Electroinjection of Colloid Particles and Biopolymers into Single

A combined electroporation and pressure-driven micro- injection method for efficient loading of biopolymers and colloidal particles into single-cell-s...
0 downloads 0 Views 470KB Size
Download PDF
Anal. Chem. 2000, 72, 5857-5862

Technical Notes

Electroinjection of Colloid Particles and Biopolymers into Single Unilamellar Liposomes and Cells for Bioanalytical Applications Mattias Karlsson,† Kerstin Nolkrantz,† Maximilian J. Davidson,† Anette Stro 1 mberg,† Frida Ryttse´n,† Bjo 1 rn Åkerman,‡ and Owe Orwar*,†

Department of Chemistry, Go¨teborg University, SE-412 96 Go¨teborg, Sweden, and Department of Physical Chemistry, Go¨teborg University, Chalmers, SE-413 45, Go¨teborg, Sweden

A combined electroporation and pressure-driven microinjection method for efficient loading of biopolymers and colloidal particles into single-cell-sized unilamellar liposomes was developed. Single liposomes were positioned between a ∼2-µm tip diameter solute-filled glass micropipet, equipped with a Pt electrode, and a 5-µm-diameter carbon fiber electrode. A transient, 1-10 ms, rectangular waveform dc voltage pulse (10-40 V/cm) was applied between the electrodes, thus focusing the electric field over the liposome. Dielectric membrane breakdown induced by the applied voltage pulse caused the micropipet tip to enter the liposome and a small volume (typically 50-500 × 10-15 L) of fluorescein, YOYO-intercalated T7phage DNA, 100-nm-diameter unilamellar liposomes, or fluorescent latex spheres could be injected into the intraliposomal compartment. We also demonstrate initiation of a chemical intercalation reaction between T2-phage DNA and YOYO-1 by dual injection into a single giant unilamellar liposome. The method was also successfully applied for loading of single cultured cells. Today, there is a growing interest in incorporating submicrometer-sized sensing, sampling, and signal-amplifying particles, as well as large biopolymers, into single cells and liposomes. Several ultrasensitive detection and sensing methods are based directly or indirectly on the use of colloidal particles. Examples are quantum dot bioconjugate sensors,1,2 the family of PEBBLE (probes encapsulated by biologically localized embedding) sensors,3 and Ag and Au colloids for use in surface-enhanced Raman spectroscopy (SERS) measurements.4-6 One of the main limitations for practically using these techniques is the difficulty of * Corresponding author: (e-mail) [email protected]. † Department of Chemistry. ‡ Department of Physical Chemistry. (1) Bruchez, M., Jr.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P. Science 1998, 281, 2013-2016. (2) Chan, W. C. W.; Nie, S. Science 1998, 281, 2016-2018. (3) Clark, H. A.; Kopelman, R.; Tjalkens, R.; Philbert, M. A. Anal. Chem. 1999, 71, 4837-4843. (4) Chourpa, I.; Morjani, H.; Riou, J.-F.; Manfait, M. FEBS Lett. 1996, 397, 61-64. 10.1021/ac0003246 CCC: $19.00 Published on Web 11/03/2000

© 2000 American Chemical Society

noninvasive and quantitative introduction of colloidal particles into the intracellular compartment of a cell.7 Furthermore, it would be attractive to direct the introduction of particles into specific subcellular compartments such as the cytosol, nucleus, or even organelles of individual cells. Giant unilamellar vesicles (GUVs) are cell-sized liposomes composed of a single lipid bilayer with an entrapped aqueous compartment.8,9 Liposomes are attractive to use as ultrasmall reaction containers in which the reaction under study is confined and separated from the external medium. As such they can be used for studies of biochemical reaction dynamics in compartments mimicking a natural intracellular-intraorganellar environment.10-13 Liposomes also have great potential for use as ultrasmall-scale reaction containers for chemical derivatizations in miniaturized separation techniques. For use as reaction containers, it is necessary to load vesicles with reactants, including biopolymers such as DNA and colloid particles or organelles (syntethic or naturally derived). Loading of liposomes can, in principle, be performed by adding the particles during the preparation of the vesicles, since they upon formation trap a part of the medium in which they are formed. The trapping efficiency for small liposomes is, however, limited even for low-molecular-weight compounds, and entrapment (5) Sharonov, S.; Nabiev, I.; Chourpa, I.; Feofanov, A.; Valisa, P.; Manfait, M. J. Raman Spectrosc. 1994, 25, 699-707. (6) Beljebbar, A.; Morjani, H.; Angiboust, J. F.; Sockalingum, G. D.; Polissiou, M.; Manfait, M. J. Raman Spectrosc. 1997, 28, 159-163. (7) Clark, H. A.; Hoyer, M.; Philbert, M. A.; Kopelman, R. Anal. Chem. 1999, 71, 4831-4836. (8) Lasic, D. D., Ed. Liposomes: from physics to applications, 1st ed.; Elsevier Science B. V.: Amsterdam, 1993. (9) Luisi, P. L., Walde, P., Eds. Giant Vesicles; Perspectives in Supramolecular Chemistry 6; John Wiley & Sons: Chichester, England, 2000. (10) Bucher, P.; Fisher, A.; Luisi, P. L.; Oberholzer, T.; Walde, P. Langmuir 1998, 14, 2712-2721. (11) 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. (12) Chiu, D. T.; Wilson, C. F.; Karlsson, A.; Danielsson, A.; Lundqvist, A.; Stro ¨mberg, A.; Ryttse´n, F.; Davidson, M.; Nordholm, S., Orwar, O.; Zare, R. N. Chem. Phys. 1999, 247, 133-139. (13) Oberholzer, T.; Nierhaus, K. H.; Luisi, P. L. Biochem. Biophys. Res. Commun. 1999, 261, 238-241.

Analytical Chemistry, Vol. 72, No. 23, December 1, 2000 5857

of larger structures such as colloids is of very low probability.14,15 Another approach for liposome loading is to introduce the materials into preformed vesicles. This is possible by utilizing microinjection techniques.16 By using microneedles made out of pulled borosilicate capillaries with outer tip diameters in the range of 200-250 nm, it is possible to mechanically penetrate the membrane wall of a liposome, or cell, and eject controlled volumes of a desired reagent into their interiors.17,18 The injection volume is typically in the femtoliter to attoliter range and controlled by regulation of injection time and injection pressure. The pressure is usually generated by utilization of pressurized-air systems. There are also reports of pressure control by means of thermal manipulation, where an injection needle first is back-filled with the materials to be introduced followed by a thermosensitive expansion medium, such as galinstan.19 The needle is then hermetically sealed in the back end, and upon heating, the medium inside the needle is forced to expand, thereby ejecting the analyte solution. With this arrangement it is possible to obtain a high degree of control over the volume injected. Whereas microinjection works well with certain cell types, they are difficult to apply to synthetic cell-sized vesicles with thin walls.17,18 A lipid membrane bilayer is, typically, very elastic and the absence of internal supporting structures in unilamellar liposomes makes them very difficult to penetrate by mechanical means. The outer diameter of a tip suitable for injection into thinwalled lipsomes and smaller cells is ∼200 nm, and the inner diameter is typically in the range of only 100 nm.17,20 Such tips are very fragile and extremely difficult to view in a light microscope, making positioning difficult. The main drawback of using small inner diameter injection tips is that they are restricted for use with materials that are much smaller than the pipet channel diameter and require ultrapure injection liquids in order to prevent clogging. Microinjection techniques are considered to be relatively invasive due to the large mechanical forces applied, that potentially can induce permanent membrane damage and even membrane rupture on cells and liposomes. An alternative approach to single-liposome loading is to use microelectroporation.21 This technique is based on the theory of electropermeabilization. When liposomes, or cells, are exposed to an electrical field of a sufficient strength, small pores will form in their membrane. In microelectroporation, the analyte to be encapsulated is added to the exterior solution of the liposomes, or cells, and an electrical field is then applied locally, using microelectrodes. The amount of analyte that enters the vesicle is dependent on the analyte concentration gradient, membrane potential, duration of the applied field, and diffusion rate of the analyte. Electropermeabilization techniques are considered to be (14) Monnard, P.-A.; Oberholzer, T.; Luisi, P. L. Biochim. Biophys. Acta 1997, 1329, 39-50. (15) Shew, R. L.; Deamer, D. W. Biochim. Biophys. Acta 1985, 816, 1-8. (16) Graessmann, A.; Graessmann, M.; Mueller, C. Methods Enzymol. 1980, 65, 816-825. (17) Wick, R.; Angelova, M. I.; Walde, P.; Luisi, P. L. Chem. Biol. 1996, 3, 105111. (18) Menger, FM.; Gabrielson, K. D. Angew. Chem., Int. Ed. Engl. 1995, 34, 2091-2106. (19) Knoblauch, M.; Hibberd, J. M.; Gray, J. C.; van Bel, A. J. E. Nat. Biotechnol. 1999, 17, 906-909. (20) Davis, B. R.; Yannariello-Brown, J.; Prokopishyn, N. L.; Luo, Z.; Smith, M. R.; Wang, J.; Carsrud, N. D. V.; Brown, D. B. Blood 2000, 95, 437-444. (21) Lundqvist, J. A.; Sahlin, F.; Åberg M. A. I.; Stro¨mberg A.; Eriksson P. S.; Orwar, O. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 10356-10360.

5858

Analytical Chemistry, Vol. 72, No. 23, December 1, 2000

mild methods since the pore formation is reversible and the membrane will basically heal itself when the electric field is removed.22 Drawbacks are difficulties of quantitative loading, and loading of structures of sizes larger than the effective pore diameter, which for electropermeabilized erythrocytes is in the range of 1-240 nm.23,24 Here we demonstrate a novel approach for introducing largemolecular-weight compounds as well as colloid particles into GUVs, and cells, by combining the concepts of micropipet-assisted injection and electroporation. The basic idea is to destabilize the membrane with electric pulses prior to the injection procedure, inducing pore formation, which facilitates penetration of the micropipet tip allowing the use of large tip diameter microneedles. The benefits of such an arrangement would be the combination of the high degree of spatial and volume control of microinjection and the noninvasiveness of electroporation, promoting quantitative introduction of analytes, including colloids, into unilamellar liposomes and cells. In addition, the micropipet and microelectrode makes a pair of tweezers allowing trapping and subsequent injection into free-floating cells and liposomes. MATERIALS AND METHODS Chemicals. N-3-Triethylammoniumpropyl-4-(4-(dibutylamino)styryl)pyridinium dibromide (FM 1-43), 3,3′-dioctadecyloxacarbocyanine perchlorate (DiO), YOYO-1 iodide (oxazole yellow dimer), and FluoSpheres (30- and 200-nm diameter) were from Molecular Probes. Fluorescein (GC grade), T2 DNA (168 000 bp25), T7 DNA (39 936 bp26), L-R-phosphatidylcholine (type II-S from soybean), potassium phosphate (>98%), and Trizma base (>99.9%) were purchased from Sigma. Chloroform, EDTA (titriplex III), magnesium sulfate, and potassium dihydrogen phosphate (all pro analysi) were obtained from Merck. Glycerol (>99.5%) from J. T. Baker and deionized water from a Milli-Q system (Millipore) was used. Formation of Small Unilamellar Vesicles (SUVs). To prepare SUVs, acetone-purified27 preparations of L-R-phosphatidylcholine lipids were diluted with chloroform to a lipid concentration of 10 mg/mL. For a standard preparation, 300 µL of this solution was transferred to a round-bottomed flask. The solvent was removed on a rotary evaporator for ∼6 h at room temperature. A thin completely dry lipid film had then formed on the walls of the flask. To this film, PBS buffer (Trizma base 5 mM, K3PO4 30 mM, KH2PO4 30 mM, MgSO4 1 mM, EDTA 0.5 mM, pH 7.8.) containing 1% v/v glycerol was carefully added to a lipid concentration of 1 mg/mL. The lipid film was allowed to swell overnight at 4 °C. Finally, the sample was sonicated in a bath-type sonicator filled with ice-water. A total sonication time of ∼10 min was normally required before the entire lipid film dissolved and a whitish opalescent mixture was formed. SUV suspensions could be stored at 4 °C for several days and still produce large yields of GUVs prepared as described below. Formation of GUVs. The formation of GUVs was performed in a two-step procedure; dehydration of the lipid dispersion followed by rehydration. (22) (23) (24) (25) (26)

Weaver, J. C. J. Cell. Biochem. 1993, 51, 426-435. Chang, D. C.; Reese, T. S. Biophys. J. 1990, 58, 1-12. Kinosita, K. Jr.; Tsong, T. Y. Nature 1990, 268, 438-441. Harpst, J. A.; Dawson, J. R. Biophys. J. 1989, 55, 12371249. Oakley, J. L.; Coleman, J. E. Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 42664270. (27) Miller, X. J. Membr. Biol. 1976, 26, 319-333.

For dehydration, a small volume (5 µL) of SUV suspension was carefully placed on a borosilicate cover slip and placed in a vacuum desiccator at 4 °C. When the sample was completely dry (no sign of “fluidness” in the microscope), the dehydration was terminated and the sample was allowed to reach room temperature before rehydration. The dry sample was first rehydrated with 5 µL of buffer. After 3-5 min, the sample was further diluted with buffer; this was done very carefully to minimize turbulence in the sample. All rehydration liquids were at room temperature. Lamellarity Investigations. To estimate the lamellarity of the liposomes formed, the method developed by Schneider et al.28 was utilized. Giant liposomes were prepared and stained with the positively charged membrane dye FM 1-43. To 5 µg of lipids, 1 ng of dye (dissolved in distilled water) was added. Fluorescence intensity profiles were obtained by using a confocal laser fluorescence scanning microscope (Leica DST, Wetzlar, Germany), performing one-dimensional (linear) scans in the equatorial plane through the liposomes. Lamellarity was also investigated through scanning electron microscopy, where a GUV preparation was immobilized by a gelatin gel-casting technique and subsequently fixed by treatment with osmium tetraoxide-glutaraldehyde before sectioning in a microtome. Fluorescence Staining of DNA. For fluorescence microscopy, T7-phage DNA was stained with YOYO-1 at a ratio of 1 dimer per 10 base pairs, using a mixing and equilibration protocol29 developed to avoid inhomogeneous staining of the DNA. More than 95% of the DNA was in its intact linear form as determined by field-inversion gel electrophoresis (T+ ) 3 s, T- ) 1.5 s; 7.5 V/cm; 1% agarose). T7 DNA and the dimer YOYO-1 of oxaxole yellow were used without further purification. Micromanipulation and Electroinjecton. All injection experiments were performed on an inverted microscope (Leica DM IRB) equipped with a Leica PL Fluotar 40× objective and a water hydraulic micromanipulation system (high-graduation manipulator, Narishige MWH-3, Tokyo; coarse manipulator, Narishige MC-35A, Tokyo). Fluorescence imaging was achieved by sending the output of an Ar+ laser (Spectra-Physics 2025-05, 488 nm) through a 488nm line interference filter followed by a spinning disk to break the coherence and scatter the laser light. The laser light was collected by a lens and was sent through a fluorescein filter (Leica I-3) into the objective to excite the fluorescent dyes. The fluorescence was collected by the objective and detected by a three-chip color CCD camera (Hamamatsu, Kista, Sweden) and recorded on VHS (Panasonic S-VHS AG-5700). Digital image editing was performed using an Argus-20 system (Hamamatsu) and Adobe Photoshop graphic software. The electroinjections were controlled by a microinjection system (Eppendorf Transjector 5246, Hamburg, Germany) and a pulse generator (Digitimer Stimulator DS9A, Welwyn Garden City, U.K.). For translation of liposomes to different locations during the experiments, carbon fiber microelectrodes (ProCFE, Axon Instru(28) Schneider, M. B.; Jenkins, J. T.; Webb, W. W. J. Phys. (Paris) 1984, 45, 1457-1472 (29) Carlsson, C.; Jonsson, M.; Åkerman, B. Nucleic Acids Res. 1995, 23, 24132420.

Figure 1. Characterization of unilamellar liposomes. (A) A single surface-immobilized vesicle viewed with DIC optics. (B) Several freefloating unilamellar vesicles stained with FM 1-43 viewed in fluorescence. (C) Electron micrograph of one small1 and one large2 liposome embedded in gelatin matrix.3 For electron microscopy, liposomes were fixed by treatment with glutaraldehyde and osmium tetraoxide, cast in gelatin gel, and sectioned into 30-nm slices before examination. This treatment forced the vesicles to collapse, explaining the irregular appearance of the membrane. Sections of the membrane were, however, preserved and appeared to consist of a single bilayer membrane (arrow). This result was also corroborated through confocal laser scanning microscopy investigations. By performing linear scans through the equatorial plane of the vesicles (D) and analyzing the corresponding intensity profiles (E), the lamellarity could be estimated.

ments, Foster City, CA) controlled by the micromanipulation system were used. By simply pushing the vesicles with the microelectrodes, they detached from the surface, adhered to the electrode tips, and could be moved over long distances to a desired target. With this technique, it was also possible to detach unilamellar protrusion vesicles from multilamellar liposomes. Preparation of Injection Tips. Injection tips were prepared from borosilicate capillaries (length, 10 cm; o.d., 1 mm; i.d., 0.78 mm; Clark Electromedical Instruments, Reading, U.K.) that were carefully flame-forged in the back ends in order to make entrance into the capillary holder easier. The capillaries were flushed with a stream of nitrogen gas before use. The tips were pulled on a CO2 laser puller instrument (model P-2000, Sutter Instrument Co., Novato, CA). The outer diameter of the injection tips varied between 0.5 and 2.5 µm. To avoid contamination, tips were pulled immediately before use. Calibration of Injection Volumes. Injection volumes were estimated by injecting buffer solution into unilamellar vesicles attached to multilamellar protrusions. The multilamellar part of these vesicles acts as a membrane reservoir, and upon injection of solutions, this type of vesicle has the ability to increase its size in order to accommodate the enlarged volume of fluids. Injection volume was controlled through modulations of injection time and pressure and by injecting a fixed volume of buffer solution into Analytical Chemistry, Vol. 72, No. 23, December 1, 2000

5859

Figure 2. Schematic drawing of the capillary holder. The main purpose of the capillary holder is to act as a interface between the microinjection system and the pulse generator. The main body (II) of the holder was constructed from Plexiglas. It was equipped with a Pt wire electrode attached to a brass pin (III) that was connected to the low-voltage power supply, and an entrance for the microinjector outlet (I). The injection tips (VI) are firmly held in place by two rubber O-rings (IV) secured by a Delrin screw cap (V).

this type of vesicle and measuring the corresponding increment in vesicle radius; the injection volume could be calculated. RESULT AND DISCUSSION Formation of Giant Vesicles. The method developed for formation of GUVs by rehydration of a SUV suspension in this work is based on the method described by Criado and Keller.30 In contrast to other methods,31-33 the protocol here described is rapid and gives high yields of unilamellar vesicles in the size range of 1-300 µm in diameter even with physiological saline buffers. From a standard preparation where 5 µg of lipid was used, several hundred cell-sized liposomes (Figure 1A,B) judged unilamellar or thin-walled by scanning electron microscopy (Figure 1C) and confocal laser fluorescence scanning microscopy (Figure 1D,E) were observed. A majority of the unilamellar vesicles formed were interconnected by tethers, either to multilamellar protrusions or to other GUVs. Also, several free-floating GUVs were found. Electroinjection Procedures. The injection tips were backfilled with a medium of choice and mounted onto an in-house constructed pipet holder shown schematically in Figure 2. Basi(30) Criado, M.; Keller, B. U. FEBS Lett. 1987, 224, 172-176. (31) Angelova, M. I.; Dimitrov, D. S. Faraday Discuss. Chem. Soc. 1986, 81, 303. (32) Akashi, K.-I.; Miyata, H.; Itoh, H.; Kinosita Jr., K. Biophys. J. 1996, 71, 3242-3250. (33) Needham, D.; Evans, E. Biochemistry 1988, 27, 8261-8269.

5860 Analytical Chemistry, Vol. 72, No. 23, December 1, 2000

Figure 3. Electroinjection of fluorescein into a giant unilamellar vesicle. (A) DIC image showing two multilamellar liposomes with two adjacent unilamellar vesicles settled on the cover slip surface. The microelectrode and injection capillary were positioned in an opposing fashion close to the target vesicle. (B) A mechanical pressure was applied on the vesicle by moving the injection tip toward the microelectrode, forcing the vesicle into a kidneylike shape. (C) By applying a rectangular dc voltage pulse (40 V/cm, 3 ms), the membrane was permeabilized, the vesicle slid onto the injection tip, and a fluorescein solution (25 µM) was injected into the vesicle. (D) The injection tip and counter electrode were removed from the vesicle. (E) Fluorescence image of the vesicles after injection. The vesicle injected with fluorescein is exhibiting strong fluorescence while the other liposomes are unaffected. The contour lines of the unilamellar vesicles were digitally enhanced.

cally, the device is a standard patch-clamp pipet holder fitted to the outlet of a microinjection system. The pipet is equipped with a Pt electrode connected to a low-voltage pulse generator. The pipet-holder was mounted on the micromanipulation system described above. A carbon fiber microelectrode with a tip diameter of 5 µm was used as counter electrode. After selecting an appropriate GUV or cell, the injection tip and the microelectrode were positioned in an opposing fashion, in close contact with the vesicle at an angle of 10-30° and 150-170° with respect to the object plane (Figure 3). By careful positioning of the electrodes, it was possible to trap free-floating vesicles and subsequently perform injections. Initially, injections were attempted by first destabilizing the membrane with electric pulses and then penetrat-

Vm )1.5rsE cos R(1 - exp(-t/τ))

Figure 4. Injection of biopolymer and colloid particles into unilamellar vesicles. Since micropipet tips with outer diameters up to 2.5 µm could be used for injections, it was possible to inject biopolymers and colloidal dispersions into GUVs. The figure is showing fluorescence images of unilamellar vesicles injected with highly concentrated solution A of 30-nm fluorescent latex spheres, (B) small (100-nm) liposomes (50 µg/mL), stained with DiO, and (C) YOYO-1-labeled T7 DNA (5 ng/mL).

ing the membrane by performing a stab injection with the pipet. It proved, however, impossible to penetrate the membrane this way, but by reversing the process, membrane penetration was easily achieved. Consequently, by applying a mechanical pressure in terms of moving the injection tip toward the microelectrode, forcing the vesicle into a kidneylike shape (Figure 3B), it was possible to penetrate the membrane by applying the electric field. When permeabilized, the vesicle slid onto the injection tip and regained its spherical form (Figure 3C). In this mode, controlled volumes of materials contained in the micropipet could be injected into the liposome. In Figure 3C, a 25 µM solution of fluorescein was injected into a single liposome. Injection volumes were controlled by the microinjection system (injection pressure, 2501000 hPa; time, 0.1-1.5 s). Typically, a volume of 50-500 fL was injected into liposomes with a diameter of 10-20 µm. Injection volumes for cells were kept as small as possible in order to prevent cell trauma. After the injection was completed, the tip was withdrawn from the interior of the vesicle without noticeable signs of vesicle damage (Figure 3D) or leakage (Figure 3E). GUVs as well as cells were permeabilized in a single-pulse mode, by applying one or several transient rectangular dc voltage pulses with a pulse duration of 1-10 ms. The electric field strength was typically in the range of 10-40 V/cm. The membrane voltage Vm, at different loci on the membrane of a vesicle during exposure to a homogeneous electric field of duration t, can be calculated from

where E is the electric field strength, rs is the radius of the sphere, R is the angle in relation to the direction of the electric field, and τ is the capacitive-resistive time constant.34 Pore formation will occur at spherical coordinates exposed to the largest potential shift, which is at the poles facing the electrodes. Since the injection capillary also acts as an electrode, the coordinates of pore formation will spatially coincide with the loci where maximum mechanical force is applied. This electromechanical permeabilization proved to be a powerful technique for penetration of lipid membranes, allowing the use of coarse micropipet tips. When this procedure was used, it was possible to inject reagents into cells and GUVs with diameters of 8-25 µm using micropipet tips with an outer diameter of ∼2 µm. Such injection tips have sufficiently large inner diameters for injection of larger structures and colloid particles at high concentrations into vesicles. This is illustrated by the fact that YOYO-1-labeled T7-phage DNA molecules (RG ) 0.56 µm), 30-nm-diameter latex spheres, and 100nm-diameter SUVs were injected into unilamellar vesicles (Figure 4). Single PC12 cells were successfully injected with fluorescein (data not shown) as well as T7-phage DNA labeled with YOYO-1 (Figure 5). Moreover, preferential but not exclusive injection of materials could be performed into the cytoplasm (Figure 5B) and the nucleus (Figure 5D) Since the electroinjection technique described here can be performed with very high success rates, it may be a powerful tool for initiation of chemical reactions inside vesicles. This is illustrated by the experiment shown in Figure 6. By performing two consecutive injections of reagents into a single vesicle, an intercalation reaction between T2-phage DNA (RG ) 1.1 µm) and YOYO-1 was initiated. When viewed in the microscope, Brownian motion of the labeled DNA could be observed. This experiment illustrates, except for the fact that chemical reactions can be initiated this way, that it is possible to sequentially inject multiple reagents into a single vesicle without noticeable leakage. Initation of complex biochemical reactions inside the confines of a liposome would therefore be feasible. This also makes it possible to perform ultra-small-scale derivation chemistry inside a liposome, or cell, for analyte labeling prior to microchemical separations CONCLUDING REMARKS AND OUTLOOK One attractive application of the presented technology is the quantitative introduction of nanosensors1-3 or colloids for SERS measurements,4-6 into cells for detection of molecules or probing of intracellular structures and chemistry. Another application is introduction of these particles into liposomal reaction containers in which the reaction under study is confined and separated from the external medium, mimicking a natural intracellular-intraorganellar environment.9-12 Such a procedure would allow studies of complex biochemical reactions where the formation of several products and intermediates simultaneously could be monitored in real time. By injecting organelles (naturally or synthetically derived) into unilamellar vesicles, it is possible to create highly advanced cellular models. This would be very attractive for studies (34) Zimmermann, U. Biochim. Biophys. Acta 1982, 694, 227-277. (35) Bhalla, U. S.; Iyengar, R. Science. 1999, 283, 381-387. (36) Weng, G.; Bhalla, U. S.; Iyengar, R. Science. 1999, 284, 92-96.

Analytical Chemistry, Vol. 72, No. 23, December 1, 2000

5861

Figure 5. Electroinjection of YOYO-1-stained T7-phage DNA into PC-12 cells. (A) and (B) shows bright-field and fluorescence images of cells injected with fluorescent DNA into the cytosol. In (C) and (D), DNA was injected preferentially into the nucleus of the cell and the cytosol displayed faint fluorescence. In both experiments, the fluorescence is concentrated locally at the site of injection, indicating that diffusion of DNA-dye complex through the cell was restricted.

of, for example, complex biochemical signaling systems that translocate between different intracellular compartments. Such systems have recently gained a lot of interest from computer simulations showing emergent properties such as oscillations and bistability.35,36 Interestingly, such bistable behaviors have the capability to store information, in the same way as static RAM devices.35 Even if it is far beyond the scope of this article, it should thus be possible to create bistable memory liposomes by using synthetic organelles and reconstitutions of the appropriate signaling systems. If combined with the ultrathin injection needles used for conventional microinjections,19,20 the method described here may be a powerful technique for introduction of low- and mediummolecular-weight compounds into smaller cells or even organelles. This is due to the efficient membrane penetrative capacity of the electroinjection technique. The method is also potentially less invasive than traditional stab-microinjection protocols since there is less movement of the injection capillary when located inside the cell. Such movements may induce severe cell trauma caused by damage to the cellular matrix. Damage to the cellular membrane, provoked by the injection capillary during the penetration, should be less of an issue, since injections are performed into porous destabilized membranes with self-healing capability.

5862 Analytical Chemistry, Vol. 72, No. 23, December 1, 2000

Figure 6. Unilamellar vesicles as reaction containers. By performing two consecutive injections into a single vesicle, it was possible to initiate the intercalation reaction between T2 DNA and YOYO-1. (A) A unilamellar protrusion from a multilamellar liposome settled on the cover slip was selected as target. (B) First, a solution containing the T2 DNA (1 ng/mL) was injected into the vesicle using a micropipet tip with an outer diameter of 2 µm (40 V/cm, 4 ms). (C) Fluorescence image of the DNA-injected vesicle displayed no fluorescence. (D) The injection capillary was withdrawn and replaced by a thinner capillary with an outer diameter of 1 µm loaded with YOYO-1 (50 µM), and a second injection was performed (20 V/cm, 4 ms). (E) Fluorescence imaging after 10-min incubation revealed the presence of fluorescent YOYO-1- intercalated DNA molecules inside the vesicle. Brownian motion of micrometer-sized structures could be observed in the microscope, strongly suggesting that the fluorescence originated mainly from YOYO-intercalated DNA. The YOYO-1 dye, however, also had affinity for the lipid membranes as shown by the strong fluorescence originating from the multilamellar liposome. In control experiments, where only YOYO-1 and no DNA was injected into unilamellar liposomes, only weak and evenly distributed fluorescence from the liposome walls and no fluorescence from the interior solution was observed.

ACKNOWLEDGMENT This work was supported by the Swedish Foundation for Engineering Sciences (TFR) and the Swedish Foundation for Strategic Research (SSF). Received for review March 17, 2000. Accepted August 14, 2000. AC0003246