Transport, Separation, and Accumulation of Proteins on Supported

Jul 22, 2010 - Cummington Street, Boston, Massachusetts 02215. ABSTRACT .... result in both a phase change due to mass adsorption and a shift in the ...
0 downloads 0 Views 3MB Size
pubs.acs.org/NanoLett

Transport, Separation, and Accumulation of Proteins on Supported Lipid Bilayers J. Neumann,†,|,⊥ M. Hennig,‡,| A. Wixforth,*,† S. Manus,‡ J. O. Ra¨dler,‡ and M. F. Schneider§ †

Nano Lett. 2010.10:2903-2908. Downloaded from pubs.acs.org by EASTERN KENTUCKY UNIV on 01/23/19. For personal use only.

Center for NanoScience CeNS, Universita¨t Augsburg, Institut fu¨r Physik Universita¨tsstrasse 1, D-86159 Augsburg, Germany, ‡ Center forNanoScience, Ludwig-Maximilians-Universita¨t, Fakulta¨t fu¨r Physik Geschwister Scholl Platz 1, D-80539 Mu¨nchen, Germany, and § Boston University, Mechanical Engineering, 110 Cummington Street, Boston, Massachusetts 02215 ABSTRACT Transport, separation, and accumulation of proteins in their natural environment are central goals in protein biotechnology. Miniaturized assays of supported lipid bilayers (SLBs) have been proposed as promising candidates to realize such technology on a chip, but a modular system for the controlled transport of membrane proteins does not exist. In this letter, we demonstrate that standing surface acoustic waves drive the in-plane redistribution of proteins on planar SLBs over macroscopic distances (3.5 mm). Accumulation of proteins in periodic patterns of about 10-fold protein concentration difference is accomplished and shown to relax into the homogeneous state by diffusion. Different proteins separate in individual fractions from a homogeneous distribution and are transported and accumulated into clusters using beats. The modular planar setup has the potential of integrating other lab-on-a-chip tools, for monitoring the membrane-protein integrity or adding microfluidic features for blood screening or DNA analysis. KEYWORDS Surface acoustic waves, Proteins, supported lipid bilayer, control, separation, accumulation

S

upported lipid bilayers (SLBs) provide a versatile platform for applying surface sensitive techniques in studying the properties of soft matter.1 A variety of tools to biofunctionalize the interface by incorporating or associating macromolecules (e.g., proteins or DNA) has been developed and was used to unravel the role of lock and key forces during cell adhesion2 or the principles of laterally confined motion of single molecules.3 However, controlled manipulation of membrane-bound proteins, an essential tool for biotechnological applications, has proven challenging and even simple spatial arrangements of membranes or proteins are only achieved with elaborate tools and expertise. Micropatterned or -structured substrates4-8 or electrical fields9,10 have been successfully applied to arrange SLBs and observe lipid rediffusion. The key feature of an analytical tool, however, active control of “trafficking”, separation and accumulation of the proteins or lipids within the membrane simultaneously, is beyond the scope of any of these methods. One promising approach for manipulating membranes with solid support is the use of an active substrate. Surface acoustic waves (SAWs), for example, can be excited under very controlled conditions and have been shown to allow for microagitation,11 droplet actuation,12 mixing,13,14 onchip drug screening,15 particle separation16 and finally, to create in-plane density modulations in lipid membranes.17

In this letter, we demonstrate how standing SAWs can be used to accumulate, transport, and separate proteins or protein complexes reversibly in the “native” environment of a SLB (thickness ∼4 nm). Employing fluorescence microscopy, we show that proteins anchored by biotin-functionalized lipids can reversibly or irreversibly be transported and aligned depending on the protein surface coverage. Homogeneous mixtures of proteins of different size are separated into individual fractions and accumulated reversibly into spatially and temporally controlled patterns. Finally, we demonstrate the advantage of our modular lab-on-a-chip approach by integrating a surface sensitive mode as a tool to monitor the formation and stability of the membraneprotein layer. SAWs are generated on piezoelectric lithium tantalate (LiTaO3) substrates using metallic interdigital transducers (IDTs)18 as shown in Figure 1a. Briefly, an applied voltage on these IDTs leads to a deformation of the substrate due to the piezoelectric effect. The characteristics of these deformations depend strongly on the crystal structure and crystal cuts used. A periodic alternating voltage results in a periodic deformation and propagation of waves, standing waves in the case of the superposition of two waves propagating in opposite directions. The substrates used have a thickness of 200-400 µm and are transparent and mirror polished on both sides, allowing for high optical resolution with fluorescence microscopy. The crystal cut (36°Y) together with conventional metallic IDTs predominantly generates shear waves.19 The active area and the IDT structures were finally coated with a layer of SiO2. This layer protects and also separates the metallic structures from the liquid and allows better supported lipid bilayer formation compared to bare

* To whom correspondence should be addressed. |

These authors contributed equally.



Current address: Biozentrum der LMU Mu¨nchen, Grosshaderner Str. 2, D-82152 Planegg-Martinsried, Germany. Received for review: 3/21/2010 Published on Web: 07/22/2010 © 2010 American Chemical Society

2903

DOI: 10.1021/nl100993r | Nano Lett. 2010, 10, 2903–2908

result in both a phase change due to mass adsorption and a shift in the sensor impedance, producing a change in transmission coefficient of the chip (Figure 1b.). A plateau in the sensor signal indicates a stable situation, for example, the formation of a SLB. The binding curve shown demonstrates the possibility of using the presented setup as a sensor system without applying any modifications. In principal, the binding and relaxation curves could be analyzed to study and interpret the binding process in more detail. This is done by several researchers, for example Gizeli et.al,23 but it always requires the calibration of the sensor, for example, by comparison with other sensing techniques, such as plasmon resonance which was beyond the focus of this work. SAWs couple to SLBs, creating local membrane density modulations, which in turn result in the transport and accumulation of labeled lipids17 (Figure 3a). This could also be combined with other methods for local SLB ordering and rearrangement, for example, electrical fields9 or microfluidic flow.24 Incorporation of a functional linker (here biotin) into the SLB allows proteins to be anchored to the membrane. Figure 2a shows the organization of Texas Red labeled neutravidin bound to the membrane by a low density of biotinylated lipids (0.1% biotin-DHPE(N-(biotinoyl)-1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine, triethylammonium salt), 99.9% DOPC(1,2-dioleoyl-sn-glycero-3phosphocholine)). With the used biotin concentration and a density of approximately one lipid per 50 Å2,25 one biotin anchor is therefore expected about every 20 nm. (The footprint of a single streptavidin is roughly 50 lipids3 and negligible compared to the distance.) The accumulation of the lipid/protein complexes occurs in the antinode region of the standing SAW. The intensity profile of Figure 2a is shown in Figure 2b (data points). Clearly, an asymmetry between narrow bright (protein enriched) and broader dark (protein depleted) regions can be observed and fitted according to the model presented in ref 17. All of these patterns appear within seconds, remain stable as long as the SAW stimulation persists (minutes to even hours; steady state) and disappear within seconds (driven by diffusion) as soon as the SAW is switched off. Identical behavior was observed for the cases of streptavidin (Texas Red labeled, Figure 2c) and the very similar protein avidin (Atto 488 labeled, Figure 2d). Proteins maintain their mobility along the membrane26 and follow the pattern of the standing SAW reversibly when the surface coverage is kept low (0.1% biotin DHPE). Higher surface coverage (e.g., 1% biotin-DHPE) resulted in increasingly immobile pattern formation and SAW-driven protein clustering. Recalling that biotin-streptavidin is applied as a modular adapter for many macromolecules, this approach has the potential to allow the alignment of larger biological objects such as viruses or cells. In addition, other anchor systems can be incorporated quite simply as well. We here show that clotting factor VIII, which is known to bind to a POPC/POPS (90/10 wt %) (1-palmitoyl-2-oleoyl-sn-glycero-

FIGURE 1. (a) Schematic representation of the setup used (not to scale). A LiTaO3 chip with IDTs (153 MHz/wavelength 26.6 µm) on opposite sides is used to generate standing SAWs. The supported membrane (∼4 nm thick) and partially aligned proteins (r ∼ 25 nm) between the IDTs can be observed by fluorescence microscopy through the optically transparent chip. (b) Measured phase of the SAW transmission signal. The setup is used as a biosensor without further modification (proof of principle). A typical transmission phase signal of such a SAW chip while loading it with reagents is shown. Clearly visible are the injection of SUVs and factor VIII solutions and the successively increasing signal with each layer. The spreading of the lipid bilayer and factor VIII binding can be observed in real time and label free. The binding and relaxation curves could be transformed into mass coverage requiring calibration by reference methods (e.g., plasmon resonance).

LiTaO3. Standing waves were excited by continuous and simultaneous stimulation of the substrate using two opposed IDTs (Figure 1a). The standing shear SAW extends over the whole distance between both transducers (3.5 mm) even in the presence of water. Displacement amplitudes are expected to be in the order of 0.1-1 nm.20-22 More common Rayleigh-wave substrates (often made of quartz or lithium niobate (LiNbO3)) are not suited for the application presented here due to their strong coupling into the liquid (acoustic streaming)14 and the resulting strong attenuation.19 Supported lipid bilayers were prepared using the vesicle fusion method1 and controlled by optical homogeneity and measurement of their diffusion constant.17 Subsequent membrane spreading as well as the binding process of proteins were monitored online using the sensing abilities of the SAW chip (Figure 1a). Changes in surface and liquid bulk conditions on the delay line and on the IDT-structures © 2010 American Chemical Society

2904

DOI: 10.1021/nl100993r | Nano Lett. 2010, 10, 2903-–2908

FIGURE 2. Fluorescence micrographs of various proteins bound to SLBs segregated by SAWs. (a) Mobile low-density neutravidin accumulations on a DOPC/0.1% biotin-DHPE membrane. (b) Profile of the micrograph shown in (a) (data points) and I(x) ∼ -exp(sin2 x) (eq 3) fit (red line). The background intensity is 140((10) measured at the intransparent gold structures. (c) Low coverage streptavidin alignment. (d) Low coverage avidin alignment. (e) Immobile high surface coverage factor VIII alignment on a POPC/10% POPS membrane by SAW. (f) Immobile high coverage annexine accumulation on POPC/POPS membrane. The stripe distance and the scale bar always correspond to half the SAW wavelength of 13.3 µm.

3-phosphocholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero3-phospho-L-serine (POPS)) SLB via hydrophobic anchors and electrostatic interactions, can be patterned in stripes when a standing wave is excited (Figure 2e). Fluorescein annexin organization, which binds to POPS via calcium ion bridges (Figure 2f) was observed as well. Both factor VIII and annexin accumulations were immobile due to their high surface coverage. All membrane-binding proteins tested could be accumulated into patterns employing the exact same setup. Importantly, these patterns can be switched on and off reversibly in the proteins’ “native” environment. We were also suc© 2010 American Chemical Society

cessful in creating stable conglomerations of protein patterns using high surface coverage of biotin-anchors (∼10%). Stripes created under these conditions, similar in their appearance to those shown in Figure 2e/f, do not disappear within approximately one hour after the SAW had been switched off. To produce a useful tool for protein manipulation and to create enriched accumulations of substantial amounts of protein on the chip surface, simple but controlled transport of the proteins is required (“membrane fluidics”). We recall that the stationary patterns of nodes and antinodes are determined by the boundary conditions of the system. 2905

DOI: 10.1021/nl100993r | Nano Lett. 2010, 10, 2903-–2908

FIGURE 4. SAW-induced segregation of membrane-bound streptavidin (red, Texas Red label) and avidin (green, Atto 488 label). Sketch (a) visualizes the proposed protein separation mechanism. In the antinode of the standing SAW, the membrane density is higher than in the node regions, so the biotinylated lipids (orange) accumulate in this region. As avidin is a larger protein than streptavidin and less capable of crystallization, a streptavidin-enriched region allows for a higher biotin density than avidin-rich regions and the system relaxes into an overall lower energetic state and avidin is pushed to the low density node region. (b) Originally homogeneously distributed avidin (green) and streptavidin (red) separate, when the SAW is applied. The background corrected profiles of both micrographs are plotted in (c), where avidin is depicted green and streptavidin (red, dashed line).

bound to a SLB. Although the increased size of the lipid/ protein complex increases the viscous drag and reduces the transport velocity by 50% compared to the pure lipid system,19 the transport remains very efficient with a persistent concentration profile and clear boundaries. One of the most common procedures in biotechnology and molecular biology is the analysis of protein distributions using gel-electrophoresis, which relies on size sensitive transport. In Figure 4, two different protein populations, originally homogeneously distributed on the SLB, are separated into individual fractions and transported into two distinct regions under the presence of SAW. The effect is dependent on the differences in protein-protein interaction (e.g., molecular weight, electrostatics, and so forth); however, it allows even for separation of proteins such as streptavidin and avidin despite their similarity in size and nature but differences in their ability to crystallize.3,27 Combining the cornerstones transport, separation and accumulation opens the avenue for protein purification in artificial and native membranes on a chip, a “Flatland Factory”. The standing shear waves cause a lateral strain field in the piezo crystal that is coupled to the membrane through the 0.5-2 nm thin water layer28 between bilayer and substrate. We assume that as a result, the membrane is subjected to a lateral density modulation. In the node regions (where strain is maximal), this results in a low area density

FIGURE 3. Protein and lipid transport. (a) Patterns of TR-labeled lipids are transported by shifting of the SAW pattern using two generators, each connected to one of the IDTs (far left and right, not shown) and with slightly detuned frequencies (∆f ) 0.1 Hz). This results in an according beat frequency of the standing acoustic wave. The pattern therefore shifts one wavelength (26.6 µm) in 10 s (ν ) 2.66 µm/s). Micrographs are depicted on the left, while the intensity profiles are shown on the right, respectively. (b) Accumulations of low-density streptavidin are moved by manual phase-shift of one generator (∆f ) 0.001 Hz). The velocity chosen was reduced because of the lower mobility of the proteins attached to the membrane (DProtein , DMembrane), as well as the needed exposure time of 5 s, caused by the low signal-to-noise ratio under the given conditions.

Hence, continuously shifting the phase between the two stimulating waves results in an equally continuous relocation of the nodes and antinodes (beat pattern). This, in turn, can be translated into homogeneous macroscopic lateral transport. In Figure 3a, a typical membrane pattern consisting of accumulations of labeled lipids (TR-DHPE, Texas Red DHPE 1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine, triethylammonium salt) is set in motion by slightly detuning the excitation frequencies of both IDTs (∆f ) 0.1 Hz, f ) 153 MHz) (see movie in Supporting Information). Figure 3b demonstrates that this is not limited to lipids but can also be used to transport proteins (e.g., streptavidin) © 2010 American Chemical Society

2906

DOI: 10.1021/nl100993r | Nano Lett. 2010, 10, 2903-–2908

while at the antinodes of the SAW (where strain is minimal, Figure 4a) it leads to high-density regions. The density gradient is sustained for the entire duration of the stimulation. It is visualized by the local accumulation of fluorescently labeled lipids, which prefer high (“high density seekers”) over low (“low density seekers”, Figure 3a) area densities (or vice versa). Embedded and originally homogenously distributed fluorescently labeled lipid molecules (e.g., TRDHPE)17 tend to segregate into lamellar patterns due to the tendency of TR-DHPE to accumulate within regions of low membrane density, or liquid disordered phases.29,27 From a thermodynamic perspective, the resulting concentration profile cF of the low density seeking molecules is expected to follow a Boltzmann distribution

cF(x) ) c0 exp(-∆G(x)/kBT)

decanoyl-sn-glycero-3-phosphoethanolamine), fluorescein DHPE (N-(fluorescein-5-thiocarbamoyl)-1,2-dihexadecanoylsn-glycero-3-phosphoethanolamine, triethylammonium salt) and Bodipy Fatty Acid (5-butyl-4,4-difluoro-4-bora-3a,4adiaza-s-indacene-3-nonanoic acid) according to ref 29. Figure 2b shows that the high density seeker follows the same distribution (eq 3), yet the sign of I0 is reversed. We conclude that both lipid and protein segregation are based on the same mechanism, a membrane density modulation imposed by the 2D strain of the acoustic field in the substrate. When two different proteins (e.g., different in molecular weight, isoelectric point, ability to crystallize) compete for the antinode position, protein separation is observed. As shown in Figure 4 for biotin DHPE-bound streptavidin and avidin, streptavidin accumulates in the antinode region. We propose the following explanation: streptavidin forms dense clusters in the antinode region due to its tendency to crystallize in a matrix of avidin.3 The biotinylated-lipid accumulates in the antinode region (like most other modified lipids examined except for TR-DHPE) dragging the proteins there. The best compromise is a high concentration of streptavidin at sites of high binding site density (Figure 4a) and an increased avidin concentration in the opposed regions. In other words, the penalty for “pushing” the avidin toward regions of lower binding site density is less than that for reversing the accumulation of the biotinylated-lipid. The mechanism can therefore be referred to as “lipid-driven”. In summary, we have shown that standing surface acoustic shear waves can be employed as a nanoscale machinery to transport, separate and accumulate proteins on flat twodimensional membranes. We demonstrate that formation of proteins into stable patterns is fully reversible and can be performed with precise spatial and temporal control while retaining the proteins’ native environment. The mechanism originates from a local, SAW-induced density modulation in the lipid membrane that is transferred to the adsorbed macromolecules. Manipulation of the excitation frequency or relative phase sets the density pattern into motion and drives protein locomotion. Proteins of varying size are selectively transported into different regions on the chip, which allows for protein separation or purification from a homogeneous mixture. Finally, we wish to point out that our approach is kept highly modular with flexible protein-membrane anchor systems as shown using proteins bound to the SLB by different binding mechanisms (Figure 2) and is not limited to membrane-bound macromolecules like receptors, peripheral proteins, or ion channels but can in principle be extended to viruses, bacteria, or entire cells bound to a support. Even spreading entire native biomembranes (like erythrocyte ghosts) followed by separation, accumulation or purification of macromolecules (“2D electrophoresis”)30,31 might be realized as DNA, for instance, can be coupled to cationic membranes and its movement directly coupled to the lipids.32 Increasing the excitation frequency to about 3

(1)

where ∆G(x) is the free energy difference a molecule at position x (between node and antinode) experiences with respect to the undisturbed case with average dye concentration c0. We assume that the free energy landscape ∆G(x) is created by a density modulation in the supported membrane, which in turn is caused by the underlying periodic piezoelectrical strainfield u(x) ∝ sin kx of the piezoelectric crystal. Assuming a first-order approximation of linear coupling between the local strain field and the force acting on the molecule in the membrane, we yield ∆G(x) ∝ |u(x)|2 and hence

∆G(x) ∝ sin2 kx

(2)

with the wave vector k ) 2π/λ given by the wavelength λ of the SAW device. Since the dye concentration profile cF is proportional to the observed lateral intensity profile IF, we expect a profile of the form

(

IF(x) ) IB + I0 exp -

α sin2 kx kBT

)

(3)

in agreement with the experimental data (Figure 2b). The fit parameters are the coupling constant R the backgound intensity, IB and the modulation intensities I0. Plotting c(x)/ c0 ) (I(x) - IB)/I0, we yield R ≈ 2.5 (Figure 2b) and hence ∆Gmax ≈ 2.5kBT, which corresponds to clearly visible stripe patterns with approximately 13-fold accumulation. In contrast to the distribution of low density seeker TRDHPE, we found that proteins bound to the membrane by functional groups accumulate in antinodes (high density seeker). Indeed, the other labeled lipids examined were observed to accumulate in the antinodes despite TR-DHPE, such as Oregon green DHPE (Oregon Green 488 1,2-dihexa© 2010 American Chemical Society

2907

DOI: 10.1021/nl100993r | Nano Lett. 2010, 10, 2903-–2908

GHz, the lateral resolution may be downscaled to less than 100 nm allowing protein organization beyond optical resolution. In particular when combined with other planar lab-ona-chip technologies, we envision an entire Flatland Factory with single molecule conveyer belts, assembly lines, and acoustically driven microfluidics for applications such as blood screening or protein and DNA analysis.

(9)

Acknowledgment. We like to thank Professor J. Kotthaus for providing rf-equipment, Professor E. Sackmann and H. Engelke for fruitful discussions, and V. Myles for improving the quality of the manuscript. Financial support by the Excellence Cluster “Nanosystems Initiative Munich (NIM)” and the Center for NanoScience (CeNS) is gratefully acknowledged. We thank the Bayer HealthCare LLC for providing factor VIII and antibody label. M.F.S. thanks the Center of Nanoscience and Nanobiotechnology (CNN) of Boston University. Financial support of the German Science Foundation under program SPP1313 is gratefully acknowledged.

(14)

(10) (11) (12) (13)

(15)

(16) (17) (18) (19) (20) (21)

Supporting Information Available. Materials and methods, principle of SAW-sensing with transmission line (delay line), micrographs showing the positions of “high- and lowdensity seekers”, additional figures, and movie. This material is available free of charge via the Internet at http://pubs. acs.org.

(22) (23) (24)

REFERENCES AND NOTES

(25)

(1) (2)

(26)

(3) (4) (5) (6) (7) (8)

Sackmann, E. Science 1996, 271 (5245), 43–48. Smith, A.; Sengupta, K.; Goennenwein, S.; Seifert, U.; Sackmann, E. Proc. Natl. Acad. Sci. U.S.A. 2008, 105 (19), 6906–6911. Horton, M. R.; Reich, C.; Gast, A. P.; Ra¨dler, J. O.; Nickel, B. Langmuir 2007, 23 (11), 6263–6269. Sanii, B.; Smith, A. M.; Butti, R.; Brozell, A. M.; Parikh, A. N. Nano Lett. 2008, 8 (3), 866–871. Hochrein, M. B.; Leierseder, J. A.; Golubovic, L.; Ra¨dler, J. O. Phys. Rev. Lett. 2006, 96 (3), No. 038103. Krishnan, M.; Mo¨nch, I.; Schwille, P. Nano Lett. 2007, 7 (5), 1270– 1275. Nabika, H.; Iijima, N.; Takimoto, B.; Ueno, K.; Misawa, H.; Murakoshi, K. Anal. Chem. 2009, 81, 2. Sto¨gbauer, T.; Hennig, M.; Ra¨dler, J. O. Biophys. Rev. Lett. 2010, in press.

© 2010 American Chemical Society

(27) (28) (29) (30) (31) (32)

2908

Groves, J. T.; Boxer, S. G.; McConnell, H. M. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 935–938, 3. Stelzle, M.; Miehlich, R.; Sackmann, E. Biophys. J. 1992, 63, 1346– 1354. To¨gl, A.; Kirchner, R.; Gauer, C.; Wixforth, A. J. Biomol. Tech. 2003, 14 (3), 197–204. Wixforth, A.; Strobl, C.; Gauer, C.; To¨gl, A.; Scriba, J.; Guttenberg, Z. Anal. Bioanal. Chem. 2004, 379, 982–991. Sritharan, K.; Strobl, C. J.; Schneider, M. F.; Wixforth, A.; von Guttenberg, Z. Appl. Phys. Lett. 2006, 88, No. 054102. Frommelt, T.; Kostur, M.; Wenzel-Schafer, M.; Talkner, P.; Ha¨nggi, P.; Wixforth, A. Phys. Rev. Lett. 2008, 100, 3. Fillafer, C.; Ratzinger, G.; Neumann, J.; Guttenberg, Z.; Dissauer, S.; Lichtscheidl, I. K.; Wirth, M.; Gabor, F.; Schneider, M. F. Lab Chip 2009, 9, 2782–2788. Shi, J.; Huang, H.; Stratton, Z.; Huang, Y.; Huang, T. J. Lab Chip 2009, 9, 3354–3359. Hennig, M.; Neumann, J.; Wixforth, A.; Ra¨dler, J. O.; Schneider, M. F. Lab Chip 2009, 9, 3050–3053. White, R. M.; Voltmer, F. W. Appl. Phys. Lett. 1965, 7 (12), 314– 316. Shiokawa, S.; Moriizumi, T. Jpn. J. Appl. Phys. 1988, 27 (1), 142– 144 (Supplemental; Proc. 8th Symp. Ultrason. Electron). Hesjedal, T. Rep. Prog. Phys. 2010, 73, No. 016102. Drafts, B. IEEE Trans. Microwave Theory Tech. 2001, 49 (4), 795– 802. Rohrbeck, W.; Chilla, E.; Fro¨hlich, H.-J.; Riedel, J. Appl. Phys. A 1991, 52, 344–347. Gizeli, E.; Liley, M.; Lowe, C. R.; Vogel, H. Anal. Chem. 1997, 69, 4808–4813. Jo¨nsson, P.; Beech, J. P.; Tegenfeldt, J. O.; Ho¨o¨k, F. J. Am. Chem. Soc. 2009, 131, 5294–5297. Nagle, J. F.; Tristram-Nagle, S. Biochim. Biophys. Acta 2000, 1469, 159–195. Gambin, Y.; Lopez-Esparza, R.; Reffay, M.; Sierecki, R.; Gov, N. S.; Genest, M.; Hodges, R. S.; Urbach, W. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 2098–2102, 7. Manley, S.; Horton, M. R.; Lecszynski, S.; Gast, A. P. Biophys. J. 2008, 95, 2301–2307. Merkel, R.; Sackmann, E.; Evans, E. J. Phys. 1989, 50, 1535–1555. Nissen, J.; Jacobs, K.; Ra¨dler, J. O. Phys. Rev. Lett. 2001, 86 (9), 1904–1907. Tanaka, M.; Kaufmann, S.; Nissen, J.; Hochrein, M. Phys. Chem. Chem. Phys. 2001, 3, 4091–4095. Daniel, S.; Diaz, A. J.; Martinez, K. M.; Bench, B. J.; Albertorio, F.; Cremer, P. S. J. Am. Chem. Soc. 2007, 129, 8072–8073. Kahl, V.; Hennig, M.; Maier, B.; Ra¨dler, J. O. Electrophoresis 2009, 8 (30), 1276–1281.

DOI: 10.1021/nl100993r | Nano Lett. 2010, 10, 2903-–2908