structural relaxation of the polymer and a reduction in the conjugation length of the polymer backbone.15 Furthermore, diacetylene polymerization is known to be highly dependent on lipid organization or packing.39-41 Polymerization behavior of lipid monolayers or bilayers on solid supports has been shown to be dependent on the substrate.13,16 Direct formation of the red polymer has been reported at both the air-water42 and solid-air interfaces.13,16 We have also observed this color change at the water surface of an LB trough when a DTPC monolayer was irradiated with UV light. The absorbance spectra for a monolayer film of DTPC that was polymerized at the air-water interface of an LB trough and then transferred onto glass substrates are shown in Figure 6. As expected, a DTPC monolayer that was transferred onto a quartz substrate without being exposed to UV light shows no discernible absorbance peak. The spectrum of a polymerized monolayer contains an absorbance peak with a maximum at approximately 470 nm. Figure 6a shows absorbance spectra at times corresponding to the increase in film pressure for DTPC (see Figure 5). The intensity of the absorbance peak increases during this time, suggesting that the extent of polymerized DTPC in the monolayer increases over the first 13.3 min of UV exposure. Figure 6b shows absorbance spectra at times corresponding to the decrease in film pressure shown in Figure 5. The intensity of the absorbance peak decreases during this time and eventually becomes undetectable, suggesting that polymer degradation occurs when UV exposure times are larger than 13.3 min. Others have also observed decomposition of diacetylenic polymers upon excessive irradiation.42,43 Polymer degradation at long exposure times is most likely due to local heating of the film, resulting in covalent bond breakage. Figure 7 shows the absorbance maximum in each of the spectra from Figure 6 overlaid with the film pressure kinetics for a DTPC (39) Jonas, U.; Shah, K.; Norvez, S.; Charych, D. H. J. Am. Chem. Soc. 1999, 121, 4580-4588. (40) Charych, D. H.; Nagy, J. O.; Spevak, W.; Bednarski, M. D. Science 1993, 261, 585-588. (41) Shostakovskii, M. F.; Bogdanova, A. V. The Chemistry of Diacetylenes; John Wiley & Sons: New York, 1974. (42) Alekseev, A. S.; Vitala, T.; Domnin, I. N.; Koshkina, I. M.; Nikitenko, A. A.; Peltonen, J. Langmuir 2000, 16, 3337-3344. (43) Ogawa, K. J. Phys. Chem. 1988, 93, 5305-5310.

Figure 7. DTPC absorbance peak intensity as a function of the time of UV exposure (open squares) overlaid with the film pressure kinetic (solid line) data.

monolayer. The good agreement between the shapes of these kinetic traces indicates that lipid polymerization and depolymerization is responsible for film pressure changes during UV irradiation. Furthermore, there appears to be a small time window during which the extent of polymerization in DTPC monolayers is at a maximum. This suggests that there might be an optimum time to expose channel-containing diacetylenic phospholipid bilayer membranes to UV light such that the mechanical stability of the lipid scaffold is maximized. Absorbance spectra at the film pressure maxima were also collected for monolayers of DTPE and PTPE deposited onto the fused silica substrates. An absorbance peak at approximately 470 nm is present for the lipid films that had been subjected to 15 min of UV irradiation, but no peaks were observed for films that were not irradiated. Although we did not collect the complete absorbance kinetic profile for DTPE and PTPE, this limited absorbance data confirms that each of these lipids undergoes polymerization during the time corresponding to the increase in the film pressure, shown in Figure 4. Polymerization of Mixed Lipid/Gramicidin Monolayers. As discussed above, the Π-A isotherms showed evidence of gramicidin dispersion within the polymerizable lipid monolayer at the air-water interface. To examine the effect of gramicidin

Photopolymerization of Mixed Monolayers and BLMs

Langmuir, Vol. 22, No. 3, 2006 1221

Figure 9. A histogram depicting the lifetime of DiPhyPC/PTPE bilayers before (gray) and after (black) 6 min of UV irradiation. Figure 8. Film pressure change due to UV irradiation for mixed monolayers of 90% DTPC (1), PTPE (2), or DTPE (3) with 10% gramicidin at the air-water interface with fixed area. Prior to UV irradiation, the film pressure was held constant below the collapse pressure at approximately 30 mN/m.

insertion on the polymerizability of the lipid monolayer, we exposed mixed films to UV light and monitored surface pressure changes at a constant area. The film pressure increased upon exposure to UV light, indicating polymerization in the mixed monolayers, as shown in Figure 8. For all lipids, the kinetics of polymerization were slower with the mixed monolayers than with monolayers of pure lipid. This makes sense if we consider that gramicidin likely disrupts the packing in the lipid films and subsequently slows polymerization. The magnitude of film pressure change also decreases with gramicidin in the monolayer for all the polymerizable lipids, as might be expected on the basis of the decreased lipid composition of the monolayer. The total pressure change during polymerization of the singlediacetylenic-tailed PTPE is affected the least by gramicidin insertion. These polymerization results suggest that the phospholipids can be polymerized as protein-containing monolayers. Effects of Polymerization on BLM Stability and Gramicidin Function. Because the LB work shows that gramicidin is not squeezed out of diacetylenic lipid monolayers at high pressure, and that its incorporation into a monolayer does not prohibit polymerization, these polymerizable lipids might, therefore, have utility for stabilizing protein channel-containing bilayers for stochastic sensing applications. To examine improvements in scaffold mechanical stability upon polymerization, we measured the lifetime of DiPhyPC/PTPE bilayers with and without 6 min of UV irradiation. This duration is less than half the UV exposure time that induces significant polymer degradation in the monolayer arrangement. Following the 6-min wait time (or the 6-min UV irradiation period), a 150-mV holding potential was applied across the bilayer. Figure 9 shows the lifetime of 20 different bilayers, half of which underwent polymerization. The average lifetime with and without UV irradiation is 30.7 min and 12.6 min, respectively. It is possible that the amount of DiPhyPC in the BLM varies from membrane to membrane and contributes to the wide fluctuation in lifetime of the UV irradiated BLMs. Nonetheless, the increased average lifetime suggests that polymerization of lipid in the black lipid membrane is a promising technique to improve stability. Because our membrane formation technique might result in some DiPhyPC prepainted lipid incorporated into the PTPE bilayer, we monitored the polymerization kinetics of a 50:50 DiPhyPC/PTPE mixture in monolayer arrangement to verify that DiPhyPC would not inhibit PTPE polymerization. As shown in Figure 10, the PTPE molecules

Figure 10. Film pressure change due to UV irradiation for monolayers of PTPE (1), DiPhyPC (2), and mixed 50:50 mole fraction DiPhyPC/PTPE (3) at the air-water interface with fixed area. Prior to UV irradiation, the film pressure was held constant below the collapse pressure at approximately 30 mN/m.

do undergo polymerization, even with a large amount of DiPhyPC incorporated into the monolayer, indicating that some PTPE polymerization is feasible, even in the unlikely event that a large amount of DiPhyPC is incorporated into the BLM. The improved bilayer stability may be a result of polymerization-induced restructuring of the lipid toward a more favorable packing arrangement. LB trough work confirmed the rearrangement of these diacetylenic lipids upon polymerization. We did observe a relatively long lifetime of one bilayer not exposed to UV irradiation, suggesting that optimal packing of the lipid layer may be achieved in rare cases without lipid polymerization, but that the polymerization process improves the probability of optimal layer packing. We did not perform stability analysis of 2-diacetylenic-tailed lipids because these membranes were prohibitively difficult to form. Shenoy and co-workers also observed that rigid lipid monolayers with a high compressional modulus did not readily form black lipid membranes.29 Nonetheless, the improvements in stability with the single-diacetylenictailed lipid suggest that a critical balance between membrane fluidity and rigidity may be required to optimize bilayer stability. Furthermore, while this cross-chain diacetylenic polymerization improves bilayer stability, polymerization in this painted bilayer arrangement does not eliminate the torus that is responsible for inherent instability in these membranes. Hence, new approaches to permanently stabilize a polymerizable lipid film across an

1222 Langmuir, Vol. 22, No. 3, 2006

Daly et al.

a highly cross-linked network of polymer. Nonetheless, the improvements in membrane stability combined with the preserved protein function after some scaffold polymerization suggest that more extensive polymerization may lead to even more robust membranes with immobilized channels. However, it might be necessary to work with channels insensitive to larger UV irradiation doses, such as synthetic ion channels or a gramicidin analogue in which all tryptophan residues have been replaced by photostable phenylalanine residues.45 Because it is very difficult to form bilayers of two-diacetylenic-tailed commercially available lipids, synthesis of new polymerizable lipid materials may be necessary for these additional improvements. Figure 11. Ion transport through gramicidin channels inserted into mixed DiPhyPC/PTPE bilayers after 2 min of UV irradiation. The bilayer was painted over a 200-µm aperture in a delrin cup immersed in 1 M KCl solution.

aperture that eliminate the torus may lead to even more dramatic increases in bilayer lifetime. We also observe that gramicidin channels remain active upon UV irradiation of the gramicidin containing a DiPhyPC/PTPE bilayer, as shown in Figure 11. This confirms that the tilting of the lipids upon polymerization does not squeeze gramicidin out of the lipid bilayer nor does our low UV irradiation dose inactivate gramicidin. McKim and Hinton have shown that the tryptophan residues in gramicidin undergo UV photolytic degradation over 24 h during exposure to a 75-W UV lamp.44 The lamp intensity in our experiments is smaller than that used by McKim and Hinton, and our UV exposure time is only two minutes. Hence, the preservation of gramicidin activity after UV irradiation in our experiments is feasible because of our small UV dose, and longer or more intense UV light exposure might alter the gramicidin channel structure and activity. In our experiments, the channel conductance is not altered by the UV light irradiation, but frequency of gramicidin channels increases following UV irradiation, as shown in Figure 10. Our polymerization experiments were conducted with gramicidin monomers in solution during UV irradiation. It is possible that the increase in channel activity is a result of increased gramicidin insertion into the bilayer during the polymerization-induced restructuring of the diacetylenic lipids. The gramicidin channel activity after UV irradiation reveals that the DiPhyPC/PTPE bilayer maintains fluidity even after some polymerization. This is expected from this bilayer containing some nonpolymerizable lipid. Furthermore, PTPE has only one polymerizable tail, so is not capable of forming (44) McKim, S.; Hinton, J. F. Biochim. Biophys. Acta 1993, 1153, 315-321.

Conclusion Herein, we showed that gramicidin molecules can be dispersed within a polymerizable lipid monolayer and that the pure and mixed diacetylene-containing lipid monolayer can undergo polymerization upon exposure to UV light. Hence, polymerizable lipids are a potential stable scaffold for ion channels and stochastic sensors. Moreover, a fundamental understanding of the polymerization kinetics was gained and a crossover from polymerization to depolymerization behavior observed. This suggests an optimum time for UV exposure. Awareness of this optimum polymerization time is crucial in the development of stochastic sensing scaffolds because long irradiation times will degrade the polymerized scaffold. The improvement in the long-term stability of the black lipid membranes upon diacetylene lipid polymerization is extremely promising and suggests that alternate schemes for polymerization of lipid may result in additional improvements in bilayer lifetime. Furthermore, the ion-channel activity of gramicidin following UV irradiation is promising because it suggests that UV irradiation does not squeeze the ion channel out of the bilayer. In the future, more extensive polymerization schemes might decrease the lateral diffusion of the ion channel and essentially lock the channel into place over an aperture. Acknowledgment. We gratefully acknowledge support from the DARPA DSO office through the Moldice Program and the Office of Naval Research. S.D. acknowledges support from a National Research Council research associateship award at the Naval Research Laboratory. L.H. is currently affiliated with the Medical University of South Carolina. LA052327P (45) Busath, D. D.; Hayon, E. Biochim. Biophys. Acta 1988, 944, 73-78.