The Electrofusion of Cells

Department o Physics, Oklahoma State University, Stillwater, Oklahoma 74078 f

HERBERT A. POHL, K. POLLOCK, AND H. RIVERA

Abstract
Many laboratories studying genetic engineering are using cell fusion to aid their research. The new and highly efficient method of using electrical fields to induce cell-to-cell fusion, or cell-toparticle fusion, or to induce entry of DNA or other compounds into cells opens up rapid ways to accomplish these ends.

Introduction
Membrane fusion, now an art controllable by man, has great potential in membrane, genetic and cell research. Somatic hybridization and genetic engineering have the means to modify plant and animal cells, which in turn offer ways to improve crops, treat disease, and improve medical science. In this review we describe a new electrical procedure of outstanding efficiency for obtaining cell-to-cell, cell-to-vesicle, vesicle-to-vesicle fusion, and of introducing chemical agents, DNA, etc., into membrane bound particles such as cells, protoplasts, or vesicles. Membrane fusion is a natural phenomenon, widely observable in nature. It is, for example, involved in fertilization, endocytosis, exocytosis, in muscle fiber formation, in secretion, in pinocytosis, in the formation of secondary lysosomes, and in the conjugation of bacteria and protozoans. Cell-to-cell fusion in v i m , opens vistas in membrane research, in genetic mapping, and in the formation of cells with new properties. The much desired production of new plants resistant to drought, high salinity, insect pests, insecticides, or fungi offers new challenges to the bioscientists. In addition, the in vitro modifications open the way to create crops able to directly assimilate nitrogen or to produce in higher yield. On the medical side, the hybridization of a permanent cell line (oncogenic cell) with antibody-producing or hormone-producing cells offers extraordinarily efficient means to produce monoclonal antibodies against selected compounds, toxins, bacteria, viruses, or even cancer cells. The new technique of electrofusion opens time-saving short cuts to these goals. It is an important new tool of research and production. In the past decade, a variety of biochemical methods have been developed to induce cell-to-cell fusion. These include the use of virus particles or of various chemical agents such as dimethyl sulfoxide or polyethylene glycol (PEG). Such fusion procedures can usually only be obtained by the application of membrane disrupting agents or by procedures which are so unphysiological as to minimize viability. The research in such chemically and viral-induced fusion has been well
International Journal of Quantum Chemistry: Quantum Biology Symposium 1 I , 327-345 (1984) 0 1984 by John Wiley & Sons, Inc. CCC 0360-8832/84/010327-19$04.00

328

POHL, POLLOCK, AND RIVERA

reviewed by a number of writers [ 1-41, The new and highly efficient electrofusion technique supplements these methods. It is evident from the literature that the nature of all these fusion processes are as yet incompletely understood. Much remains to be learned. Practical progress in fusion research depends upon the realization that the optimum fusion conditions need at present to be empirically determined for each cell system, for they vary considerably from species to species. Moreover, in the viral or chemical induction of fusion, the exact time of fusion onset cannot be determined, or followed directly under the microscope. The latter procedure is particularly desirable in the case of interspecies fusion. The yields of fusion in such cases is generally low and the number of cells to be fused cannot be preselected. The electrofusion process, on the other hand, can be done so as to greatly avoid these disadvantages. Here, the cells are subjected to a sharp electrical pulse which temporarily opens up pores between contacted cells, following which the membranes unite and cause fusion of the cells. In brief, the new electrofusion technique offers three possible ways to operate on cells. One can (a) fuse like cells to form larger entities such as cell pairs, triplets, etc., on up to giant many-celled objects having a common outer membrane; (b) fuse unlike cells such as desired in creating hybridomas; or (c) help drive external objects or chemical agents such as DNA into cells (or vesicles). It is a powerful new tool for the life scientists. The electrofusion technique is best done with the aid of gentle ac dielectrophoresis (a nonuniform electric field effect on neutral matter, called DEP for short) [ 5 ] . This forces the cells to stack nicely upon each other, so as to bring about tight membrane contact. Following this, one or more brief dc pulses is applied to open up and fuse the cells. The procedure is quite gentle, is synchronous among the cells, and can lead to high yields of fused particles. It is perhaps most important to appreciate that electrofusion does not necessarily perturb the whole cellular surface, but acts principally at the small area (ca. 100 in diameter) of contact, and then only for a brief instant (ca. 40 ks). The size of the affected area can be increased if desired. The number of cells electrofused at one time can be controlled by the field arrangements, by the suspension density, by sequential loading of the fusion chamber by different types of cells, etc. For what is very desirable, the electrofusion technique can be made to maximize the selectivity of the fusion of AB fusions, where A and B represent different cell types. It should be remarked that in most, but not all cases of cell-to-cell electrofusion observed to date, it has been desirable to maximize the fusion yield by decorticating the cells to form protoplasts, as by use of enzymes such as dispase, neuraminidase, or pronase for avian or mammalian cells, or of enzymes such as cellulase or macerase for plant cells. Such enzymatic treatments do indeed alter the cellular surface temporarily during the brief treatment. In the main, however, the viability of electrofused cells appears to be excellent. Electrofusion has been successful in all types of cells tested so far, including bacterial and plant protoplasts, yeast cells, sea urchin ova, avian and mammalian

ELECTROFUSION OF CELLS

329

cells [6-111. Giant cells, i.e., those made by electrofusing many cells to form a giant entity, have been prepared from yeast mammalian red blood cells, Indian muntjac fibroblasts, and cow lung macrophage cells [ 10,l I ] . The resulting giant cells respond favorably to vital stain (methylene blue) tests for vitality. It is reported, however, that giant cells containing more than about four original cells do not survive attempts at division upon further culture [6]. Electrofusion of cells in vitro was first observed by Pohl and Buckner in 1972 during studies [12] on the dielectrophoresis (DEP) of canine erythrocytes and thrombocytes. Pohl also noted that human erythrocytes suspended in deionized 5% sucrose solution, and then subjected to DEP at 2.5 MHz coalesced to form a single cell following the application of a brief pulse of 200 V peak-to-peak. Similar results on electro-coalescence were obtained with canine thrombocytes suspended in 0.25M sucrose on using 1 MHz in pulses. The first published morphological evidence for electrically mediated fusion of the protoplasts of different plants was that of Senda et al. [ 131. Neumann, Gerisch, and Opatz [7] first described electrofusion using high voltage pulses, as applied to the eucaryotic microorganisms, Dictyosteliurn discoideurn. In Europe, electrofusion done with the aid 13 DEP was, one author (H. Pohl) was told, first observed by Maja Mischel [ 141, an observation that was quickly recognized for its value by Lamprecht and Zimmermann. This was followed by a spate of research on electrofusion. The publications of Neumann et al. [7], Berg et al. [6], and Zimmermann et al. [8,10,14] stimulated much interest in electrofusion and its implications for the interactions of electromagnetic fields with living systems. The electrofusion procedure is a relatively gentle one, provides a sharp synchrony of the onset of fusion, and can lead to a high yield (20-80%) of fused cells. The presently preferred technique is one combining DEP [5,15-181 to obtain placement of cells into close contact by the formation of pearlchains or stacks, and of controlled fusion following the application of dc pulses [7-111.

Dielectrophoresis
Dielectrophoresis, first theoretically recognized, quantitatively analyzed, and named by Pohl [15], is the translational motion of neutral matter induced by the action of a nonuniform electric field (see Fig. 1). It can be considered to arise in two steps. First the electric field causes the material to be polarized, creating an electric dipole. The two equal charge distributions of this material dipole are, however, in a nonuniform field. Since those charges in the stronger field region experience a stronger local force [ F = q X E (local)] than their opposite numbers in the weaker field, a net force arises which tends to propel the matter into the stronger field. It is called dielectrophoresis (DEP) since it depends upon the dielectric properties, the polarizability. The action of an electric field on charged particles, i.e., those carrying free and excess charge, is by convention called electrophoresis. The phenomenon of DEP is an ancient one. It was doubtless through it that Thales of Miletus in about 600 B.C. is reported to have observed the attraction

330

POHL, POLLOCK, AND RIVERA

Figure 1. Comparison of the behaviors of neutral and charged bodies in an alternating nonuniform electric field. (a) The positively charged body moves toward the negative electrode. The neutral body is polarized, then is attracted towards the region of strongest field. This is because the two charge regions on the neutral body are equal in their amount of charge, but one of the charge regions is in a stronger field. The force on a charge is proportional to the local field. (b) Here the direction of the field is reversed. The positively charged object again moves toward the negative electrode. The neutral body is again polarized, but does not reverse its direction although the field is reversed, for it still seeks the region of highest field intensity.

of charged (rubbed) amber for small (neutral, we can now realize) lint particles, and discovered electricity [ 5 ] . The distinction between dielectrophoresis and electrophoresis is important. When an ac field is applied, for instance, DEP tries at each field reversal to drive the particle into the region of higher field strength, for the induced dipole reverses each time. The particle then tends to go into the region of stronger field continually. On the other hand, during electrophoresis, the charged particle is pulsed towards the electrode of opposite charge sign. At high frequencies it merely shudders and makes little net motion towards either electrode. DEP, then, can act to separate neutral from charged particles. Neutral particles such as cells, can be made to collect into the region of strongest field while charged ones are more or less ignored. A caveat is necessary here. Since one uses DEP on suspension of cells or particles in a fluid medium, there is necessarily a competition of DEP forces acting upon the material of the medium and that of the suspended particles. If the polarizability of the particles exceeds that of the medium at the field frequency chosen, the particles will tend to be collected in the region of strongest field. This is positive DEP. If conversely the medium is the more polarizable, then the particles will be pushed into the region of weaker field. This is negative DEP. The procedure is quite analogous to the familiar floating or sinking as in gravitational fields. If the particle is asymmetrical, or inhomogeneously made, the electrical polarization can produce a torque so as to realign the particle in the field lines and

ELECTROFUSION OF CELLS

33 I

minimize its energy. Thus dielectrophoretic effects can produce both a purely translational and an orientational result. The term electrostriction is nowadays reserved to describe only the distortional effects of fields, such as would arise in the piezoelectric effect, the Kerr effect, or bulk distortions of matter. A further fundamental difference between electrophoresis and DEP requires mention. In mixtures such as suspensions of cells in aqueous media, successful DEP requires a substantial difference between the permittivities (total polarizabilities) of the medium and particles. The dielectric properties of living cells and their parts change with frequency. There are four or more distinct regions of this relaxation response, reflecting, as the frequency increases, the various components of the cell (see Fig. 2 ) . At the frequency range of about 100-10,000 Hz, the very loosely bound counterions of the outer layer of the ionic double layers associated with membranes and cell walls strongly respond. At about 10100 kHz the more tightly bound counterions of the ionic double layers respond. If associated with the outer layers of spherical particles, for example, these response frequencies vary as the inverse of the square of the particle radius. At about 100-10 MHz, the predominent response is that due to the bulk-bulk interfacial (Maxwell-Wagner-Sillars) polarization [ 5 ] . In this range, the DEP response of all cells is a common occurrence, reflecting the common presence of membranal interfaces with the medium. It is probably the most useful frequency range for producing the cell-to-cell stacking and assuring membrane-tomembrane contact. In the frequency range above about 3 MHz the polarization responses of dissolved proteins, DNA, RNA, and other polar molecules can be

\
0
I

log,,

FREQUENCY

Hz

Figure 2. A simulated DEP force, or effective dielectric constant spectrum such as observed in living cells. The regions of frequency below which the various polarization mechanisms are active are: (A) responses of loose counterions of the ionic double layers; (B) responses of tightly bound counterions in ionic double layers; (C) Maxwell-Wagner, bulk-bulk interfacial response region; (D) molecular dipolar responses, of large macromolecules; (E) molecular dipolar responses, of small molecules.

332

POHL, POLLOCK, AND RIVERA

observed. For further details on these points, the reader is referred to other sources [ 18,201.

Mutual Dielectrophoresis
Up to this point the possible interactions of particles in an electric field have not been discussed. The cells normally have a different polarizability than the medium, and therefore perturb the field. A second cell will therefore be subject to that field nonuniformity and be impelled toward the region of highest field density, at the end of the first cell. The net result is alinement of the cells along the field lines in a little pearl chain (see Fig. 3). The same situation occurs if the polarizability of the cells are less than that of the medium for then the medium seeks minimum energy and forms regions lined up along the field, driving the cells into pearl chains by default. There is one exception. If two freely suspended particles of approximately equal size are present, one with a higher, one with a lower permittivity than the fluid medium, then the two particles tend to form a short pearl chain sitting at right angles to the field lines. The formation of the stacks of cells by DEP is quite useful for the process of electrofusion, for it puts the cells in contact as desired. The attractive force of mutual DEP now can help overcome the natural repulsion between the adjacent membrane surfaces due to their hydration and to their ionic double layers. The mutual DEP and cell-stacking is best done with the use of high frequency ac fields. As mentioned above there is an almost universal positive DEP response of cells in the frequency range of about 200 to 900 kHz due to the bulk-bulk interfacial (Maxwell-Wagner) polarization response, so this is the frequency usually used to evoke the cell-to-cell contact prior to the application of the higher voltage dc pulse to produce the fusion. DEP and its concommitant cell stacking normally has to be done in media which are nearly non-conducting (specific conductivity less than about 0.0001 mho/cm or S/cm) so as to avoid undue heating, turbulences, and other deleterious consequences for the cells and the electrofusion process [ 5 ] . The use of such media with low ionic content poses little or no problem for those cells having walls such as bacteria, yeasts, and algae. But then the electrofusion, which

Figure 3. Diagram of pearl chaining by cells or other particles produced by the action of DEP is due to an ac field arising between parallel wire electrodes.

ELECTROFUSION OF CELLS

333

requires intimate membranal contact is also hindered. The matter is eased by the use of the cells in protoplast form with the medium now made properly isoosmotic by the use of nonionizing solutes such as sucrose, glucose, mannose, mannitol, or sorbitol, or of the zwitterionic materials such as histidine, or glycine. Experiments with mammalian, avian, yeast protoplasts, or with plant protoplasts shows them to experience but a few deleterious side effects on their viability upon the brief exposure (ca. 2-30 min) to these conditions during the entire procedure required for electrofusion.

Membrane Puncture and Function

The application of a strong electrical field across a bilayer membrane can cause it to be punctured. This poration as it is sometimes called can be temporary (reversible), or can be so severe as to induce irreversible breakdown and disruption of the whole membrane-supported assembly. In the process of electrofusion, we shall be most interested in the gentler, reversible type of membrane puncture. It provides the condition from which fusion between closely positioned similar structures can follow. The more damaging irreversible cell rupture is also of interest, for although it can cause trouble in the normal electrofusion operations, it can also be used to selectively destroy cells. The type of membranal puncture one obtains is governed largely by the amplitude of the field, and its duration. It is also strongly affected by the chemical composition of the membrane region. The presence, for instance, of pronase is reported to improve the durability [8] of the membranes to overvoltage during the electrofusion process. As representative of the effects of field amplitude and duration on the electrofusion of erythrocytes, a look at Figure 4 shows that for short dc pulses applied to the cells formed into stacks or pearl chains by a gentle ac field, there is little observable effect by pulses shorter than about 25 ps in the low fields used here, but there is more and more effect as the field duration of the pulse lengthens. There is also observable a range of field amplitudes in which the desired fusion takes place, and a range, which if exceeded, causes almost total destruction (lysis) of the cells. It is this type of information which needs to be determined in each case as one studies the electrofusion of differing types of cells. There are some guiding principles, however. For one, the potential required to induce the reversible puncture of most cellular membranes is known to be about 0.85 V for pulse lengths of about 20-100 ps. It was also shown by Fricke in 1924 [18,19] that the potential across the cellular membranes due to the application of a static external field acting on a spherical cell modelled as a conductive region inside a nonconductive membrane and lying in a rather conductive medium should follow the relation:

V,

1.5Er,

where V , is the potential across the membrane region, E is the external field which would pertain if the cell were absent, and r is the radius of the spherical cell. To account for angular variations of the potential around the cell, one need

334
Volts

POHL, POLLOCK, AND RIVERA

KV cm
-6

-150
LY

0
2
A

-5
\

+ 4

i -loo

\.

-4

\ \

* '*-,
\

Ly

A u7

- -1-

2-

- - - - -5 3
21

\
\

n -50 .

*\

CELL E L E C T R O - F U S I O N

-- 60

.----80

20

40

PULSE

DURATION

(PSec)

Figure 4. Comparison of the fields able to evoke fusion (lower curve) and able to evoke lysis (upper curve) of rabbit erythrocytes. The dc pulse amplitudes versus duration are shown as the cells were gently held in pearl chains by a small ac field. Along the lower curve, fusion of the cells occurs in about 5-15% of the cells. The electric field is that due to parallel R wires of 250-pm diameter separated by 120 pm. The cells are pearl-chained away from the electrodes, while on the floor of the chamber, and are in 0.3M sucrose containing added CO, at pH 4.8.

only add a multiplying factor of the cosine of the angle 0, away from the axis provided by the field line going into the middle of the cell, to the formula, viz. (see Figs. 5 and 6). V,,, = 1.5Er cos0. (2)
To take into account the time variations requires more complicated analysis. In the Debye format, [201 one simply inserts a factor of (1 + j w T ) in the denominator and then obtains for the real part:

V , = (1.5Er cos0)/(1

+ w2T2),

(3)

Figure 5 . Diagram of cell in a field showing the coordinates of interest for pulsing fields.

ELECTROFUSION OF CELLS
Pt ELECTRODE

335

'POL A R" FUSION

Figure 6 . Diagram of polar and lateral fusion processes among cells or vesicles situated between wire electrodes and subject to strong, brief electrical pulses.

where w = 2nf, the angular frequency, f is the cyclic frequency, and T is the relaxation time of the system. The relaxation time T for frequencies in the range of the Maxwell-Wagner mechanism and where DEP is most often used for fusion, can be evaluated by the relation [21,22]

( P , + P,/2)rC,,

(4)

where P, and P, are the specific resistivities of the internal and external liquid media, and C , is the specific capacitance (usually about 1 pF/cm2). For our purposes, the product wT is usually rather less than unity, so that the simpler formula (2) can be used with reasonable confidence. It can be used immediately as a guide for preliminary electrofusion experiments by making the rough assumption that the field E is that provided by parallel plates (rather than the more realistic and practically used set of parallel wire electrodes, say). The E V l d . where V is the applied potential on the electrodes, and d is their separation, and we have
2

V , = 1.5Er
or

l.SV(r/d),

(5)

V -- V,d/( 1.5r),

(6)

as the estimate for the voltage to apply across the electrodes to achieve fusion. Using the value V , = 0.85 V, we have as our approximate suggested voltage V:

V = 0.85(d/r)/1.5

0.6(d/r).

(7)

For example, should we wish to try a fusion on some cells with a diameter

336

POHL, POLLOCK, AND RIVERA

of, say, 10 pm in our fusion chamber having a spacing of d = 200 pm, the simple formula suggests a voltage of some 12 V. The actual value in practice will vary from this, of course, for the real conditions will not be so ideal, but the suggested value is helpful as an initial guide. One of the factors determining the actual field at the cells to be electrofused is the electrode shape and spacing. A helpful guide in this is the calculations of Lafon and Pohl [23] for the field intensities about electrodes of the sphere-sphere and wire-wire types. For the cylinder-cylinder pair, the field, E9Z) along the plane connecting the centers is shown in Figure 7. Here the field with 1 V applied across the electrode pair is shown for a variety of electrode dimensions. The calculation is based upon the relation

E(Z)

V 1 2 In ( R , I R )Z - R 2 / R ,

+-R , 1 Z -

(8)

where V is the applied voltage, R , is the radius of the electrode wire, Z is the distance between electrode centers,
2
=

R,

+ R2/RI,

(9)

E. v l m

Figure 7. The electric field along the line between centers of a pair of cylindrical and parallel electrodes between which a potential of 1 V is applied. Here, R is the electrode radius, and the field (in V/m) is shown as a function of the distance between the electrodes closest approach. Also indicated are the critical fields, which if just the 1 V is applied across the electrodes, might be expected to evoke appreciable poration of the cell of the diameter indicated.

ELECTROFUSION OF CELLS

331

is the center-to-center cylinder separation, and R , is the radius of the cylinders. To use the chart, simply multiply the values by the voltage to be used, if it is not to be just 1 V. To aid the experimenter wishing to use electric fields to handle living cells, we have also calculated the critical field required by relation (3, assuming and the breakdown voltage V,, to be 1 .O V, and have plotted these values for several frequently encountered cell sizes (diameters) as shown in Figure 7. If, for example, it is desired to merely handle and examine cells in an electric field without subjecting them to critical fields liable to cause disruptive puncture, then the appropriate conditions can be read off from the graph.

Aligning Cells
The genius of electrofusion consists of bringing cells into intimate contact, then applying a short sharp pulse of field to evoke small breaks in the contacted areas following which fusion across the temporary ruptures occurs. Membrane contact between the properly prepared cellular surfaces is best achieved by DEP and mutual surface interactions in a mild ac field. The field nonuniformities DEP force upon an object depends upon how all the charges in the object and in its surrounding medium respond to the applied nonuniform field. This means that not just the normal polarization (dielectric constant), but also the conduction responses are involved. As a consequence, one must look after the conductivities to achieve maximal success. In a nutshell, successful DEP of cells and other biological particles requires the use of aqueous media of low conductivity. Normally a specific conductivity of about 3 x mhoicm or less is required (i.e., a specific resistivity of about 30,000 ohm cm or higher). Best results are obtained with specific conductivities below about 5 x mho/cm (i.e., specific resistivities higher than 200,000 ohm cm). A feel for the relative importance of the polarization and conductivity factors can be had from looking at the equation for the DEP force upon a simple sphere suspended in a fluid while in a nonuniform electric field. One relation derived some time ago [ 5 ] is

( V / 2 )Re{e&T(K2 - K I ) / ( K 2+ 2 K , ) / V (&I2},

(10)

where V is the volume of the sphere, en is the permittivity of free space, (8.854 X l o - F / m or C/V m), K , and K2 are the complex relative dielectric constants of the medium and the sphere respectively, (see below). Re refers to taking the real part of what follows, and V (En) is the gradient of the square of the electric field which would be at the center of the sphere were it absent. A more exact expression due to Prof. Dr. Freidrich A . Sauer of the Max-Planck Institute for Biophysics in Frankfurt, Germany, and derived not on the less precise basis of energy conservation as was the above expression, but on the conservation of momentum so as to better account for frictional losses, etc., is

(3V/16)[6*/(3 b*) +

+ bl(3 + b)](KT + KI)(EO),

(11)

338

POHL, POLLOCK, AND RIVERA

where b = (K& - Kl,tot)/Kl,,,,,, = K + j s / ( e , W ) , the complex relative K,,, dielectric constant, b* is the complex conjugate of b, and s is the specific conductivity. The two expressions agree fairly well for situations in which the conductivity is not too high. As can be seen, however, both the polarization (dielectric constant) and the conduction factors are important. Said another way, . the presence of minute ionic conduction (e.g., ca. 0.0005M salt) in water drastically raises its effective dielectric constant when considering DEP forces. As a consequence, the effective dielectric constant of the suspended cells, and their DEP attractions to each other or to the desired electrode positions can be overridden by the high effects of effective dielectric constant of the conductive medium, resulting in poor dielectrophoretic control of the cells. This emphasizes the need for maintaining low conductivity in the aqueous medium. Another good reason for maintaining low conduction in the support medium lies in the need to minimize Joule heating with its consequent thermal upsets, stirring effects, and outright danger to the cells. Dielectrophoresis and the pearl-chaining of cells, then, usually must be done in nearly nonconductive media, but where the osmotic problems are met with by using nonconductive solutes such as glucose, mannose, sucrose, histidine, and the like. Mechanism of Electrofusion A brief comment on the way in which electrofusion is thought to occur is in order. From the electrical point of view, the thin membrane is a good insulator situated between two conductors, i.e., the aqueous inner and outer media. A field imposed across the membrane-water interfaces will therefore produce its major potential drop (field strength and electric stress) across the membrane itself. When the cells are brought into contact, as by DEP, the ensuing strong (dc) electric pulse produces small punctures in the membrane regions, especially where the field is most intense, at the poles of the cells along the lines of the field, i.e., toward the electrode surfaces, or toward the usual contact areas of the cells in the pearl-chains. In living cells there already exists a strong field across the outer membrane (plasmalemma), amounting to about 100,000 V/cm. As the external field pulse is applied in a particular direction, this adds to the field across one of the two membranes in contact, and subtracts from the other. The membrane with the greater field is thought to then break down first. The pores and defects in its surface then act to focus the remaining field on the second membrane causing it to develop punctures in the locale. The time required for this early event is judged to be in the order of a small fraction of microsecond. The ensuing time of the pulse duration of some 10-100 k s , say, is then involved in holding open the communal intercellular puncture while the new edges knit or seal to each other. Following this the field is turned off and the newly sealed edges develop under surface forces to form part of a single unitized membrane enveloping the original cells. An example of the sequence of events is shown in Plates 1-6. Here murine

ELECTROFUSION OF CELLS

339

Plate 1

Plate 2

340

POHL, POLLOCK, AND RIVERA

Plate 3

Plate 4

ELECTROFUSION OF CELLS

341

Plate 5

Plate 6

342

POHL, POLLOCK, AND RIVERA

myeloma cells are first seen to be formed into pearl-chains between the two parallel Pt electrodes plates by the gentle ac chaining field. They are then subjected to the sharp dc pulse which evokes fusion. Following the dc pulse the ac voltage was reduced in value so as to permit the cells more easily to fuse into each other. The entire sequence shown covers a time interval of about three minutes. The living cells contain positive and negative ions inside, able to move when an electric field is applied to them. The mobility of hydrogen ions is about an order of magnitude higher than the mobility of potassium, sodium, magnesium or calcium ions. In response to the externally established electric field those internal ions move against the membrane. The double membrane layer at the contact zone of the two cells becomes in effect a capacitor charged at the potential which originally was across the distance separating the ionic charges in the two identical cells. By raising the externally applied electric field E one can reach the dielectric breakdown threshold of this double membrane capacitor and thus initiate the electrofusion of cells. Since ions inside the cells have a finite mobility the charging of this membrane capacitor is concentration and time dependent during application of the rectangular electric field pulse. That is why in the data of Figure 4 no fusion is observed even with an externally applied field of 4 kV/cm until the duration of the pulse is at least 25 p s . Up to that point ions had not moved in sufficient number to charge the capacitor and provide a potential across it (Q = CV) reaching the breakdown threshold. As the pulse duration is increased the lower curve of Figure 4 shows that fusion threshold is achieved with lower and lower externally applied electric fields, because more ions had time to charge the double membrane capacitor to the breakdown potential. Beyond 7 0 - i ~ ~ duration the external electric field pulse producing the fusion threshold is about constant, indicating that beyond that point the electric field inside the cells is almost zero due to the now completed drift ions under the externally applied electric field. At the other end of the spectrum one might have cells that have lost so much of their internal ions that there are not enough left, even with a long duration pulse, to charge the double-membrane capacitor to the breakdown point even at very high applied external fields. By partial bleeding of ions, cells become desensitized or able to stand higher fields than the one that would normally lyse them. Once the contact double membranes are broken down the cells are still kept in contact by dielectrophoresis to insure the establishment of a single outside envelope. In Plates 1-6 one can observe the evolution of the outside envelope of several cells fused together into one unit. The giant cell in Plate 4 is still in an early stage of fusion with a narrow waist. In Plate 5 the diameter of the waist is increased despite the stretching dielectrophoretic forces still applied due to the RF fields. Once the RF field is removed Plate 6 shows the giant cell on its way to become a sphere. Without the holding RF it has drifted against the

ELECTROFUSION OF CELLS

343

electrode. In Plates 1-3 even a larger number of cells is seen electrofusing. In Plate 1, cells are aligned into a pearl chain by dielectrophoresis before the fusion pulse. After the fusion pulse Plates 2 and 3 show the evolution of the fusion process. The cells of the Plates 1-6 are myeloma cells P 3 x 63-Ag8-653.

Electrofusion Experiments
Typically, the electrofusion is carried out in electrode chambers in which the fusion operation can be observed under the microscope. The commercially available ones (D.E.P. Systems, Inc., 4350 Barber Road, Metamora, MI 48455) are of two basic types: open top and closed. The open top chambers are typically used for small batch type experiments. They permit easy loading and access to the active fusion area between the platinum electrodes and are used where sterile conditions can be relaxed, although their operation can be done with sterile conditions readily maintained, as in a proper hood, etc. The closed chamber can be operated in a semi-continuous manner to fuse cells. Here the maintenance of sterile conditions is somewhat easier. In a typical experiment, the cells to be fused, or to be subject to field pulses to help drive in desired agents (e.g., DNA) from the surrounding medium, are first removed from their medium by gentle centrifugation, resuspended in deionized and isoosmotic sucrose, say, and again spun out. This process of spinning out and resuspending in the deionized and isoosmotic medium is repeated until the conductivity of the suspension is less than about 0.00001 mho/cm (10 ps/ cm). If it is desired to pretreat the cells with proteases or cellulases to decorticate them and produce protoplasts, this can also be done at this time. Following this, the cells (or perhaps now protoplasts) can be injected into the fusion chamber and the fusion carried out. The cells are first gathered onto each other with gentle DEP by the application of the ac field, then subject to the dc pulse(s) of preselected amplitude and duration. In a typical experiment, an ac voltage of some 7 V at 250 kHz is used to gather the cells and produce the desired interfacial contact, following which a pulse of, say, 60 V lasting for 40 p s is made. (It is usually desirable to reduce the ac chaining voltage to about f value during the application of the dc pulse.) If this process of DEP plus a pulse had produced the desired fusion result, the fused cells would be let sit for a few moments so as to permit the fused portion to grow and knit the cells together. A gentle influx of medium may be injected to help induce better overall shaping of the fused cells. The fused cells are then drawn out, the chamber cleaned and readied for the next use. Fusion can be used as a technique to introduce DNA, plasmids, enzymes, proteins, markers, etc., inside cells. In a preliminary step the molecules of interest are absorbed from solution on the surface of the cells. Then the cells are brought into contact by DEP. If the molecule of interest is trapped in the cell-to-cell contact zone it will be introduced by the fusion process inside the new envelope

344

POHL, POLLOCK, AND RIVERA

of the double cell. In another version the dilute suspension of cells is simply subject to high field pulses (without DEP) and the resulting poration of the cells allows some molecules of interest to enter the cells. The fusion of like cells becomes a technique to introduce desirable molecules into the new unit.

Conclusions We have described how cells and other biological particles can be made to fuse together by the use of the highly efficient electrofusion technique. Its application to the production of giant cellular entities, to the production of intercellular fusion (e.g., hybridomas), and to the incorporation of desired chemical agents (such as DNA, etc.) into cells or vesicles will provide the life scientists with powerful new means for genetic engineering and other research and production approaches.

Acknowledgments This work has been supported by the Pohl Cancer Research Laboratory, Inc. Thanks are due to Dr. John P. Biscar for assistance in making the microphotographs.

Bibliography
[ I ] P. J. Edelson and Z. A. Cohn, in Membrane Fusion, G. Poste and J. L. Nicholson, Eds. (Elsevier, Amsterdam, 1978), pp. 369-385. [2] J . A. Lucy, in Ref. 1, pp. 267-304. [3] G. Poste and C. A. Pastemak, Cell Surface Reviews, Membrane Fusion V (Elsevier, Amsterdam, 1978), pp. 305-367. [4] H. &la and D. Brooks, in Monoclonal Hybridom Antibodies, J. G. R. Hurrell, Ed. (CRC Press, 1982), pp. 1-29. [ 5 ] H. A. Pohl, Dielectrophoresis, The Behavior of Mutter in Nonuniform Electric Fields (Cambridge University Press, London, 1978). [6] W. Weber, W. Foerster, H. E. Jacob, and H. Berg, Z. Allg. Mikrobiol. 21, 555 (1981). [7] E. G. Neumann, G . Gerisch, and K. Opatz, Natunvissenschaften 67, 414 (1981). [8] U. Zimmermann and J. Vienken, J. Membrane Biol. 67, 165 (1982). [9] J. Teissie, V. P. Knutson, X. Tsong, and M. D. Lane, Science 216, 537 (1982). [lo] U. Zimmermann, G. Pilwat, and H. A. Pohl, J. Biol. Phys. 10, 43 (1982). 11 I] H. Rivera, H. A. Pohl, P. Czerski, and M. L. Swicord, J. Biol. Phys. 11, 63 (1983). [I21 J. E. Rhoads, H. A. Pohl, and R. E. Buckner, J. Biol. Phys. 4, 93 (1976). [I31 M. Senda, J. Takeda, Sh. Abe, and T. Nakamura, Plant Cell Physiol. 20, 1441 (1979). (141 P. Scheurich, U. Zimmermann, M. Mischel, and I. Lamprecht, Z. Naturforsch. 35c, 1081 (1980). [I51 H. A. Pohl, J. Appl. Phys. 22, 869 (1951). [I61 H. A. Pohl, Sci. Am. 203, 106 (1960). 1171 H. A. Pohl, J. Biol. Phys. 1, 1 (1973).

ELECTROFUSION OF CELLS

345

[I81 R. Pethig, Dielectric and Electronic Properties o Biological Materials (Wiley, New York, f 1979). [I91 H.Fricke, Phys. Rev. 24, 575 (1924). [20] H.Fricke, Phys. Rev. 24, 678 (1924). [21] A . H.Von Hippel, Dielectrics and Waves (Wiley, New York, 1954). p. 175. [22] H.P. Schwan, Adv. Biol. Med. Phys. 5, 147 (1957). 1231 H.P. Schwan, Ann. N.Y. Acad. Sci. 103, 198 (1977). 1241 E. E. Lafon and H.A. Pohl, J . Electrostat. 9, 209 (1981). 1251 F. A. Sauer (personal communication, 1982).