Optical Trapping of Unilamellar Phospholipid Vesicles: Investigation of

Anal. Chem. 2004, 76, 4920-4928

Optical Trapping of Unilamellar Phospholipid Vesicles: Investigation of the Effect of Optical Forces on the Lipid Membrane Shape by Confocal-Raman Microscopy Daniel P. Cherney, Travis E. Bridges, and Joel M. Harris*

Department of Chemistry, University of Utah, 315 South 1400 East, Salt Lake City, Utah 84112-0850

Optical trapping of liposomes is a useful tool for manipulating these lipid vesicles for sampling, mechanical testing, spectroscopic observation, and chemical analysis. Through the use of confocal Raman microscopy, this study addresses the effects of optical forces on the structure of unilamellar, dipalmitoylphosphatidylcholine (DPPC) vesicles, both optically trapped in solution and adhered to a coverslip. The energy and forces involved in optical trapping of lipid vesicles were derived in terms of the dielectric contrast between the phospholipid membrane and the surrounding solution; reflection forces at the membrane/water interface were found to be negligible. At optical powers of 9 mW and greater, unilamellar liposomes trapped in bulk solution experience a gradient force sufficiently strong to bend the vesicle membrane, so that a second bilayer from the same vesicle is drawn into the optical trap, with an energy of ∼6 × 10-13 erg. For vesicles adhered to a coverslip, the confocal probe can be scanned through the attached vesicle. Optical forces are insufficient to detach the bilayer that is adhered to the glass; however, the upper DPPC bilayer can be manipulated by the optical trap and the shape of the vesicle distorted from a spherical geometry. The effect of calcium ion on the flexibility of membrane bilayers was also tested; with 5 mM calcium ion in solution, the lipid bilayer of a surface-attached liposome is sufficiently rigid so that it cannot be distorted at moderate laser powers.

Since the development of optical trapping on dielectric spheres by Ashkin in the early 1970s,1 the popularity of the “optical tweezers” technique has grown considerably and has recently been extended to the analysis of biological samples such as viruses, bacteria, chromosomes, and nerve cells.2-5 The tool has evolved from the two-laser beam arrangement, utilized in Ashkin’s * To whom correspondence should be addressed. E-mail: harrisj@ chemistry.chem.utah.edu. (1) Ashkin, A. Phys. Rev. Lett. 1970, 24, 156-159. (2) Ashkin, A.; Dziedzic, J. M. Science 1987, 235, 1517-1520. (3) Ashkin, A.; Dziedzic, J. M.; Yamane, T. Nature 1987, 330, 769-771. (4) Berns, M. W.; Wright, W. H.; Tromberg, B. J.; Profeta, G. A.; Andrews, J. J.; Walter, R. J. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 1539-4543. (5) Dai, J.; Sheetz, M. P. Biophys. J. 1995, q68, 988-996.

4920 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

first study,1 to a simpler, single-laser beam trap.6-8 It has been previously demonstrated that phospholipid vesicles, or liposomes, can be optically trapped with a single focused laser beam and then held or manipulated for sampling, mechanical testing, spectroscopic observation, and chemical analysis.9-11 Studies of the forces exerted on particles within optical traps have focused primarily on solid, spherical particles but, however, have not addressed the trapping of liposomes. Unlike solid particles, the thin, bilayer membranes of phospholipid vesicles can be deformed and changes in their structure under the influence of optical forces should be considered in designing an analysis method based on optical tweezers methods. The forces responsible for the trapping of solid, homogeneous particles greater than the wavelength of the radiation (typically several micrometers in diameter) are primarily due to the refractive index difference between the particle and the surrounding solution leading to reflections and refractions of the light at the particle/solution interfaces.1,6 Reflections produce a force at the particle/solution interface due to the change in the photon direction and momentum caused by the difference in the indices of refraction of the sphere and the surrounding solution. Refraction can also produce a force on the particle due to the change in the direction, and thus momentum, of the photons due to transmission through the particle/solution interface. Conservation of momentum in the system dictates that the changes in momentum of the reflected and refracted photons are canceled by equal and opposite forces on the particle within the trap. Particles much smaller than the wavelength of light can also be optically trapped (even atoms and large molecules),12-14 and the reflection/refraction model of trapping forces does not apply to samples of this small size. Several studies have demonstrated that the forces primarily responsible (6) Ashkin, A. IEEE J. Quantum Electron. 2000, 6, 841-856. (7) Wright, W. H.; Sonek, G. J.; Tadir, Y.; Berns, M. W. IEEE J. Quantum Electron. 1990, 26, 2148-2157. (8) Stout, A. L. Biophys. J. 2001, 80, 2976-2986. (9) Ichikawa, M.; Yoshikawa, K. Appl. Phys. Lett. 2001, 79, 4598-4600. (10) Chiu, D. T.; Wilson, C. F.; Ryttse´n, F.; Stro ¨mberg, A.; Farre, C.; Karlsson, A.; Nordholm, S.; Gaggar, A.; Modi, B. P.; Moscho, A.; Garza-Lo´pez, R. A.; Orwar, O.; Zare, R. N. Science 1999, 283, 1892-1895. (11) Chiu, D. T.; Hsiao, A.; Gaggar, A.; Garza-Lo´pez, R. A.; Orwar, O.; Zare, R. N. Anal. Chem. 1997, 69, 1801-1807. (12) Ashkin, A. Phys. Rev. Lett. 1978, 40, 729-732. (13) Svoboda, K.; Block, S. M. Opt. Lett. 1994, 19, 930-932. (14) Chaloupka, J. L.; Fisher, Y.; Kessler, T. J.; Meyerhofer, D. D. Opt. Lett. 1997, 22, 1021-1023. 10.1021/ac0492620 CCC: $27.50

© 2004 American Chemical Society Published on Web 07/29/2004

for holding neutral atoms, molecules, and particles in the Rayleigh limit (r , λ) in an optical trap were due to the dipole gradient force ∼1/2 R3E2.15-17 This development of trapping forces for large and small particles has not, however, been applied to the trapping of inhomogeneous samples, typical of biological particles. There have been several studies that use an optical trap to measure forces induced either on objects within the trap or between an object in the trap and an object outside the trap. When dielectric objects smaller than the beam waist of the trap were held in the optical tweezers, interactions between aggregates of the dielectric matter within the trap could be observed.18 An optical tweezers experiment has also been used to characterize the strength of noncovalent protein/protein interactions between molecules surface-linked to polystyrene and glass.8 In this experiment, an immunoglobulin G (IgG)-coated polystyrene microsphere was held in the optical tweezers and brought to a surface-bound Staphylococcus protein A (SpA) in order to characterize the strength of the IgG-SpA bond. Despite the lack of data in the literature on the magnitude of optical forces acting on lipid vesicles, lipid bilayers, or other small biological structures, there have been numerous successful examples of optical trapping and manipulation of biological samples including synaptosomes, red blood cells, myosin, and DNA molecules reported in the literature where the small size of the object in the trap is relevant.19-22 Although these studies simply used tweezers as a means to manipulate or hold a sample in place while another analysis was performed, there was no mention of possible effects due to the optical forces acting on the sample in the trap. There have been observations of phospholipid vesicles in optical traps where smaller vesicles are expelled from a multilamellar lipid membrane.23,24 It was postulated that the expulsion was due to internal pressurization of a vesicle in the optical trap; however, the postulated change in internal pressure was not compared with the rapid permeability coefficient of water across a lipid membrane, on the order of 4.7 × 10-6 cm/s for dipalmitoylphosphatidylcholine (DPPC) in the gel phase.25 The “pearling” of long, tubular vesicles in an optical trap has also been reported,26-29 and causes of the morphology changes have been attributed to effects of optical forces on the membrane in the trap. A confocal Raman microscope30 can be a useful tool to study the interaction of optical forces with the lipid bilayer of an optically (15) Harada, Y. Asakura, T. Opt. Commun. 1996, 124, 529-541. (16) Smith, P. W.; Ashkin, A.; Tomlinson, W. J. Opt. Lett. 1981, 6, 284-286. (17) Ashkin, A.; Dziedzic, J. M.; Bjorkholm, J. E.; Chu, S. Opt. Lett. 1986, 11, 288-290. (18) Burns, M. M.; Fournier, J.-M.; Golovchenko, J. A. Phys. Rev. Lett. 1989, 63, 1233-1236. (19) Ajito, K.; Torimitsu, K. Lab Chip 2002, 2, 11-14. (20) Svoboda, K.; Schmidt, C. F.; Branton, D.; Block, S. M. Biophys. J. 1992, 63, 784-793. (21) Finer, J. T.; Simmons, R. M.; Spudich, J. A. Nature 1994, 368, 113-119. (22) Perkins, T. T.; Quake, S. R.; Smith, D. E.; Chu, S. Science 1994, 264, 822826. (23) Bar-Ziv, R.; Frisch, T.; Moses, E. Phys. Rev. Lett. 1995, 75, 3481-3484. (24) Menger, F. M.; Gabrielson, K. J. Am. Chem. Soc. 1994, 116, 1567-1568. (25) Deamer, D. W.; Bramhall, J. Chem. Phys. Lipids 1986, 40, 167-188. (26) Bar-Ziv, R.; Moses, E. Phys. Rev. Lett. 1994, 73, 1392-1395. (27) Nelson, P.; Powers, T.; Seifert, U. Phys. Rev. Lett. 1995, 74, 3384-3387. (28) Granek, R.; Olami, Z. J. Phys. II (Fr.) 1995, 5, 1349-1370. (29) Bar-Ziv, R.; Menes, R.; Moses, E.; Safran, S. A. Phys. Rev. Lett. 1995, 75, 3356-3359. (30) Houlne, M. P.; Sjostrom, C. M.; Uibel, R. H.; Kleimeyer, J. A.; Harris, J. M. Anal. Chem. 2002, 74, 4311-4319.

trapped vesicle,31 because the laser used to trap the sample can also provide Raman scattering analysis on the material in the trap. The confocal optical design provides submicrometer spatial selectivity,32 which allows the contents of a vesicle to be distinguished from the surrounding solution.31 This combination of attributes allows the dependence of the trap forces acting on a lipid bilayer to be analyzed. In the present work, we have utilized confocal Raman microscopy to examine the effect of optical forces on single, unilamellar liposomes that are optically trapped in solution (∼3 µm in diameter) and adhered to a glass coverslip (∼5 µm in diameter). The laser power- and distance-dependent data suggest changes in shape of the lipid membrane arise from drawing the membrane into the center of the trap. The results are compared with an estimate of the lowering of the energy of the system as a result of trapping a lipid bilayer and are compared with the expected energies required for trapping and for bending the membrane out of a spherical shape. THEORY Since the thickness of the phospholipid bilayer is much smaller than the wavelength of the laser light, 5 nm compared to 647.1 nm, the electric field across the lipid bilayer can be approximated as being constant, analogous to the treatment of a Rayleigh particle in an optical trap.33 With this assumption, two forces act on a lipid bilayer in an optical trap arising from (1) the gradient force caused by a reduction in total free energy when more polarizable phospholipid molecules replace water in the center of the strong optical field and (2) the reflection of the incoming light from the interfaces of the thin film. Refraction force is negligible because the water/lipid interfaces are parallel and the path length through the bilayer is small; the transmitted rays, therefore, do not deviate in direction preserving their original momentum. Reflection of radiation from the water/lipid interfaces can produce an upward force on the vesicle. The fraction of incoming light reflected by the water/lipid bilayer interfaces can be predicted from the Fresnell equations, which for reflectance from a thin film is given by34

R)

[(r12)2 + (r21)2 + (2r12r21 cos(β))] [1 + (r12)2(r21)2 + (2r12r21 cos(β))]

(1)

where r12 ) [(n1 - n2)/(n1 + n2)], n is the index of refraction (1 ) water and 2 ) lipid), r21 ) -r12, and β is the optical thickness of the sample, defined as34

β ) 4πn2h/λ

(2)

where h and n2 are the thickness and refractive index of the bilayer and λ is the wavelength of the radiation. Using values for a phospholipid film that is 5 nm thick, the reflectivity is estimated to be R ) 4.7 × 10-5. This value allows for the force from the optical reflection of the lipid bilayer to be determined from the (31) Cherney, D. P.; Conboy, J. C.; Harris, J. M. Anal. Chem. 2003, 75, 66216628. (32) Bridges, T. E.; Houlne, M. P.; Harris, J. M. Anal. Chem. 2004, 76, 576584. (33) Bartlett, P.; Henderson, S. J. Phys.: Condens. Matter 2002, 14, 7757-7768. (34) Born, M.; Wolf, E. Principles of Optics; Pergamon: New York, 1980.

Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

4921

change in momentum of light striking the bilayer. The change in the photon momentum leads to a reflection force given by1

Fs ) 2PRL/c

(3)

where P is the power of the incoming light, R is the reflectance, c is the speed of light, and L is the average z component of the radiation from the objective reaching the sample, which can be determined by

L)

2π

∫

67.5

0

2π

cos(θ) sin(θ) dθ

∫

67.5

0

(4)

sin(θ) dθ

and has a value of 0.69, compared to unity if the light were collimated and at normal incidence to the bilayer. The limits of integration given by the half-angle of the trap, 67.5°, span the angular range of light coming into the center of the beam waist formed by the 1.4 NA objective. With an incident laser power of 30 mW, the greatest used in this study, the reflection force on the bilayer is estimated to be 6.5 × 10-15 N, which is much smaller than the gradient force (see below). Calculation of the gradient force requires the derivative of the free energy lost by trapping the lipid bilayer. This free energy lost by displacing water from an optical trap with the more polarizable lipid bilayer is given by27

∆G ) -Uδaoh

(5)

where U is the laser energy density, δ is the dielectric contrast between water and the phospholipid with a value of 0.23,27 ao is the illuminated area of the lipid bilayer, and h is the bilayer thickness. Due to tight focusing of the laser beam by the microscope objective, the optical energy density and the illuminated amount of the bilayer vary with distance from the center of the focus. This also makes the free energy change depend on this distance as well, giving

∆G(z) ) -U(z)δao(z)h

(6)

where the distance-dependent laser energy density, U(z), is32

U(z) ) P/πcw2(z)

(7)

where P is the laser power, w(z) is the radius of the beam profile in the xy plane in the confocal probe, and c is the speed of light. The radius of the beam profile size is defined as32

w2(z) ) w2o[1 + (z/zc)2]

(8)

where z is the distance away from the beam waist and zc is the confocal distance, which is defined as

zc ) πω2o/λ

Figure 2. Gradient optical force on a phospholipid bilayer. This is an attractive force toward the center of the optical trap. The x axis corresponds to the distance from the center of the trap in the z direction.

confocal probe volume. The distance-dependent free energy given by eq 6 is plotted in Figure 1. By locating a lipid bilayer film at the center of the laser focus, the free energy of the system is lowered by 4.2 × 10-20J or ∼11 kT at a laser power of 6 mW. Taking the derivative of the free energy change with respect to z (the direction of light propagation), the gradient force is described by

Fg(z) ) δh(d/dz)ao(z)U(z)

(10)

An evaluation of this equation can be found in Figure 2. The maximum force is found at 0.34 µm from the center of the trap, or zc/x3, corresponding to where the second derivative of the beam profile size goes to zero. The maximum gradient force was calculated to be 5.1 × 10-14 N for a laser power of 6 mW and 2.6 × 10-13 N at 30 mW. Note, this is >40 times greater than the reflection force under the same illumination conditions, so that the reflection force can be neglected. The values of gradient force and energy will be used to interpret the experimental observations of bilayer deformation and compared to energies reported in the literature for this phenomenon.

(9)

where ωo is the radius of the beam waist at the center of the 4922

Figure 1. Free energy gradient created by the optical trap with 6-mW laser power. The x axis of the graph corresponds to the distance from the center of the optical trap in the z direction. The horizontal line at -4.1 × 10-21 erg corresponds to kT at room temperature. Where the trapping energy exceeds kT in magnitude, a phospholipid bilayer should be trapped at this laser power.


EXPERIMENTAL SECTION Reagents and Materials. 1,2-Dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) was purchased as a powder from Avanti Polar

Lipids (Alabaster, AL) and dissolved in chloroform purchased from Sigma (St. Louis, MO); the lipid solution was used without further purification. An extruder was also purchased from Avanti Polar Lipids. Polycarbonate membranes with a pore size of 5.0 µm were made by Nucleopore (Pleasanton, CA). Polystyrene and Tris(hydroxymethyl)aminoethane were purchased from Aldrich (Milwaukee, WI). Tetramethylammonium perchlorate was purchased from Sigma. Sodium phosphate (dibasic), hydrochloric acid and sodium hydroxide were purchased from Mallinckrodt (Paris, KY). Calcium chloride was purchased from Fischer (Fair Lawn, NJ), and coverslips were purchased from Fisher Scientific (Pittsburgh, PA). All water used was doubly distilled and then filtered with a Barnstead (Boston, MA) NANOpure II and had a minimum resistivity of 18.2 MΩ‚cm. Confocal Raman Microscope. The Raman microscope is based on a previous design30 with minor modifications. Briefly, the 647.1-nm line from a Kr+ laser (Innova 90, Coherent Inc.) with an output power of up to100 mW was sent via a series of mirrors through a band-pass filter (F10-647.1-4, CVI Laser Corp.) and a 4× beam expander (model 50-25-4X-647, Special Optics Inc.) mounted on the back of a Nikon TE 300 inverted fluorescence microscope. The expanded beam was sent through the rear aperture of the microscope, through a another band-pass filter (D647/10, Chroma Tech Inc.) and dichroic beam splitter (655DCLP, Chroma Tech Inc.). The excitation light was then directed through a 100×, 1.4 NA oil immersion microscope objective (CFL PLAN APO, Nikon Inc.) and focused to a 0.60-µm-diameter spot inside the flow cell. The Raman scattering was collected by the same objective and passed through the dichroic beam splitter, a longpass filter (E660LP, Chroma Tech Inc.), and a holographic notch filter (Kaiser) and focused onto the 50-µm entrance slit of a monochromator (250IS, Chromex Inc.), which defined the confocal volume in the horizontal dimension. A charge-coupled device (CCD) camera (DV420, Andor Inc.) was used to collect the spectrum over a 3-pixel or 66-µm vertical region, which defined the confocal volume in the vertical dimension. Bright-field illumination from a 0.52 NA overhead lamp assembly and a digital camera (CoolPix 950, Nikon, Inc.) were employed to acquire images of the liposomes. The laser power entering the objective was measured after the dichroic beam splitter through an aperture that matched the entrance pupil of the objective using a Scientech model AA30 power meter. Sample Preparation. Thin Polystyrene Film. A polystyrene film was dip-coated onto a coverslip by withdrawing it from a chloroform solution with a concentration of 2.17 mg/mL polystyrene. The following equation was used to estimate the thickness, l, of the film:35

l ) 0.944(ηv)2/3/(Fg)1/2γ1/6

(11)

where η is the viscosity of the solution, v is the rate of withdrawal from the solution, F is the density of the solution, g is the acceleration due to gravity, and γ is the surface tension of the solution. With a withdrawal rate of 1 cm/s, the film is predicted to be ∼10 nm in thickness. Liposomes. One milligram of DPPC/chloroform solution was transferred to a 15-mL vial and dried under a stream of nitrogen (35) Landau, L.; Levich, B. Acta Physiochim. U.R.S.S. 1942, 17, 42-54.

for 15 min. The resulting film was placed under vacuum overnight to remove any trace quantities of chloroform. The dried lipid film was hydrated for 1 h above the lipid transition temperature (41 °C) with 1 mL of an aqueous solution containing 50 mM phosphate buffer (pH 7.0) with 50 mM NaCl or with a 5 mM CaCl2 solution with 92.5 mM NaCl for additional ionic strength. Approximately 100 µL of the hydrated lipid suspension was extruded several times through a polycarbonate membrane with a 5-µm pore size to produce large unilamellar vesicles.36-39 Following extrusion, the samples were diluted by a factor of 100 with the same solution used for hydrating the particular sample in order to introduce a number of liposomes into the flow cell and minimize potential liposome/liposome interactions. The samples were injected into a flow cell constructed in-house30 having a coverslip for a lower window and allowed to settle for ∼30 min. Single liposomes were located visually using bright-field illumination. Once several liposomes had adhered to the surface, thereby immobilizing them, a solution containing 50 mM phosphate buffer (pH 7.0) and 50 mM perchlorate ion was slowly injected into the flow cell, passing over the adhered liposomes, which had not been exposed to calcium during sample preparation. Spectral Collection and Analysis. The xy location of the laser trap could be changed by manually manipulating the position of the sample stage. The location of the trap in the z direction within the flow cell was changed by manually turning the fine-focus knob on the side of the microscope. The objective was moved in steps of 0.5 µm for analysis. The laser power directed into the objective for depth scans was either 6 or 30 mW. The spectra were collected with an entrance slit width of 50 µm and an integration time of 3 min. All spectra were collected at 23.5 ( 0.5 °C. In addition to the Raman spectra of the lipids, a spectrum was acquired of the tungsten source from the overhead bright-field illuminator to remove the wavelength-dependent modulations in the detector sensitivity. The Raman spectra were flat-field corrected for the CCD wavelength-dependent response by taking their ratio to that of the spectrum of the tungsten source. Because the tungsten source exhibited a falloff in intensity between 750 and 800 nm, this effect was corrected by fitting the tungsten source spectrum to a fourth-order polynomial in MATLAB (Mathworks, Natick, MA) and multiplying the flat-fielded spectrum by this smooth polynomial to remove the intensity bias at longer wavelengths. RESULTS AND DISCUSSION Effects of Optical Forces on an Optically Trapped Liposome in Solution. The optical forces on a liposome membrane (see eqs 6-8) depend on the energy density of the incident light, U(z), which in turn depends on its distance away from the center of the optical trap, z. As the distance between the bilayer and the center of the trap, z, increases, so does the effective spot size, w(z) (eqs 7 and 8), which exposes the bilayer to a lower energy density. In these first experiments, individual DPPC liposomes of ∼3-µm diameter were optically trapped in solution where the (36) Olson, F.; Hunt, C. A.; Szoka, F. C.; Vail, W. J.; Papahadjopoulos, D. Biochim. Biophys. Acta 1979, 557, 9-23. (37) Szoka, F.; Olson, F.; Heath, T.; Vail, W.; Mayhew, E.; Papahadjopoulos, D. Biochim. Biophys. Acta 1980, 601, 559-571. (38) Macdonald, R. C.; MacDonald, R. I.; Menco, B. Ph. M.; Takeshita, K.; Subbarao, N. K.; Hu, L.-R. Biochim. Biophys. Acta 1991, 1061, 297-303. (39) Hope, M. J.; Bally, M. B.; Webb, G.; Cullis, P. R. Biochim. Biophys. Acta 1985, 812, 55-65.


4923

Figure 3. Raman spectrum of an optically trapped, 3-µm-diameter DPPC liposome. Peaks noted by arrows are (a) perchlorate Cl-O stretch, (b) C-C stretching region, (c) CH2 twist, (d) CH2 symmetric stretch, and (e) CH3 symmetric stretch.

location of the trap was maintained ∼10 µm above the coverslip to avoid interactions with the glass surface. This arrangement mimics trapping conditions for an optical tweezers experiment, where the trap is used primarily to manipulate a liposome or simply hold it in place for experiments in solution. Figure 3 shows a typical Raman spectrum acquired from a DPPC liposome in an optical trap in a solution containing 50 mM tetramethylammonium perchlorate with a laser power of 6 mW reaching the objective. The peak corresponding to the CH2 twist (∼1300 cm-1)31,40,41 in the acyl chains was used as the marker band for the phospholipid membrane. To monitor the relative amount of buffer solution outside the vesicle that was within the confocal probe volume, scattering from the symmetric stretch of perchlorate ion (933 cm-1) was used as a marker. It has been previously demonstrated that DPPC liposome membranes are impermeable to perchlorate ion transport at 24 °C over the time course of these experiments,31 so the perchlorate ion signal reports detection of only the solution outside the vesicle. To observe the power dependence of Raman scattering from the liposome sample, a unilamellar liposome (∼3 µm in diameter) was optically trapped in a solution containing 50 mM perchlorate ion and the laser power entering the objective was varied from 3 to 24 mW. The top panel of Figure 4 shows the power-dependent response from the phospholipid bilayer(s) of the trapped liposome. The number of counts from the CH2 twist peak was divided by the time for signal collection to determine an average number of counts per minute from the phospholipid of the trapped liposome. The lower x axes in Figure 4 indicate the laser power that reached the objective, which is an upper bound to the laser power at the sample. At laser powers of e6 mW, the slope of the trend line is lower by a factor 0.61 compared to the response from the same liposome excited at laser powers of g9 mW. The upper x axis on the top panel of Figure 4 corresponds to the predicted lowering of the free energy (eq 6) by transferring phospholipid for water in the center of the optical trap at the corresponding incident laser powers. Figure 4 displays the power-dependent perchlorate ion Raman signal arising from the ions in solution outside of the ∼3-µmdiameter vesicle. Dislocations in the linear trends of the data (40) Brown, K. G.; Peticolas, W. L.; Brown, E. Biochem. Biophys. Res. Commun. 1973, 54, 358-364. (41) Gaber, B. P.; Peticolas, W. L. Biochim. Biophys. Acta 1977, 465, 260-274.


Figure 4. Power-dependent Raman scattering response of a liposome (∼3-µm diameter) trapped in free solution, containing 50 mM perchlorate. The number of counts from the CH2 twist band (b, top plot) and perchlorate stretch (0, bottom plot) were divided by the total integration time for each observation. The break in the continuity of the line suggests a change in the shape of the liposome. The slope of the line fitted through the points at powers greater than 9 mW is ∼1.65 times greater than the line fitted at powers below 6 mW.

appear in both phospholipid and perchlorate ion signals, where between 6 and 9 mW, a nearly 2-fold jump in the signals of both the phospholipid and the perchlorate ion solution probe are observed. A control experiment, carried out using an optically trapped 3-µm-diameter polystyrene latex particle with the same range of laser powers, showed no discontinuity; the polystyrene Raman scattering is linear with laser power, r2 ) 0.994. Since the DPPC bilayer is impermeable to the perchlorate ion under these experimental conditions,31 the discontinuity in both the phospholipid and perchlorate signals indicates that there is a jump in the amount of phospholipid and the amount of external solution within the volume of the optical trap. We interpret these results as indicating that a second bilayer is being drawn into the center of the optical trap at higher trapping powers, as illustrated in Figure 5. This change in liposome shape would nearly double the signal from the phospholipid bilayer since twice the area would be sampled, although the sensitivity per bilayer should be slightly lower since both membranes are not likely to be drawn into the exact center of the trap (recall that the excitation and collection efficiencies falls off rapidly with z due to the confocal optics32). Note that this change in liposome shape also nearly doubles the volume of external solution that is located within the optical trap, which coincides with the jump in the perchlorate signal shown in the bottom panel of Figure 4. The signal discontinuities for both the membrane and external solution

Figure 5. Cartoons depicting one and two bilayers in an optical trap. The dots represent the perchlorate ions in the solution outside the vesicle. The horizontal lines represent the approximate boundaries of the confocal volume, 1 µm above and below the center of the trap.

occur when the trapping energy is between 5 × 10-13 and 7 × 10-13 erg. The bending rigidity of spherical phosphatidylcholine vesicles has been reported in the literature to correspond to energies on the order of 8 × 10-13 erg,42 which is very close to these experimental results. The change in liposome shape, therefore, arises from the second bilayer being drawn into the center of the optical trap when there is sufficient lowering of the free energy from the integral of the gradient force to overcome the energy cost of deforming the membrane from its spherical shape. The deformation of the liposome shape, such as that illustrated in Figure 5, would change either the surface area of the liposome or its interior volume. If the interior volume remained constant, the change in surface area of the liposome required to distort it from a spherical shape would require a significant increase in the free energy, due to the large surface tension of phospholipid membranes.44-46 The energy cost of increasing the membrane area by only 15% (less that the fractional area increase required to (42) Fernandez-Puente, L.; Bivas, I.; Mitov, M. D.; Me´le´ard, P. Europhys. Lett. 1994, 28, 181-186. (43) Thompson, T. E.; Huang, C. Ann. N. Y. Acad. Sci. 1966, 137, 740-744.

produce the distorted shape in Figure 5 from a sphere of the same volume) is predicted from measured phospholipid bilayer surface tensions44-46 to be ∼10-7erg. This value is more than 105 times greater than the free energy available from the optical trap, at laser powers that are capable of distorting the vesicle shape (see above). We conclude, therefore, that the surface area of the liposome membrane is not increased by the deformation; this conclusion requires that the interior volume of the vesicle decreases under the influence of the gradient trapping force on the bilayer. In order for the interior volume of the vesicle to change in response to the gradient optical force on the bilayer, the hydrostatic pressure must relax by efficient transport of water through the vesicle membrane. The time required for the volume to change in response to a pressure gradient can be estimated from the permeability coefficient of water through a phospholipid membrane, an average value of which has been reported to be kp ) 4 × 105 s-1.43 The flux of water through the membrane can be estimated from this value, J ) kp[H2O]o ) 2.4 × 10-4 mol cm-2 s-1, where [H2O]o ) 0.055 mol cm-3 is the concentration of water at the membrane/solution interface. The surface area of a 5-µm liposome membrane is A ∼ 8 × 10-7 cm-2, giving a rate of transfer of water of 2 × 10-10 mol/s across the bilayer. The interior volume of a 5-µm-diameter liposome is ∼1 fL, so the vesicle contains ∼6 × 10-14 mol of water. Given the rate of transfer and the water content of the vesicle, the time required for complete exchange of the water in the vesicle is only 0.3 ms, much shorter than the time needed to change the laser power and acquire a Raman spectrum in these studies. This fast water permeability should lead to equilibration of any pressure gradients with a rapid change in volume, thereby allowing the bilayer to bend and distort without the large energy cost of increasing its surface area. A liposome could be oriented in an optical trap with lipid bilayer perpendicular to the propagation of the laser radiation (as illustrated in Figure 5) or with the lipid bilayer parallel to the laser beam. In this latter case, when the laser power is low and shape distortion is negligible, trapping a single bilayer would hold the liposome asymmetrically in the beam since the untrapped bilayer would lie off to the side of laser beam center. Images acquired from liposomes of several sizes held in an optical trap show the vesicles centered axially in the laser beam,31 lending support to the idea that the plane of the membrane is perpendicular to the laser beam propagation (Figure 5); this orientation would align the acyl chains of the phosphatidylcholine molecules within 20° of the propagation (z) axis.52 The relative intensities of several features in the Raman spectra of long alkyl chains50,51 and phosphatidylcholine molecules52,53 are sensitive to their orientation (44) Evans, E.; Heinrich, V.; Ludwig, F.; Rawicz, W. Biophys. J. 2003, 85, 23422350. (45) Gulgliotti, M.; Politit, M. J.; Chaimovich, H. J. Colloid Interface Sci. 1998, 198, 1-5. (46) Israelachvili, J. Intermolecular and Surface Forces, 2nd ed.; Academic Press: San Diego, 1992; p 315. (47) Gottleib, M. H.; Eanes, E. D. Biophys. J. 1972, 12, 1533-1548. (48) Koppel, D. E.; Axelrod, D.; Schlessinger, J.; Elson, E. L.; Webb, W. W. Biophys. J. 1976, 16, 1315-29. (49) DiSalvo, E. A. Biochim. Biophys. Acta 1987, 905, 9-16. (50) Snyder, R. G. J. Mol. Spectrosc. 1970, 36, 222-231. (51) Harrand, M. J. Mol. Struct. 1989, 214, 71-91. (52) Harrand, M. J. Chem. Phys. 1983, 79, 5639-5651. (53) Harrand, M. J. Chem. Phys. 1984, 81, 1-5.


4925

with respect to the electric field vector of the incident radiation. Notable differences in the shape of the Raman spectra can be found in the C-C stretching region and the C-H stretching region depending on the orientation of the molecules. If the acyl chains were lying the xy plane, the trans/gauche ratio in the C-C vibrational region is e1 and the CH2/CH3 symmetric stretch ratio is .1.52 Aligned along or near the z axis (acyl chains parallel to the direction of light propagation), the trans/gauche ratio in the C-C region is >1 and the CH2/CH3 symmetric stretch ratio is ∼1.52 The Raman spectrum of a trapped liposome (as shown in Figure 3), which not change with increased laser power, shows peak ratios for both the C-C and C-H stretching regions that are consistent with molecules oriented along the axis of light propagation. These results are consistent with bright-field images that show the liposome centered in the optical trap31 and, therefore, with the cartoons in Figure 5. Distance Dependence of Optical Forces on Phospholipid Membranes. The capture of a second liposome bilayer by the optical trap at higher laser powers has allowed the bending energy barrier of a phospholipid membrane to be estimated. Optical trapping in free solution, however, does not allow the distance dependence of the trapping energy to be investigated. To gain control and interrogate the distance dependence of the optical forces on the membrane, ∼5-µm-diameter liposomes were adhered to a coverslip, thereby immobilizing them. The focal point of the objective was lowered beneath the coverslip surface and raised in small increments through an adhered liposome to control the distance between the center of the trap and the bilayer of the liposome. This distance dependence can, thus, provide information about the mobility of the lower and upper bilayers of the surfaceadhered liposome. To characterize the z-axis response of the confocal probe to an immobile membrane, a thin (∼10 nm) polystyrene film was deposited on a coverslip by dip coating at a controlled withdrawal rate from a dilute solution of polystyrene in chloroform. The z-axis response to the thin film serves as an immobile standard to be compared to the response of the lipid bilayer. The z position of the probe was moved in steps of ∼0.5 µm for each successive spectrum that was acquired. Since the thickness of a phospholipid bilayer (∼5 nm47) and the 10-nm polystyrene film is 2 orders of magnitude smaller than the expected confocal depth resolution of the microscope,32,48 the thin polymer film results should mimic the response of an immobilized phospholipid bilayer. The collection efficiency of the confocal Raman microscope has been previously measured for small particles;32 those results are similar to the depth-dependent response of the thin, polystyrene film, shown in Figure 6. The full width at half-maximum of the thin film response is ∼1.6 µm and agrees with the expected Lorentzian shape of the confocal collection efficiency function as shown in Figure 6.48 To test the scan of the confocal probe through a rigid liposome (control) attached to the glass coverslip, 5-µm-diameter DPPC liposomes were prepared by hydration in a calcium ion-containing buffer. Calcium ion is known to increase significantly the rigidity of phospholipid bilayers,49 thus presenting the optical trap with a less deformable vesicle structure for scanning. A plot of a representative depth scan of a ∼5-µm-diameter liposome, hydrated in 5 mM Ca2+ buffer and adhered to the coverslip, is displayed in 4926 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

Figure 6. Depth-dependent Raman scattering response from a polystyrene film (∼10 nm in thickness) dip-coated onto a coverslip. The solid line is a fit to a Lorentzian function describing the z dependence of the confocal collection efficiency.48

Figure 7. Depth-dependent Raman response from a 5-µm-diameter DPPC liposome (b) in buffer containing 5 mM calcium ion. The peak on the left side (2.5 µm) corresponds to the response from the bilayer adhered to the coverslip. The peak on the right side (8 µm) corresponds to the bilayer extended into solution. The polystyrene thin film response (+) is overlaid for comparison. The Z locations correspond to the confocal probe as noted in Figure 8.

Figure 7. The Raman signal from the CH2 twisting mode is normalized and plotted. Figure 7 also shows an overlay of the depth-dependent Raman signal from the polystyrene film on the depth scan for comparison. The rigidity in the membrane is clear from these results; the response from each of the two bilayers overlaps the response of the completely immobile polystyrene film. This nearly identical response leads us to the conclusion that, in the presence of calcium ion, neither phospholipid bilayer (the layer adhered to the coverslip nor the one that extends into solution) is significantly pulled or pushed by the optical forces available from the focused laser beam. Therefore, the optical trap simply profiles the unperturbed spatial distribution of phospholipid at the interface, as illustrated in Figure 8A. Based on the optical trapping results presented earlier, we would anticipate that forces acting on a deformable liposome (without calcium ion) should perturb the shape of the free membrane of an attached liposome. To test this expectation, ∼5µm-diameter DPPC liposomes were prepared in phosphate buffer, diluted into a solution containing 50 mM perchlorate ion, and allowed to adhere to a coverslip. Raman scattering from the perchlorate ion could be monitored as a marker for the amount of external solution in the confocal probe. Figure 9A shows the depth-dependent Raman scattering (CH2 twisting mode) from a liposome scanned with 6 mW of laser power; the plot also includes a comparison to the response from the polystyrene film. It is easily

Figure 8. Shapes of surface-attached liposomes in response to the forces of an optical trap. (A) Cross section of a liposome adhered to a coverslip that is rigid and does not respond to optical forces acting on the lipid bilayer. Z locations correspond to the probe at the coverslip surface, in the middle of the liposome and at the top portion of the bilayer that is extended into solution (see Figure 7). (B) Cross section of a liposome adhered to a coverslip with the center of the confocal probe inside the liposome. The gradient force from the laser light is pulling the upper bilayer into a bowl-like shape. (C) Cross section drawing of a liposome adhered to a coverslip with the center of the confocal probe above the equilibrium position of the bilayer. The bilayer is drawn further into solution by the gradient force of the optical trap.

seen that the bilayer attached to the coverslip is not affected by the optical forces. However, the upper bilayer extending into solution appears to be manipulated by the optical forces, as evident in the significant broadening of the z-axis Raman response from this mobile portion of the bilayer (as illustrated in the sketch in Figure 8B,C). The Raman scattering response of the CH2 twisting mode of the phospholipid membrane is complemented by changes in the intensity of perchlorate ion scattering from the outer solution that is detected within the volume of the optical trap, as shown in Figure 9B. The signal from the perchlorate marker in solution is approximately half-maximum when the top bilayer is in the middle of the optical trap, indicating that the volume of the optical trap is half-filled with solution outside of the vesicle. The effect of optical forces on the organization of the phospholipid molecules can be tested by comparing the spectrum of the bilayer that is adhering to the coverslip with the bilayer that is extended into solution, to determine whether the optical forces cause a change in the orientation of the molecules. Figure 10 displays spectra corresponding to the two maximums seen in

Figure 9. Monitoring the deformation of a surface-attached lipsosome. (A) Depth-dependent Raman scattering response from a 5-µmdiameter DPPC liposome in buffer without calcium ion ([); the polystyrene thin-film response (+) is overlaid for comparison. (B) Depth-dependent Raman scattering response from the 5-µm-diameter DPPC liposome ([) overlaid with the signal from the perchlorate ion in the solution outside the liposome (O).

Figure 10. Raman spectra of a DPPC bilayer adhered to a coverslip (lower line) and a bilayer in contact with solution (upper line). An offset of 3000 counts was added to the response from the bilayer in contactfree solution. The only difference between the spectra is the increase in the perchlorate peak at 933 cm-1, indicating that greater solution volume outside to the vesicle is within the confocal probe volume when observing the bilayer in contact with free solution.

Figure 9 that correspond to the bilayers adhered to the coverslip and extended into solution. The resulting spectra in Figure 10 are indistinguishable except for the much smaller perchlorate scattering peak at 933 cm-1 for the surface-adhered bilayer, which would be expected. The results indicate that no structural Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

4927

differences are detectable in the Raman spectrum for the supported versus the free bilayer of the vesicle. The z-axis responses of the signals from the solution perchlorate and from the phospholipid bilayer as well as the orientation data allow us to reach conclusions as to what changes occur in the shape of the phospholipid bilayer of a liposome when optical forces are introduced. The lipid bilayer is pulled into the trap from the top and/or bottom of the liposome and not the sides. Several factors lead to this conclusion. First, the distribution of light in the optical trap produces no significant radiation in the xy plane outside of the beam spot due to the rapid falloff of the Gaussian radial distribution. In the absence of light, no optical force can exist, and lateral forces on the membrane that would draw material in from the sides of the trap are negligible. Second, since the confocal optics allow for the distinguishing of contents inside and outside the probe volume, the nature of the deformation caused by the optical trap can be determined by analyzing the data acquired on adhered liposomes. The perchlorate signal is observed to increase as the center of the probe is moved upward from the center of the surface-adhered vesicle. At the point where the upper bilayer becomes trapped, the perchlorate signal rises from zero to approximately half of its maximum value (Figure 9B), indicating that about half of the final external volume is present in the volume of the optical trap at this point. This is indicative of the upper bilayer being drawn downward toward the center of the liposome (Figure 8B). Third, the liposome does not move laterally in the trap (see Figure 11). If only one bilayer were trapped and that bilayer was drawn in from the side of the vesicle, then the liposome would move laterally in the x,y plane to relieve some of the curvature strain that would be induced in the vesicle. The images in Figure 11 (with and without the laser beam on) clearly show that the liposome has not moved in any direction in the x,y plane of the image. Finally, the response from the liposomes prepared without calcium compared to those prepared with calcium, confirms the expectation that calcium significantly increases the energy barrier for deformation of the bilayer relative to the trapping energy available at the moderate laser powers used for this analysis. There is also a significant difference between the mobility of the surfaceattached bilayer and the bilayer in contact with solution. If the bilayer adhered to the coverslip may be manipulated by the light and detached from the surface as easily as the bilayer that is in contact with solution, then the width of the peaks from each bilayer would be the same. It is easily seen from Figure 9 that this is not the case. Summary. The effects of laser forces on optically trapped, unilamellar phospholipid vesicles have been determined by confocal Raman microscopy on liposomes trapped in free solution as well as those adhered to a coverslip. Vesicles trapped in free solution show that deformation of the spherical shape of the lipid bilayer by laser light is possible at milliwatt laser powers where the trapping energy is greater than the energy needed to bend the membrane from its spherical shape. For vesicles adhering to a glass surface, scanning the confocal probe volume below the free membrane can draw it toward the center of the optical trap,


Figure 11. Bright-field images of a 5-µm-diameter DPPC liposome adhered to a coverslip. (A) Laser off. (B) Laser on. The bright spot in the center of the liposome is the laser light scattered by the lipid bilayer. There is no detectable lateral motion of the liposome when the optical forces are introduced to the sample via the laser.

again distorting the vesicle from its spherical shape. It was verified that the rigidity of a membrane bilayer is significantly increased by the addition of calcium. Finally, the bending rigidity of a DPPC vesicle in the absence of calcium was determined experimentally to be ∼6 × 10-13 erg, a value that closely matches the reported membrane rigidity found in the literature. ACKNOWLEDGMENT The authors thank John Conboy for many helpful discussions. This research was supported in part by the National Science foundation under Grant CHE-0137569.

Received for review May 20, 2004. Accepted July 8, 2004. AC0492620

Optical Trapping of Unilamellar Phospholipid Vesicles: Investigation of

Recommend Documents