Cell Motility and the Cytoskeleton 35345-357 (1996) Effects of Electroporation on the Tubulin Cytoskeleton and Directed Migration of Corneal Fibroblasts Cultured Within Collagen Matrices Damien G. Harkin and Elizabeth D. Hay Department of Cell Biology, Harvard Medical School, Boston, Massachusetts Electroporation provides a useful method for loading fibroblasts with fluorescent probes for the cytoskeleton, but the possible deleterious effects of this loading technique on cell motility are unknown. We have used conventional and confocal microscopy of living cells and immunohistochemistry to examine the migration and cytoskeleton of chick embryo corneal fibroblasts electroporated while cultured within collagen gels. Fibroblasts cultured in collagen ( 1 mg/ml) are successfully electroloaded (0.5-1.0 kVcm-'/960 pF in DMEM/F12/20 mM Hepes, pH 7.2) with dextran (4-150 kDa) and immunoglobulin, but subsequently display uncoordinated pseudopodia and hence are unable to migrate effectively in any one direction. The lack of directed movement is due to depolymerization of microtubules and/or a perinuclear collapse of vimentin filaments, seemingly caused by millimolar levels of Ca2+ ions derived from culture medium following electroporation. Fibroblasts loaded in a buffer which resembles intracellular fluid (510 pM Ca2+)maintain their cytoskeleton and continue to migrate, when returned to culture medium within 10 min. Using this novel approach, we have loaded fibroblasts migrating through extracellular matrix (ECM) with rhodamine phalloidin and monitored the behavior of the labeled actin cortex by confocal microscopy. During migration phalloidin-actin accumulates near the base of pseudopodia and at the rear of the cell where it is subsequently left behind. We conclude that electroporation is a valuable technique for loading fibroblasts to study migration within ECM, provided that the conditions used support stability of the tubulin cytoskeleton. 0 1996 Wiley-Liss, Inc. Key words: cell motility, extracellular matrix, microtubules, vimentin, F-actin, live confocal microscopy INTRODUCTION bedded in collagen gels are difficult to load with probes for the cytoskeleton by using traditional methods such as The migratory behavior of fibroblasts, as for that of microinjection. We therefore chose, in the present study, many cell types, is readily studied using cells attached to to investigate the use of electroporation as a method for planar substrata [Abercrombie et al., 1970a, 1970b, loading fibroblasts suspended in collagen gels with mac1972; Vasiliev et al., 1970; Heath, 1983; Wang, 1985; romolecules. Theriot and Mitchison, 1992; Jones et al., 19931, but this Electroporation (electro-permeabilization, electromethod invariably introduces artifacts associated with injection) is widely used as a method for transfecting the cells having an upper free surface [Bard and Hay, 19751. Reconstituted collagen matrices [Elsdale and Bard, 19721 therefore provide a valuable means for Received May 20, 1996; accepted August 9, 1996. studying fibroblast migration within a 3-D substrate of Address reprint requests to Damien G. Harkin, Ph.D., Department of extracellular matrix (ECM) components similar to that Cell Biology, Harvard Medical School, 220 Longwood Avenue, Bosfound in vivo [Hay, 1985, 19911. However, cells em- ton, MA 021155729, 0 1996 Wiley-Liss, Ine. 346 Harkin and Hay cells with nucleic acid, but this method has also been used to load cells with a variety of other molecules including phalloidin [Hashimoto et al., 1989; Glogauer and McCulloch, 19921 and antibodies against the cytoskeleton [Wilson et al., 1991al. Although the precise mechanism of cell electro-permeabilizationis debatable, it is apparent that brief electrical pulses temporarily open plasma membranes to molecules which are normally too large or hydrophilic to enter cells by diffusion [Tsong, 1991; Neil and Zimmermann, 1993; Weaver, 1994, 19951. However, the conditions required for electroporation of cell membranes are often associated with a decrease in cell viability. Moreover, methods commonly used to assess the health of cells following electroporation, such as dye exclusion [Liang et al., 1988; Wilson et al., 1991b;Glogauer and McCulloch, 1992; Baum et al., 19941 or proliferation in culture [Tatsuka et al., 1988; Lukas et al., 1994; Wolf et al., 19943, provide a poor indication of the motile capacity of cells immediately following loading. The suitability of electroporation as a tool for studying cell migration is therefore unclear. In the present article, we examine the effects of electroporation on the permeability, migration and cytoskeleton of chick embryo corneal fibroblasts cultured within hydrated collagen gels. Corneal fibroblasts (keratocytes), like other types of cranio-facial mesenchyme, are derived from motile neural crest cells, and continue to migrate in the chick until 12 days of embryonic development [Hay and Revel, 19691. We establish optimal parameters for electroporating motile corneal fibroblasts by loading various sizes of fluorescent dextran and by measuring retention of cytoplasmic esterase activity with fluorescein diacetate. The motile behavior of fibroblasts following loading is monitored by both conventional and confocal time-lapse microscopy. We develop an electroporation buffer which preserves the normal migratory behavior of fibroblasts in collagen gels and facilitates stability of the tubulin cytoskeleton as demonstrated by immunohistochemistry. The effectiveness of this approach for studying the cytoskeleton, is demonstrated by monitoring the F-actin cortex of living cells electroloaded with rhodamine phalloidin. MATERIALS AND METHODS Isolation of Corneal Stroma Six-day-old White Leghorn chick embryos (Spafas, Norwich, CT) were dissected in Hanks' balanced salt solution containing 20 mM Hepes (HBSS-Hepes, pH 7.2) and their corneas removed using a 2 mm trephine blade (Storz Instrument Co., St. Louis, MO). Excised corneas were digested in 0.25% trypsidHBSS-Hepes for 15 min at 37°C to remove epithelia, then placed in DMEM/Fl2 medium supplemented with 5 % fetal bovine serum (EBS). Corneal stromas were then cut into several pieces and suspended in collagen gels. Preparation of Collagen Gels Type I collagen from rat tail (RTC) was obtained from Collaborative Biomedical Products (Bedford, MA) as a solution in 0.02 N acetic acid, and dialyzed at 4°C against a 10% solution of F12 medium (without bicarbonate; pH 4.0). The concentration of dialyzed RTC was determined by Sirius Red precipitation [Marotta and Martino, 19851and adjusted to 3.0 mg/ml with 10% F12 medium. A stock solution of polymerizing collagen was prepared on ice by combining dialyzed RTC, with a 10 times concentrated solution of F12 medium (without bicarbonate), 140 mM sodium bicarbonate and FBS, in the ratio of 7:l:l:l [Elsdale and Bard, 19721. The collagen solution was further diluted to 1.0 mg/ml with cold DMEM/F12/5% FBS and immediately dispensed as 50 pl aliquots into rings of Whatman No. 1 grade filter paper (inner diameter, 6 mm; outer diameter, 10 mm) placed in a plastic petri dish. One piece of corneal stroma was added to each ring of filter paper and set in place by polymerizing the collagen at 37°C. The corneal stromas in collagen were then cultured for 18-24 h in DMEM/ F12 supplemented with 5 % FBS, 0.4 mM L-glutamine, 50 pg/ml ascorbic acid, 2.5 p,g/ml fungizone, 5.0 pglml gentamicin, and 1% ITS + medium supplement (Collaborative Biomedical Products). Preparation of Macromolecules Tetramethyl rhodamine isothiocyanate (TRITC) dextrans (4 kDa and 76 kDa) and fluorescein isothiocyanate (FITC) dextrans (4, 40, 70, and 150 kDa) were obtained commercially (Sigma Chemical Co., St. Louis, MO) and dialyzed in PBS to remove trace amounts of unbound fluorophore. Fluorescein isothiocyanate (FITC; Sigma) was conjugated to chicken IgG (Sigma) as previously described [F'ryzwansky, 19821. All preparations of fluorescent macromolecules were equilibrated with electroporation buffer by dialysis, sterilized by filtration, and stored at 4°C. Rhodamine phalloidin (RPh; Molecular Probes, Eugene, OR) was dissolved in methanol at a concentration of 6.6 pM and stored at -20°C. For electroporation, 500 p1 of RPh stock was dried, dissolved in 1 ml of electroporationbuffer, and sterilized by filtration prior to use. Electroporation Buffers Three electroporation buffers were tested for effects on the motility and/or cytoskeleton of fibroblasts cultured in collagen gels: (1) serum-free DMEM/F12 without phenol red (Gibco BRL Cat. No. 21041-025; Life Technologies, Grand Island, NY), buffered with 20 mM Hepes, pH 7.2 (DFH); (2) cytoskeleton stabilizing Effects of Electroporation buffer (CSB) as used for immunostaining (refer below); and (3) a buffer referred to as “intracellular” electroporation buffer or ICEB, which we designed to mimic the ionic composition of intracellular fluid (125 mM KOH, 4 mM NaOH, 73 mM Pipes free acid, 34 mM myoinositol, 10 mM NaHC03, 5 mM K,HPO,, 5 mM KH,PO,, 5 mM D-glucose, 4 mM MgCl,, 1 mM MgSO,, 1-10 p M CaCl,, pH 7.0). 347 Video Microscopy A preliminary assessment of the effects of electroporation on the speed and direction of fibroblast migration were made using conventional light microscopy. Collagen gels containing corneal fibroblasts were placed in a 35 mm petri dish which had been modified for microscopy by cutting a 16 mm diameter hole in the bottom and attaching a 22 mm2 no. 0 glass coverslip with silicon glue. A 12 mm diameter round coverslip was then placed Electroporation on top of each gel and allowed to settle before adding 2 Collagen gels containing explants of corneal ml phenol red-free culture medium buffered with 20 mM stroma were removed from culture and washed three Hepes (pH 7.2) followed by 2 ml silicon oil. Cultures times for 10 min each in 10 ml serum-free DMEM/ were maintained at 37”C, covered with 5% CO, in air F12/20 mM Hepes (pH 7.2), then rinsed for an additional and mounted on a Zeiss IM-35 inverted microscope us5 rnin in three changes of electroporation buffer. Washed ing an LU-CB-1 tissue culture chamber and TC- 102 temgels were then incubated for 15 rnin at room temperature perature controller (Medical Systems Corp., Greenvale, (25°C) in electroporation cuvettes (0.4 cm electrode gap, NY). Time-lapse recordings of cell migration were made Bio-Rad Laboratories, Hercules, CA) containing 750 under phase (one every min for up to 2 h) using a Zeiss p1 of electroporation buffer with macromolecules. Cul- 63 X A.25 n.a. Neofluar oil immersion lens and DAGEtures were electroporated at room temperature using a MTI NC-66 X video camera (DAGE-MTI Inc., Michisingle exponential pulse delivered from a Bio-Rad Gene gan City, IN) interfaced with Metamorph image analysis Pulser transfection apparatus (Bio-Rad). Various combi- software (Universal Imaging, West Chester, PA). Loadnations of field strength (500-1,000 Vcm-’) and capac- ing of cells was confirmed at the end of each recording itance (125-960 FF) were tested for their effects on the by switching to epifluorescence. Distances migrated by uptake of macromolecules and cell viability. Immedi- cells were defined by the location of nuclei and deterately following electroporation, gels were removed from mined using the measure function of the Metamorph procuvettes and incubated for 10 min in fresh electropora- gram, calibrated with a micrometer slide. tion buffer at 37°C to facilitate resealing of cell membranes, then returned to room temperature and washed Live Confocal Microscopy for 1 h in several changes of DMEM/Fl2/Hepes (pH 7.2) Cultures of corneal fibroblasts in collagen were without phenol red, to remove excess macromolecules. prepared as above for video microscopy, then mounted Cell Loading and Viability Assay on a Zeiss LSM confocal microscope equipped with enCollagen gels containing electroporated fibroblasts vironmental chamber (37”C, 5% CO,). Paired pseudowere placed on a glass slide in phenol red-free culture Nomarski and confocal fluorescence images (average of medium and covered with a no. 0 glass coverslip. Slides 8 line scans) were obtained using a 40 x /1.3 n.a. Neowere mounted on a Zeiss (Thornwood, NY) Axioskop fluar lens and 488 nm or 568 nm krypton-argon laser line fluorescence microscope equipped with 200 watt mer- attenuated to 1%. The pinhole aperture for confocal flucury lamp and examined using a 63 x /1.25 n.a. Plan- orescence was adjusted to produce a ‘Z’ resolution of Neofluar oil immersion lens. At least 100 cells were approximately 1 pm. Images were collected manually at examined per gel and the percentage of fluorescent 15-min intervals for up to 2 h. loaded cells calculated. The cytoplasmic distribution of loaded macromolecules was confirmed by live confocal lmmunostaining microscopy. Collagen gels containing corneal fibroblasts were Cell viability was assessed by measuring the pro- rinsed briefly in cytoskeleton stabilizing buffer (CSB; portion of cells with sufficient cytosolic esterase activity 100 mM Pipes free acid, 160 mM NaOH, 20 mM myoto convert fluorescein diacetate to its fluorescent prod- inositol, 5 mM MgCl,, 5 mM EGTA, pH 6.8), extracted uct. Collagen gels containing fibroblasts were incubated with 1% Triton X-100/4%polyethylene glycol-8000 in for 10 min at 37°C in a 1 FM solution of fluorescein CSB (15 rnin), fixed with 2.0% (w/v) paraformaldehyde diacetate in phenol red free culture medium, then washed in CSB (15 min), washed in PBS (two times for 15 min), briefly in fresh medium before being examined as above. then incubated overnight at 4°C in 0.15 M Tris buffer Simultaneous measurements of loading and viability containing 50 mM NH,Cl and 1% normal goat serum were achieved by using cells electroporated in the pres- (NGS). Gels were then immunostained for tubulin or vimentin by incubation for 1 h at 37°C in 250 (1.1 of E7 ence of TRITC-dextran (4m a ) . 348 Harkin and Hay hybridoma supernatant containing IgG 1 raised against drosophila p tubulin, or H5 hybridoma supernatant containing IgGl raised against chick vimentin (both obtained from Developmental Studies Hybridoma Bank, Iowa City, IA; refer to "Acknowledgments"), followed by 1 h at 37°C in 250 pl of 1:lOO affinity purified, goat anti-mouse rhodamine-IgG (Boehringer Mannheim, Indianapolis, IN) in PBS with 1% NGS. Gels were washed for 1 h in PBS following incubation with each antibody. Stained gels were mounted in glycerol-PBS and viewed using a Zeiss LSM confocal microscope equipped with 100 x /1.3 Neofluar oil immersion lens. Staining of Fixed Preparations Cultures were fixed for 15 min in CSB containing 2% paraformaldehyde, then extracted for 5 min in CSB containing 1% Triton X-100 and 4% PEG. Fixed preparations were then washed in PBS and stained for 1 h at room temperature with 0.3 pM rhodamine phalloidin in PBS. Stained gels were subsequently washed in PBS and examined by confocal microscopy as above for immunostaining. Image Processing All micrographs were recorded digitally as TIFF format files and printed using a Fujix Pictrography 3000 thermal transfer printer (Crimson Tech, Cambridge, MA). Brightness and contrast modifications were made using Adobe Photoshop image processing software (version 3.0, Adobe Systems Inc., Mountain View, CA). RESULTS Electroporationof Fibroblasts in Collagen Gels Up to 80% of fibroblasts cultured in collagen gels are loaded with dextran (4 kDa) following electroporation in either DFH (serum free culture medium) or ICEB (intracellular buffer), using field strengths between 500 and 1,000 Vcm-' and a capacitance of 960 p F (Fig. 1). The time constants (time required for voltage to decay to 37% of its initial value) generated under these conditions are between 10 and 20 ms. Field strengths greater than 1.0 kVcm-' detach collagen gels from the supporting ring of filter paper. Capacitance settings between 125 and 500 pF do not support loading within the range of field strengths tested. Cell viability, assessed by fluorescein diacetate (FDi) conversion, declines with increasing voltage from 95% (no electroporation) to 61.7 2 7.4% (mean k s. d.) for DFH (1,000 Vcm-'), and to 30.0 15.6% for ICEB (1,000 Vcm-'). Parameters for optimal loading are 875 Vcm-'/960 p F for cells in DFH (viability, 71.7 ? 15.4% of total cells counted; loading, 53.3 2 30.6% of total cells counted) and 750 Vcm-'/960pF for cells in ICEB (viability, 72.3 2 7.8%; loading, 50.0 * f 8.9%).ICEB supports electro-loading of fibroblasts in collagen gels with different molecular weight dextrans ranging from 4-150 kilodaltons, as well as immunoglobulin (Table I). Non-viable (FDi negative) cells rarely retain low molecular weight dextran (TRITC-dextran, 4 m a ) . The cytoplasmic distribution of loaded macromolecules is confirmed by live confocal microscopy (Fig. 2). Effects of Electroporation of Fibroblast Migration Fibroblasts begin to migrate out of corneal stroma within 6-8 h after placing the tissue inside a collagen gel. The traction forces exerted by emigrating fibroblasts cause a radial alignment of nearby collagen fibers which, in combination with contact inhibition between the cells, promotes directed migration of cells away from the explant (Fig. 3). Following mitosis, however, one daughter cell may return towards the explant until contact inhibition reverses its direction. After 18-24 h in culture, 73% (n = 45) of cells continue to migrate in a directed manner 0.40 (s.d.) p d m i n . with an average speed of 0.56 During migration, filopodia (small cylindrical extensions of the cytoplasm [Trinkaus, 19731) continually extend and retract from a larger and more stable pseudopodal extension [Trinkaus, 19731, which is maintained in the direction of cell migration. Pseudopod branching is often observed, but additional branches wither and collapse back towards the cell. Additional filopodia are occasionally displayed at the rear during the approach of neighboring cells, but the forwards movement of cells is usually associated with detachment of cellular material from the rear. This behavior of control cells is consistent with that reported previously for corneal fibroblasts migrating through collagen matrices [Bard and Hay, 19751. Collection of confocal images at intervals of 15 min did not reveal noticeable changes in the motility of cells as compared with cells monitored by conventional light microsCOPY. Cells electroporated in culture medium (DFH), display similar filopodial activity to control cells, but in 90% of loaded cells (n = 11) the direction of pseudopodal extension continually changes in a random manner (Fig. 4). These changes in the direction of pseudopod protrusion mostly occur along the long axis of the cell, but extensions occasionally emerge at the side of the cell (Fig. 4; 45 min). Cells electroporated in culture medium therefore fail to migrate more than 10-15 pm in the first 2 h following their return to culture at 37°C. This behavior is observed following loading of cells in culture medium with either phalloidin or dextran. Extended observations of cell recovery reveal that 62% of cells electroporated in culture medium display evidence of directed migration after 3 h at 37"C, and 84% by 4 h. The motility of cells following electroporation in cytoskeleton stabilizing buffer (CSB; as used for immu- * Effects of Electroporation I 00 $T T 90 T I a * II 1 b T l T 349 T * T /I 30 20 10 r L 625 750 40 30 20 r 875 0 4 I 500 1000 0 Vcm-l/960 p F Fig. 1. Viability and loading of corneal fibroblasts cultured in collagen gels following electroporation in (a) culture medium (DMEM/ F12/20 mM Hepes, pH 7.2) or (b) intracellular electroporation buffer (ICEB). Viability (white bars) is demonstrated by fluorescein diacetate conversion following electroporation in the presence of TRITC- 1 625 1i 875 750 Vcrr1-~/960pF 1000 dextran (4 m a ) . Loading (black bars) is illustrated by uptake of fluorescent dextran (4 m a ) . Bars represent mean (? standard deviation) percentage of the total number of cells observed (2100). Asterisks indicate electroporation parameters used in subsequent studies of cell motility and the cytoskeleton. TABLE I. Electroloading of Fibroblasts in Collagen Gels With Dextran and IgG % Cells loaded, mean (SD) Molecule FITC-dextran 40 kDa 70 kDa 150 kDa FITC-IgG mg/ml Control 750 Vcm-’ 875 Vcm-’ I ,ooOVcm-’ 10 20 40 5 0.5 (0.7) 2.5 (3.5) 1.5 (2.1) 0.0 (0.0) 20.5 (10.6) 65.0 (9.9) 59.5 (3.5) - 65.0 (1.0) 62.3 (15.0) 67.7 (30.6) - 52.7 (7.8) 54.3 (15.9) 64.0 (21.2) 41.3 (4.2) nostaining) was not assessed due to poor viability, but cells loaded in CSB containing low calcium (1 pM), instead of EGTA (5 mM), display random pseudopod activity and, hence, fail to migrate effectively in any given direction (n = 6). However, following electroporation in intracellular buffer (ICEB), 83% of loaded cells display directed migration (0.46 0.19 pm/min, n = 18) that is similar to that performed by non-electroporated cells. The opportunity was therefore taken to examine the feasibility of studying changes in the F-actin cortex of migrating cells following loading of phalloidin (Fig. 5). Cells electroporated in ICEB containing 3 pM rhodamine phalloidin initially display a pale, even staining of the F-actin cortex (labeled in Fig. 5 as “C”), but leading cytoplasmic extensions are poorly stained when compared with fixed preparations (refer to Fig. 8) due to new actin polymerization in the time (30-45 min) between loading phalloidin and observing the cells. This phenomenon has been noted previously for cells micro- * injected with phalloidin [Wehland et al., 1977, 19801). During migration, phalloidin-stained F-actin accumulates at the rear of the cell (labeled “X” in Fig. 5) and at the base of pseudopodia (labeled “Y” in Fig. 5 ) . Sites of accumulation at the base of additional pseudopodia remain following the collapse of these extensions, and sites which develop at the rear remain fixed to the substrate as the cell migrates forwards. Effects of Electroporation on the Cytoskeleton Prior to electroporation, fibroblasts cultured in collagen gels display a dense network of microtubules which extend the length and breadth of the cell. However, 10 min after electroporation in culture medium (DFH), the majority of cells (275%) display only a few microtubules radiating from an organizing center (Fig. 6a). This degree of microtubule disruption is similar to that produced by treating cells for 1 h at 37°C with 2 350 Harkin and Hay Fig. 2. Effect of electroporation on uptake of fluorescent dextran by corneal fibroblasts cultured within collagen gels. Control cells (a,b) were incubated for 15 min at 25°C in “intracellular” electroporation buffer containing 10 mglml TRITC-dextran (76 m a ) , and washed for 90 min at 25°C in phenol red-free culture medium. These cells have failed to take up detectable levels of fluorescent dextran (b). Loaded cells (c,d) were treated as controls but electroporated at 750 Vcm-’/ 960 (LFprior to washing. The cytoplasmic uptake of fluorescent dextran by electroporated cells is confirmed by live confocal fluorescence (d). Cultures were incubated at 37°C for 30 min prior to image acquisition. pg/ml nocodazole and is only slightly improved by electroporation in the presence of 2 mM EGTA (data not shown). In contrast, cells electroporated in either cytoskeleton stabilizing buffer (CSB) or intracellular electroporation buffer (ICEB) display little or no evidence of microtubule disruption 10 min after electroporation (Fig. 6b,c, respectively). However, microtubules are disrupted when cells are electroporated in ICEB containing 2100 pM calcium chloride. One hour after returning electroporated cells to culture medium, cells electroporated in DFM display a more extensive microtubule network than that observed 10 min after electroporation, but few microtubules extend more than 10 pm from the organizing center (Fig. 6d). Almost complete recovery of microtubules is achieved within 3-4 h culture at 37°C. Cells electroporated in CSB dis- play completely disrupted microtubule networks following their return to culture medium (Fig. 6e). In contrast, cells electroporated in ICEB show only slight disruption of their microtubule networks which appears to be mainly restricted to the outermost parts of the cell (Fig. 6f). However, the degree of microtubule disruption is increased if cells electroporated in ICEB are maintained in this buffer for more than 5-10 min prior to their return to culture medium. A perinuclear collapse of vimentin intermediate filaments is associated with electroporation conditions which cause the disruption of microtubules (Fig. 7). This effect is reversible and follows a similar recovery period to microtubules. However, F-actin filaments are stable following electroporation of fibroblasts in culture medium (Fig. 8). Effects of Electroporation Fig. 3. Directed migration of chick embryo corneal fibroblasts away from explant of corneal stroma. This piece of corneal stxoma (bottom left) from 6-day-old chick embryo was cultured for 24 h inside a hydrated collagen matrix (1 mg/ml) before fixation with 2% paraformaldehyde in PBS. The orientation of cells, especially those furthest from the corneal stroma, is indicative of their directed migration away from the explant prior to fixation. DISCUSSION In summary, we demonstrate that corneal fibroblasts cultured within hydrated collagen matrices can be loaded with dextran (4-150 m a ) , immunoglobulin, and rhodamine phalloidin, using electroporation parameters (500-1,000 Vcm-'/960 pF) similar to those previously reported to permeabilize fibroblasts suspended in medium [Glogauer and McCulloch, 19921. However, the subsequent motility of these cells within ECM is dependent upon the choice of electroporation buffer, and the conditions which follow its use. Fibroblasts electroporated in serum-free culture medium fail to perform directed migration, seemingly because pseudopodia are projected randomly. Microtubules are disrupted and vimentin intermediate filaments adopt a pennuclear distribution, but actin filaments remain intact following electroporation of cells in culture medium. However, cells electroporated in a buffer which mimics the ionic composition of intracellular fluid (ICEB, 5 10 kM), continue to migrate through collagen with coordinated pseudopod formation and display little or no change in their cytoskeleton, when returned to culture medium within 10 min. Thus, ICEB supports loading of cells within collagen gels with low concentrations of rhodamine phalloi- 351 din and facilitates studies of the F-actin cortex during cell migration through ECM . We initiated this research, because we believe that it is important to study the migration of fibroblasts within a 3-D substrate of ECM. Planar substrata facilitate studies of cell migration by assisting optical resolution of cellular structures, but the existence of an upper free surface limits our understanding of how mesenchymal cells actually migrate when surrounded by ECM in vivo. Fortunately, developments in confocal microscopy now provide opportunities for a detailed analysis of the motile behavior of fibroblasts within more natural 3-D substrata. Indeed, we were able to record the migration of living fibroblasts in collagen gels at 15-min intervals for up to 2 h using attenuated laser light and a temperaturecontrolled incubation chamber. However, collagen gels hamper the microinjection of probes for labeling the cytoskeleton, because the fine injection needles required often clog or break upon entering the gel. Therefore, in the present study, we explored the possibility of loading corneal fibroblasts suspended within collagen gels by using electroporation. Electroporation proved to be the method of choice for introducing macromolecules into the cytoplasm of cells embedded in ECM. In addition to providing a noninvasive method for loading cells surrounded by ECM, electroporation loads large numbers of cells in a fraction of the time required for microinjection of a single cell. Since electroporation can affect cell viability [Baum et al., 1994; Wolf et al., 19941, we began our study by establishing optimal parameters for loading fibroblasts in collagen gels by this method. We soon found that although chicken fibroblasts are readily loaded in DMEMI F12 culture medium (DFH), the lack of directed migration for up to 3-4 h following electroporation makes this medium an unsuitable choice of electroporation buffer for studies requiring immediate observations of cell migration. For example, cells loaded with rhodamine phalloidin in culture medium may completely redistribute this F-actin probe before they resume normal migration. Vasiliev and others [Vasiliev et al., 1970; Bershadsky et al., 1991; Liao et al., 19951 have concluded that microtubules are necessary for the polarized morphology and directed migration of fibroblasts attached to planar substrata. Moreover, it is known that high levels of calcium ions such as those present in the DFH culture medium (-1 mM) inhibit the stability of microtubules in vitro [Borisy et al., 19761. We therefore considered that the lack of directed movement displayed by cells in collagen after electroporation in DFH might be due to depolymerization of microtubules following the entry of Ca2+ ions from the surrounding medium. The correlation that we found between microtubule stabilitylreformation and directed movement clearly supports this idea. 352 Harkin and Hay Fig. 4. Typical behavior of corneal fibroblast in collagen gel following electropration in culture medium. This cell was electroprated in DhtEM/F12/20mM Hepes (pH 7.2) containing 3 pM rhodamine phalloidin, then cultured at 37°C for 90 min prior to examination by confocal microscopy. While pseudopodia (p) and filopodia (f) are produced, the shifting location of these extensions results in little net migration. This poor migratory behavior is apparent for 2-3 h following electroporation in culture medium and is also displayed by cells loaded with fluorescent dextran. Loading of this cell was confirmed by confocal fluorescence (not shown). This conclusion is further supported by the fact that cells display apparently normal directed migration following electroporation in ICEB which contains low concentrations of Ca2+ ions (1 pM). However, factors other than calcium may also contribute to the disruption of microtubules since stability was only partially improved in the presence of EGTA (2 mM). Furthermore, a role for vimentin intermediate filaments in directed migration cannot be ruled out since they collapse to a perinuclear location in the absence of microtubules. It also remains possible that electroporation may have deleterious effects on organelles required for directed movement such as the Golgi apparatus [Bershadsky and Futerman, 19941. Interestingly, collagen fiber lattices do not support directed migration of fibroblasts in the absence of microtubules as do grooves cut into silicon wafers [Oakley and Brunette, 19951. In contrast to microtubules and vimentin filaments, F-actin filaments are apparently unaffected by electroporation in culture medium (Fig. 8). This finding is not surprising since cells continue to produce pseudopodia following electroporation in this medium. Corneal fibroblasts suspended in collagen gels display a dense cortical (ectoplasmic) network of F-actin filaments surrounding an endoplasm containing myosin 11, microtubules, and vimentin filaments [Tomasek et al., 1982; Tomasek and Hay, 19841. We have suggested that cortical F-actin behind the leading pseudopodium becomes fixed to the surrounding ECM via integrins, thus allowing the myosin-rich endoplasm supported by microtubules and/or vimentin filaments to slide forwards on the cortex [“fixed cortex” theory; Hay, 1985, 1989; Bilozur and Hay, 19891. The deposition of cortical cytoplasm containing plasma membrane, F-actin, talin [Daniels and Hay, 19901, and pl integrins [Regen and Horwitz, 1992; Palecek et al., 19961 in the wake of migrating cells is Effects of Electroporation Fig. 5 . Directed migration of corneal fibroblast through collagen following electroporation in intracellular buffer (ICEB). This cell was electroporated in ICEB containing 1 pM added CaCl, and 3 pM rhodamine phalloidin, then cultured at 37°C for 30 min prior to examination by confocal microscopy. Note that maintenance of an antenor pseudopod facilitates directed movement of the cell. The actin cortex (C) is initially stained evenly towards the middle and rear of the 353 cell, but the anterior extensions are poorly stained due to new actin polymerization. During subsequent forwards migration, phalloidinlabeled actin filaments accumulate at the rear of the cell (X) and near the base of pseudopodia (e.g., Y). These sites of F-actin accumulation appear to remain relatively fixed with respect to the surrounding collagen gel. 354 Harkin and Hay Fig. 6. Effect of electroporation buffer on microtubule stability. Corneal fibroblasts cultured in collagen gels were electroporated in either serum-free culture medium (DFH), cytoskeleton stabilizing buffer (CSB; as used for immunostaining), or intracellular buffer (ICEB), then placed at 37°C for 10 min to assist resealing of plasma mem- branes. Immunostaining for tubulin was then carried out immediately (a, DFH; b, CSB; c, ICEB) or 1 h after returning cells to 25°C culture medium (d,e,f, respectively). Each image is a projected composite of 10 1 pm optical sections. The apparent fragmentation of microtubules observed in b and c is an artifact of the optical sectioning. Effects of Electroporation 355 Fig. 7. Electroporation conditions which disrupt microtubules also cause a perinuclear collapse of vimentin intermediate filaments; (a) corneal fibroblasts immunostained for vimentin after 48 h culture inside a hydrated collagen gel; (b) cells immunostained for vimentin 1 h following electroporation in serum-free culture medium. Staining for F-actin (Fig. 8) reveals that these cells retain bipolar morphology after electroporation in culture medium. Fig. 8. Actin filaments are resistent to electroporation in culture medium. This cell was fixed and stained 10 min after electroporation in serum free culture medium. The phalloidin staining for F-actin in cytoplasmic extensions suggests that the stability of actin filaments is unaffected by electroporation conditions which cause disruption of microtubules and collapse of vimentin networks. consistent with this theory, as is the fact that actin mRNA [Lawrence and Singer, 19861 and newly synthesized plasma membrane locate near the leading edge [Kupfer et al., 19871. Along these lines, it is tempting to speculate that the lack of directed migration observed in the absence of microtubules might be mediated by an inability of the myosin-rich endoplasm to move on the actin cortex in their absence. Given the apparent requirement for microtubules and/or vimentin filaments for directed migration of fibroblasts through collagen lattices, we applied ourselves to the task of developing the best electroporation proto- 356 Harkin and Hay col to load cells without adversely affecting the cytoskeleton. Since Pipes/EGTA buffers support the cytoskeleton when cells are permeabilized with detergent prior to immunostaining, we decided to develop an electroporation buffer based on these ingredients. The stability of microtubules immediately following electroporation in CSB initially suggested that this might be a suitable buffer for our experiments, but unfortunately cells loaded in this medium displayed disrupted microtubules when returned to culture medium. The most likely explanation for this result is that CSB does not enable electroporated cell membranes to reseal effectively before their return to culture medium (DFH). Therefore, while CSB proved to be an unsuitable electroporation buffer for studies of directed cell migration, the stability of microtubules immediately following electroporation in this buffer at least indicated that microtubules are not directly disrupted by the electric currents used. Following studies with CSB, we considered making several modifications to this buffer including substitution of EGTA with low calcium (1 pM) and increasing pH. In the end, we designed a buffer (ICEB) which mimics as closely as possible the ionic composition of intracellular fluid (i.e., high potassium, low sodium, and low calcium). The problem of how to add high quantities of potassium ions without accompanying anions was solved by neutralizing the required amount of potassium in the form of KOH with Pipes free acid. While a pH of 6.8 is optimal for microtubule assembly [Borisy et al., 19761, preliminary results suggested that a pH of 7.0 improved the ability of cell membranes to reseal following electroporation. Myo-inositol was used [Neil and Zimmermann, 19931 to adjust osmolarity without changing the ionic strength or pH. ATP was omitted from the electroporation medium because it is known to permeabilize plasma membranes [McNeil, 19891 to low molecular weight molecules ( I1 kDa) and may therefore increase the risk of exposing microtubules to extracellularCa2 ions following the return of cells to culture medium. It should be noted that while only 1-10 pM of CaCl, were added to ICEB, these concentrations are less than, or equal to, that for Ca2+ ions already present in standard preparations of distilled deionized water. In the end, ICEB is a buffer that assists electroporation of cell membranes, supports stability of the cytoskeleton, and enables fibroblasts to migrate normally when returned to culture medium. While other types of intracellular buffers have been used previously for cell electroporation [Knight and Scrutton, 19861, we believe that ICEB is the first buffer to be designed with attention to the requirements for studies of fibroblast motility. We believe that the confocal and electroporation methods developed in the present study should promote future studies on the behavior of the F-actin cortex during the directed migration of fibro+ blasts within ECM. Furthermore, the large size range of macromolecules loaded in the present study, suggests that the technique may be suitable for loading cells in ECM with a variety of other molecules of current interest including small GTP-binding proteins [Hall, 19941, and cell lineage markers [Artinger et al., 19951. ACKNOWLEDGMENTS We thank Eric Berthiaume, Dr. Joel Swanson, and Dr. Stephen Baer for their advice on microscopy, Megan Maultsby for technical assistance, and Dr. Charles Vanderburg for his helpful comments throughout the preparation of this article. The E7 hybridoma supernatant developed by Dr. M. Klymkowsky of the University of Colorado, and the H5 hybridoma supernatant developed by Dr. J. Sanes of Washington University School of Medicine were both obtained from the Developmental Studies Hybridoma Bank maintained by the Department of Pharmacology and Molecular Sciences, The Johns Hopkins University School of Medicine, Baltimore, MD 21205, and the Department of Biology, University of Iowa, Iowa City, IA 52242, under contract NOl-HD-23144 from the NICHD. This work was supported by U.S. Public Health Service grants R01-EY09721 and ROl-DE11142. REFERENCES Abercrombie, M., Heaysman, J.E.M., and Pegrum, S.M. (1970a): The locomotion of fibroblasts in culture, I: Movements of the leading edge. Exp. Cell Res. 59:393-398. Abercrombie, M., Heaysman, J.E.M., and Pegrum, S.M. (1970b): The locomotion of fibroblasts in culture, 111: Movements of particles on the dorsal surface of the leading lamella. Exp. Cell Res. 62:389-398. Abercrombie, M., Heaysman, J.E.M., and Pegrum, S.M. ( 1972): Locomotion of fibroblasts in culture, V: Surface marking with concanavalin A. Exp. Cell Res. 73536-539. Artinger, K.B., Fraser, S., and Bronner-Fraser, M. (1995): Dorsal and ventral cell types can arise from common neural tube progenitors. Dev. Biol. 172591-601. Bard, J.B.L., and Hay, E.D. (1975): The behavior of fibroblasts from the developing avian cornea. .I. Cell Biol. 67:400-418. Baum, C., Forster, P., Hegewisch-Becker, S., and Harbers, K. (1994): An optimized electroporation protocol applicable to a wide range of cell lines. Biotechniques 17:1058-1062. Bershadsky, A.D., and Futerman, A.H. (1994): Disruption of the Golgi apparatus by brefeldin A blocks cell polarization and inhibits directed cell migration. Roc. Natl. Acad. Sci. U.S.A. 9115686-5689. Bershadsky, A.D., Vaisberg, E.A., and Vasiliev, J.M. (1991): Pseudopodal activity at the active edge of migrating fibroblast is decreased after drug-induced microtubule depolymerization. Cell Motil. Cytoskeleton 19:152-158. Bilozur, M.E., and Hay, E.D. (1989): Cell migration into the neural tube lumen provides evidence for the “fixed cortex” theory of cell motility. Cell Motil. Cytoskeleton 14:469-484. Borisy, G.G., Johnson, K.A., and Marcum, J.M. (1976): Self-assembly and site-initiated assembly of microtubules. In Goldman, Effects of Electroporation R., Pollard, T., and Rosenbaums, J. (eds.): “Cell Motility, Book C: Microtubules and Related Proteins.” Cold Spring Harbor, N Y Cold Spring Harbor Laboratory, pp. 1093-1 121. Daniels, K.J., and Hay, E.D. (1990): Distribution of p integrins, matrix, actin, and actin-associated proteins during migration of elongated fibroblasts in 3D collagen gels. J. Cell Biol. 111: 140a. Elsdale, T., and Bard, J. (1972): Collagen substrata for studies on cell behavior. J. Cell Biol. 54:626-637. Glogauer, M., and McCulloch, C.A.G. (1992): Introduction of large molecules into viable fibroblasts by electroporation: Optimization of loading and identification of labeled cellular compartments. Exp. Cell Res. 200:227-234. Hall, A. (1994): Small GTP-binding proteins and the regulation of the actin cytoskeleton. Annu. Rev. Cell Biol. 1031-54. Hashimoto, K., Tatsumi, N., and Okuda, K. (1989): Introduction of phalloidin labeled with fluorescein isothiocyanate into living polymorphonuclear leukocytes by electroporation. J. Biochem. Biophys. Methods 19:143-154. Hay, E.D. (1985): Interaction of migrating embryonic cells with extracellular matrix. Exp. Biol. Med. 10:174-193. Hay, E.D. (1989): Theory for epithelial-mesenchymal transformation based on the “fixed cortex” cell motility model. Cell Motil. Cytoskeleton 14:455-457. Hay, E.D. (1991): “Cell Biology of Extracellular Matrix.” New York: Plenum Press. Hay, E.D., and Revel, J.P. (1969): “Fine Structure of the Developing Avian Cornea.” New York S. Karger. Heath, J.P. (1983): Behaviour and structure of the leading lamella in moving fibroblasts, I: Occurrence and centripetal movement of arc-shaped microfilament bundles beneath the dorsal cell surface. J. Cell Sci. 60:331-354. Jones, G., Wigley, C., and Warn, R. (eds.) (1993): “Cell Behaviour: Adhesion and Motility. Symposia of the Society for Experimental Biology.” Cambridge, UK: The Company of Biologists Limited. Knight, D.E., and Scrutton, M.C. (1986): Gaining access to the cytosol: The technique and some applications of electropermeabilization. Biochem. J. 234:497-506. Kupfer, A., Kronebusch, P.J., Rose, J.K., and Singer, J. (1987): A critical role for the polarization of membrane recycling in cell motility. Cell Motil. Cytoskeleton 8: 182-189. Lawrence, J.B., and Singer, R. (1986): Intracellular localization of messenger RNAs for cytoskeletal proteins. Cell 45:407-415. Liang, H., Purucker, W. J., Stenger, D.A., Kubiniec, R.T., and Hui, S.W. (1988): Uptake of fluorescence-labeled dextrans by 1OT 1/2 fibroblasts following permeation by rectangular and exponential-decay electric field pulses. Biotechniques 6:550-558. Liao, G., Nagasaki, T., and Gundersen, G.G. (1995): Low concentrations of nocodazole interfere with fibroblast locomotion without significantly affecting microtubule level: Implications for the role of dynamic microtubules in cell locomotion. J. Cell Sci. 108:3473-3483. Lukas, J., Bartek, J., and Straws, M. (1994): Efficient transfer of antibodies into mammalian cells by electroporation. J. Immunol. Methods 170:255-259. Marotta, M., and Martino, G. (1985): Sensitive spectrophotometric method for the quantitative estimation of collagen. Anal. Biochem. 150:86-90. McNeil, P.L. (1989): Incorporation of macromolecules into living cells. Methods Cell Biol. 29:153-173. Neil, G.A., and Zimmermann, U. (1993): Electroinjection. Methods Enzymol. 221:339-361. 357 Oakley, C., and Brunette, D.M. (1995): Topographic compensation: Guidance and directed locomotion of fibroblasts on grooved micromachined substrata in the absence of microtubules. Cell Motil. Cytoskeleton 31:45-58. Palecek, S .P., Schmidt, C.E., Lauffenburger, D. A,, and Horwitz, A.F. (1996): Integrin dynamics on the tail region of migrating fibroblasts. J. Cell Sci. 109:941-952. Pryzwansky, K.B. (1982): Applications of double-label immunofluorescence. In Bullock, G.R., and Petruszs, P. (eds.): “Techniques in Immunocytochemistry.” London: Academic Press, Inc., pp. 77-90. Regen, C.M., and Horwitz, A.F. (1992): Dynamics of p l integrinmediated adhesive contacts in motile fibroblasts. J. Cell Biol. 119:1347-1359. Tatsuka, M., Orita, S., Yagi, T., and Kakunaga, T. (1988): An improved method of electroporation for introducing biologically active foreign genes into cultured mammalian cells. Exp. Cell Res. 178:154-162. Theriot, J.A., and Mitchison, T.J. (1992): Comparison of actin and cell surface dynamics in motile fibroblasts. J. Cell Biol. 119: 367-377. Tomasek, J.J., and Hay, E.D. (1984): Analysis of the role of microfilaments and microtubules in acquisition of bipolarity and elongation of fibroblasts in hydrated collagen gels. J. Cell Biol. 99536-549. Tomasek, J.J., Hay, E.D., and Fujiwara, K. (1982): Collagen modulates the shape and cytoskeleton of embryonic corneal and fibroma fibroblasts: Distribution of actin, alpha-actinin, and myosin. Dev. Biol. 92:107-122. Trinkaus, J.P. (1973): Surface activity and locomotion of Fundulus deep cells during blastula and gastrula stages. Exp. Cell Res. 30:68-103. Tsong, T.Y. (1991): Electroporation of cell membranes. Biophys. J. 60:297-306. Vasiliev, J.M., Gelfand, I.M., Domnina, L.V., Ivanova, O.Y., Komm, S.G., and Olshevskaja, L.V. (1970): Effect of colcemid on the locomotory behaviour of fibroblasts. J. Embryol. EXP. Morph. 24:625-640. Wang, Y .-L. (1985): Exchange of actin subunits at the leading edge of living fibroblasts: Possible role of treadmilling. J. Cell Biol. 101597-602. Weaver, J.C. (1994): Molecular basis for cell membrane electroporation. Ann. N.Y. Acad. Sci. 720:141-152. Weaver, J.C. (1995): Electroporation theory: Concepts and mechanisms. Methods Mol. Biol. 47:l-26. Wehland, J., Osborn, M., and Weber, K. (1977): Phalloidin-induced actin polymerization in the cytoplasm of cultured cells interferes with cell locomotion and growth. Proc. Natl. Acad. Sci. U.S.A. 745613-5617. Wehland, J., Osborn, M., and Weber, K. (1980): Phalloidin associates with microfilaments after microinjection into tissue culture cells. Eur. J. Cell Biol. 21:188-194. Wilson, A.K., Gorgas, G., Claypool, W.D., and de Lanerolle, P. (1991a): An increase or a decrease in myosin I1 phosphorylation inhibits macrophage motility. J. Cell Biol. 114:277-283. Wilson, A.K., Horwitz, J., and de Lanerolle, P. (1991b): Evaluation of the electroinjection method for introducing proteins into living cells. Am. J. Physiol. 260:C355<363. Wolf, H., Rols, M.P., Boldt, E., Neumann, E., and Teissie, J. (1994): Control by pulse parameters of electric field-mediated gene transfer in mammalian cells. Biophys. J. 66524-531.