Langmuir 2007, 23, 8129-8134
8129
Chemically Encapsulated Structural Elements for Probing the Mechanical Responses of Biologically Inspired Systems Ying Zhang,† Chao-min Cheng,† Brian Cusick,‡ and Philip R. LeDuc*,† Department of Mechanical and Biomedical Engineering and Biological Science, Department of Chemistry, Carnegie Mellon UniVersity, 5000 Forbes AVenue, Pittsburgh, PennsylVania 15213 ReceiVed February 19, 2007. In Final Form: April 16, 2007
A living cell has a crowded environment with a dense distribution of molecules that requires structured organization for its efficient functioning. One component of this structure, the actin cytoskeleton, is essential for providing mechanical support and facilitating many response activities, including the contraction of muscle cells and chemotaxis. Whereas many investigations have provided insight into the mechanical response from either an in vivo or in vitro perspective, a significant gap exists in determining how the living cell response and the polymer physics response are bridged. The understanding of these systems involves studying their components, including the individual cytoskeletal elements versus the higher-order organism organization in a living cell. Here, we leverage this organization in nature by using a chemistry-based approach to mimic the cytoskeleton in an artificial environment composed of spherically distributed lipid bilayers. This construct bears similarities to the cell membrane. To create a structurally regulated environment, we encapsulate G-actin into giant unilamellar vesicles and then polymerize actin filaments within individual liposomes. We visualize these vesicles with epifluorescence microscopy and confocal microscopy. Atomic force microscopy is then used to probe the mechanical properties of these artificial cells. This polymer cytoskeletal network appears to connect with the lipid bilayer and span the internal space within the liposomes in a manner similar to what is observed in living cells. This work will have implications in a variety of fields, including chemistry, polymer physics, structural biology, and engineering mechanics.
Introduction The mechanical response of a living cell is highly tuned to the structure provided by the cytoskeleton. This architectural arrangement also encompasses organizational regulation that is essential for complex cellular responses. Whereas the cytoskeleton is composed of actin filaments, microtubules, and intermediate filaments, the actin filaments play a major role in providing mechanical support for the cell and facilitate many physiological activities, including chemotaxis, phagocytosis, and mitosis.1-3 Attempts to understand the effects of the mechanical influence on the structural response have been made in a variety of fields ranging from biology to physics, but all have focused on producing an explanation that illuminates the core processes behind this behavior. However, significant gaps in understanding remain as a byproduct of the separate approaches pursed by each field. The “top-down” approach, favored by biological research, chooses to examine the complexity of cell structure behavior in vivo,4-6 whereas the “bottom-up” approach, favored by polymer physics research, elects to investigate cytoskeletal networks in vitro.7-9 * To whom correspondence should be addressed. E-mail: prleduc@ cmu.edu. Tel: 412-268-2504. Fax: 412-268-3348. † Department of Mechanical and Biomedical Engineering and Biological Science. ‡ Department of Chemistry. (1) Alberts, B.; Johnson, A.; Lewis, J.; Raff, M.; Roberts, K.; Walter, P. Molecular Biology of the Cell, 4th ed.; Garland Science: New York, 2002. (2) McGrath, J. L. Curr. Biol. 2006, 16, R326-327. (3) Sato, M.; Ohashi, T. Biorheology 2005, 42, 421-441. (4) Heidemann, S. R.; Wirtz, D. Trends Cell Biol. 2004, 104, 160-166. (5) Radmacher, M. Methods Cell. Biol. 2002, 68, 67-90. (6) Fabry, B.; Maksym, G. N.; Butler, J. P.; Glogauer, M.; Navajas, D.; Fredberg, J. J. Phys. ReV. Lett. 2001, 87, 148102. (7) Gardel, M. L.; Nakamura, F.; Hartwig, J. H.; Crocker, J. C.; Stossel, T. P.; Weitz, D. A. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 1762-1767. (8) Ziemann, F.; Radler, J.; Sackmann, E. Biophys. J. 1994, 66, 2210-2216.
Although such analyses have provided insight into the understanding of cell structure, developing a chemically based hybrid intermediate that merges the two bounded systems would enable a novel avenue for understanding the complexity of this process. A variety of approaches and methods have been employed to examine the filamentous actin networks through both in vivoand in vitro-based pursuits. The mechanical properties of actin, for example, have been measured in vitro.7,10,11 However, these actin filaments exhibit Young’s moduli that are several orders of magnitude below that of bulk actin and may be variable depending on their geometry and density.12 Furthermore, the Young’s moduli of F-actin networks are orders of magnitude larger than those measured in vivo.13 Whereas the values themselves are variable depending on a variety of factors including the experimental technique, advances in technology have improved the quality and accuracy of these property measurements. One of these techniques, atomic force microscopy (AFM), provides a unique approach to measuring the mechanical property of cells.14-16 The AFM has several modes of operation. The most appropriate of these modes for living cells is tapping mode liquid-AFM because of its improved force resolution and characteristic aqueous environment, which is more suitable to (9) Guzman, C.; Jeney, S.; Kreplak, L.; Kasas, S.; Kulik, A. J.; Aebi, U.; Forro, L. J. Mol. Biol. 2006, 360, 623-630. (10) Ruddies, R.; Goldmann, W. H.; Isenberg, G.; Sackmann, E. Biochem. Soc. Trans. 1993, 21, 37S. (11) Wachsstock, D. H.; Schwarz, W. H.; Pollard, T. D. Biophys. J. 1994, 66, 801-809. (12) Gardel, M. L.; Shin, J. H.; MacKintosh, F. C.; Mahadevan, L.; Matsudaira, P.; Weitz, D. A. Science 2004, 304, 1301-1305. (13) Gardel, M. L.; Nakamura, F.; Hartwig, J.; Crocker, J. C.; Stossel, T. P.; Weitz, D. A. Phys. ReV. Lett. 2006, 96, 088102. (14) Rosenbluth, M. J.; Lam, W. A.; Fletcher, D. A. Biophys. J. 2006, 90, 2994-3003. (15) Fernandez, J. M. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 9-10. (16) Costa, K. D. Dis. Markers 2003, 19, 139-154.
10.1021/la700488p CCC: $37.00 © 2007 American Chemical Society Published on Web 06/22/2007
8130 Langmuir, Vol. 23, No. 15, 2007
Zhang et al.
Figure 1. Artificial cell fabrication and analyses. (A) Schematic of the electroformation device for inducing the formation of the vesicles. A chamber was clamped by two indium tin oxide glass electrodes. An electric field was connected and applied for 4 h at 20 °C. (B) Schematic for investigating surface-attached F-actin liposomes with a fluorescence microscope or an AFM. (C) Schematic of the process for creating the artificial cells where the lipid film was dried in a chamber. G-actin in G-buffer was added to create preartificial cells, after which potassium ions triggered G-actin to polymerize into actin filaments. Two different forms of actin distribution result from this process.
an examination of living cells. One issue with this approach, however, is that solution viscosity presents a measurement challenge. Recently, these issues have been addressed, and the Young’s modulus of the lipid bilayer has been measured.17 In this work, we probe the complex mechanical and structural response by working on a hybrid system that mimics the cell (17) Legleiter, J.; Park, M.; Cusick, B.; Kowalewski, T. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 4813-4818.
structure by introducing the actin cytoskeleton into an artificial environment composed of lipid bilayers. We then examine structural properties through fluorescence and atomic force microscopy, which are well suited for our hybrid cell-like systems.
Results and Discussion To create our hybrid system, we first chemically synthesize membranes that are similar to cells in structure and size. Giant
Mechanical Response of Biological Systems
Langmuir, Vol. 23, No. 15, 2007 8131
Figure 2. Organization within the artificial cell-like systems. Optical microscopy images of (A) rhodamine-labeled lipid membrane fluorescence and (B) Alexa Fluor 488 phalloidin-labeled actin filaments before and (C) after deconvolution ImageJ filtering. This is the cortex-distributed actin filament system. These artificial cells were formed with DOPC/rhod-DOPE/K+ ionophore (98:2:0.1 molar ratio). Scale bar, 15 µm.
unilamellar vesicles (5-200 µm in diameter), which can be made on the size scale of cells, have been formed in previous research;18,19 this approach is well positioned to mimic biological structure. We use electroformation to prepare our cell-like structures (Figure 1A) and then visualize them with fluorescence, confocal, and atomic force microscopy (Figure 1B). We hydrate a lipid film, in the presence of a low-voltage electric field, using actin to obtain G-actin-loaded vesicles. We then initiate the polymerization of G-actin by introducing potassium ions through ionophores in the F-actin polymerization buffer (Figure 1C). These artificial structures are round with diameters of 9.6 ( 5.7 µm (n ) 200). To localize specific molecules in our system, we use fluorescence microscopy with Rhod-DOPE for labeling the lipid membranes and Alexa Fluro 488-phalloidin for the filamentous actin (Figure 2). Figure 2A is a rhodamine isothiocyanate-channel image of the lipid bilayers that reveals unilamellar structures; fluorescein isothiocyanate (FITC) of our control (liposomes without actin) displays similar patterns. When phalloidin is added during the initial film composition, the location of the actin can be determined using the FITC channel. When using raw images with FITC imaging (Figure 2B) followed by deconvolution with ImageJ because of the out-of-plane fluorescence (National Institutes of Health),20 F-actin can be observed to co-localize to the vesicle membrane to form a cortex (Figure 2C). Because the systems are 3-D and the actin filaments may be out of the plane, visualizing a single plane within the cell-like structures may not adequately capture the distribution of linear filaments. To determine the distribution of actin filaments inside the vesicles in three dimensions, we used a laser-scanning confocal microscope. This technique reveals a second distribution of actin within our artificial system. Figure 3A,C shows images of the top of the vesicles. Figure 3B,D shows images that are deconvolved through the method described previously for Figure 2C. Figure 3A demonstrates that actin filaments appear to be co-localized with the lipid membrane and suggests that these filaments are associated with the inner surface of the membrane. Figure 3C displays a second mode where the filaments appear to be distributed within our artificial cells through the interior; this is similar to general actin networks that have been observed both in vitro and in vivo. This is accomplished without the addition of actin-binding proteins such as R-actinin or filamin in order to demonstrate the organization in a relatively unmodified form without the added complication of binding proteins within these vesicles. To quantify the distribution of our molecules in this artificial cell system, we analyzed the intensity distribution with ImageJ. Figure 3E is an intensity distribution corresponding to the line in Figure 3B. Two separate peaks (labeled 1 and 2) are (18) Menger, F. M.; Keiper, J. S. Curr. Opin. Chem. Biol. 1998, 2, 726-732. (19) Meleard, P.; Gerbeaud, C.; Bardusco, P.; Jeandaine, N.; Mitov, M. D.; Fernandez-Puente, L. Biochimie 1998, 80, 401-413. (20) Yi, Q.; Coppolino, M. G. Biotechniques 2006, 40, 745-746.
Figure 3. Laser scanning confocal microscope images of Alexa Fluor 488 phalloidin-labeled actin filaments. (A and C) Images captured in the x-y plane. (B and D) Images after processing with deconvolution filtering in ImageJ. (E-G) Intensity profiles for the lines depicted in B and D, respectively. The peak locations (1-4) are labeled in the images that correspond to the labeled peaks in the intensity plots. Scale bar, 5 µm.
8132 Langmuir, Vol. 23, No. 15, 2007
Zhang et al.
Figure 4. Characterization of the cell-like structures using tapping mode AFM. Images of (A, C, E) actin-loaded liposomes and (B, D, F) control liposomes lacking actin filaments. The (A, B) height, (C, D) amplitude, and (E, F) 3-D images are shown.
observed, which reveal the location of the actin filaments relative to the membranes. The double peaks are indicative of the boundary of our system from the lipid bilayer. Figure 3F,G shows intensity distributions for the lines in Figure 3D. A single peak is found in each of these profiles (labeled 3 and 4), indicating that the actin filaments are not attached to the vesicle membranes because of the location of the peaks. Previously, actin filaments have been introduced into liposomes and anchored through electrostatic forces to form a shell.21 However, actin filaments exist not only at the membrane but also in other forms, such as bundles and networks, distributed through the internal space (i.e., the cytoplasm in a living cell). In our system, we simply introduce the actin and phalloidin in the initial configuration. The actin filaments either appear to be associated with the inner membrane or are alternately distributed through the interior of the vesicle. To develop a cell-like system that mimics the mechanical properties of a living cell, our approach focuses on having both membrane and actin filament components and then characterizing (21) Limozin, L.; Roth, A.; Sackmann, E. Phys. ReV. Lett. 2005, 95, 178101.
the organization. We examine our cell-like system through the use of an AFM. A series of AFM images (height, amplitude, and 3-D) are shown in Figure 4. Comparing the height and amplitude of an actin-loaded cell-like system (Figure 4A,C) to those of a non-actin-loaded (control) system (Figure 4B,D) reveals an increase in height for the non-actin-loaded liposomes. The actinloaded liposomes appear to be more highly distributed over the surface of the substrate. It is likely that internal actin filaments not only provide mechanical support but also alter inner structures to allow for increased surface contact area to occur, which induces the additional spreading. To understand the mechanical properties of these artificial cells and to analyze their structure comparatively in both in vivo and in vitro studies, we used the AFM to produce forceindentation curves on our cell-like structures. Because of the characteristics of our hybrid system and the associated deformation of our structurally reinforced liposomes, tapping mode AFM was determined to be most appropriate for examining the mechanically linked properties of our artificial cells. A series of
Mechanical Response of Biological Systems
Figure 5. Representative deflection-distance curves for the (‚‚‚) control liposomes, (---) actin-loaded liposomes, and (-) substrate. This data was used to characterize the mechanical properties of our artificial cell-like systems.
force-deflection curves were obtained (Figure 5). Young’s modulus, E, is determined for these AFM studies through a Hertz model, which relates the applied force, F, to the indentation, d, of a tip of radius R.17
E≈
x3π/4R(1 - γ2)F
x3 d2 0.5,22
Because the Poisson ratio, γ, is this yields a Young’s modulus of E ) 6.3 ( 0.6 MPa for control liposomes without actin filaments. This value is relatively close to that in a previous report for pure DOPC large unilamellar vesicles, which was determined using quasi-elastic light scattering,22 and other composition liposomes.23,24 Furthermore, using this same approach, we find that the Young’s modulus for our artificial celllike structure is 26.3 ( 15.4 MPa. Note that whereas these experiments were conducted with phalloidin, researchers have previously shown that phalloidin does not affect the mechanical properties of actin filaments.25,26 In our system, the actin filaments increase the Young’s modulus in these vesicles by adding structural elements. When comparing these values with the reported values of actin filaments for in vitro and in vivo studies, a challenging issue occurs as a result of the wide range of reported values for different cell lines and experimental techniques.27 Young’s modulus of bulk actin in vitro is relatively consistent and has been reported to be 2.4 GPa through stretching experiments of actin filaments28 and 1.8 GPa through torsional micromanipulation.29 However, the elasticity of in vivo cells has a tremendous range as evident by Young’s moduli values that are orders of magnitude lower than that for bulk actin for cells (22) Hantz, E.; Cao, A.; Escaig, J.; Taillandier, E. Biochim. Biophys. Acta 1986, 862, 379-386. (23) Liang, X.; Mao, G.; Ng, K. Y. J. Colloid Interface Sci. 2004, 278, 53-62. (24) Mao, G.; Liang, X.; Ng, K. Y. In Dekker Encyclopedia of Nanoscience and Nanotechnology; Schwarz, J. A., Contescu, C. I., Putyera, K., Eds.; Marcel Dekker: New York, 2004; pp 923-932. (25) Prochniewicz-Nakayama, E.; Yanagida, T.; Oosawa, F. J. Cell Biol. 1983, 97, 1663-1667. (26) Kishino, A.; Yanagida, T. Nature 1988, 334, 74-76. (27) Mathur, A. B.; Collinsworth, A. M.; Reichert, W. M.; Kraus, W. E.; Truskey, G. A. J. Biomech. 2001, 34, 1545-1553. (28) Kojima, H.; Ishijima, A.; Yanagida, T. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 12962-12966. (29) Tsuda, Y.; Yasutake, H.; Ishijima, A.; Yanagida, T. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 12937-12942.
Langmuir, Vol. 23, No. 15, 2007 8133
such as those in skeletal muscle (i.e., 24.7 ( 3.5 kPa), which was measured using atomic force microscopy.27 Our Young’s modulus values, which are for a hybrid system between actin in vitro and cells in vivo, are intermediate ones. In this study, we investigated the structural and mechanical response of an artificial cell-like system in order to gain further insight into cell and polymer behavior. We accomplished this by using a system that mimics the cell cytoskeleton in an artificial environment composed of giant liposomal vesicles. By visualizing our system with epifluorescence and laser confocal microscopy, we observed that actin filaments remained inside the artificial cell membrane near the bilayer or were distributed as filaments within the interior of our vesicles; these structural modes are related to the response of these artificial cells. Furthermore, by using an AFM to capture images and force curves, the actin filaments appear to affect the mechanical properties so that they are between a cellular and a polymer response. Our results indicate that this approach has a strong potential for bridging technology between the cell and biopolymer worlds. Whereas our system does not condense all of the studies into one general theory, our work does provide a significant advance using a novel approach to understand the properties and organization within living cells and biopolymers by developing and employing a system that incorporates two major model elements in cell mechanics: the cell membrane and the actin cytoskeleton. These results will be useful in a wide range of fields, including polymer physics, cell biology, and biotechnology. Materials and Methods Chemicals. Lipids 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and 1,2-dioleoyl-sn-glycero-3-phosophoehtanolamine-N(lissamine rhodamine B sulfonyl) (DOPE-Rhod) as CHCl3 solutions are from Avanti Polar Lipids (Alabaster, AL). Potassium ionophore is from Sigma (St Louis, MO). Skeletal muscle actin (99%), general actin buffer, ATP stock, and actin polymerization buffer (F-buffer) are from Cytoskeleton (Denver, CO). G-buffer was obtained from general actin buffer containing 0.2% (v/v) ATP stock. Alexa Fluor 488 phalloidin and DNase I are from Invitrogen (Carlsbad, CA). Liposome Preparations. Vesicles containing actin were made through electroformation.30 In brief, a mixture of DOPC, DOPERhod, ionophores, and Alexa Fluor 488 phallodin (98:2:0.1:0.1 molar ratio) in the solvent chloroform/methanol (5:3 v/v) was dried in a dry N2 stream. Either G-actin in G-buffer (0.4 mg/mL) (for structured cells) or G-buffer (for control liposomes) was added to the chamber, which was clamped together between two pieces of glass covered with a conductive indium tin oxide (ITO) film, resulting in a final concentration of 0.25 mg/mL. An ac electric field (1 V, 8 Hz) was applied for 4 h at 20 °C, and then the G-actin was polymerized by introducing potassium ions through ionophores. The polymerization of unencapsulated G-actin in the solution was inhibited by DNase I. Microscopy. For the optical microcopy images, an Axiovert inverted Zeiss microscope was used with epifluorescnce and differential interference contrast. The images were captured with an Insight digital camera and 63× (1.4 numerical aperture) oil objectives with tetramethyl rhodamine isothiocyanate and fluorescein isothiocyanate filters for the fluorescent images. Laser scanning confocal microscopy was performed on an Olympus IX-50 microscope with an Olympus 60× (1.4 numerical aperture) oil objective and a Laser Physics argon laser. Images were collected and processed with QED Imaging Software (Pittsburgh, PA) on a Macintosh G4 computer. AFM imaging and force measurements were performed with a Nanoscope III multimode scanning probe microscope (Digital Instruments, Santa Barbara, CA) using a tapping fluid cell with an (30) Hackl, W.; Barmann, M.; Sackmann, E. Phys. ReV. Lett. 1998, 80, 17861789.
8134 Langmuir, Vol. 23, No. 15, 2007
Zhang et al.
O-ring. A V-shaped oxide-sharpened Si3N4 cantilever was used with a nominal spring constant of 0.5 N/m. Freshly cleaved mica cleaned with double-sided tape was used as the substrate. The imaging and data processing were performed on the basis of previously published techniques.17
This work was supported in part by a National Science Foundation Career Award, the National Academies Keck Foundation Futures Initiative, the Pennsylvania Infrastructure Technology Alliance, the Department of Energy-Genome to Life, and a Beckman Young Investigators Award.
Acknowledgment. We especially thank Tomasz Kowalewski in the Department of Chemistry at CMU for the AFM studies.
LA700488P
