Elementary [Ca21]i signals generated by electroporationfunctionally mimic those evoked by hormonalstimulation

FEDJA BOBANOVIC,*,† MARTIN D. BOOTMAN,*,‡,1 MICHAEL J. BERRIDGE,*,‡

NICOLA A. PARKINSON,*,2 AND PETER LIPP**Laboratory of Molecular Signalling, Babraham Institute, Cambridge, England, U.K.; †Laboratory ofBiocybernetics, Faculty of Electrical Engineering, University of Ljubljana, Slovenia; and ‡Departmentof Zoology, University of Cambridge, Cambridge, England, U.K.

ABSTRACT The generation of oscillations andglobal Ca21 waves relies on the spatio-temporal recruit-ment of elementary Ca21 signals, such as ‘Ca21 puffs’.Each elementary signal contributes a small amount ofCa21 into the cytoplasm, progressively promotingneighboring Ca21 release sites into an excitable state.Previous studies have indicated that increases in fre-quency or amplitude of such hormone-evoked elemen-tary Ca21 signals are necessary to initiate Ca21 wavepropagation. In the present study, an electroporationdevice was used to rapidly and reversibly permeabilizethe plasma membrane of HeLa cells and to allow alimited influx of Ca21. With low field intensities (100–500 V/cm), brief (50–100 ms) electroporation trig-gered localized Ca21 signals that resembled hormone-evoked Ca21 puffs, but not global signals. With suchlow intensity electroporative pulses, the Ca21 influxcomponent was usually undetectable, confirming thatthe electroporation-induced local signals representedCa21 puffs arising from the opening of intracellularCa21 release channels. Increasing either the frequencyat which low-intensity electroporative pulses were ap-plied, or the intensity of a single electroporative pulse(>500 V/cm), resulted in caffeine-sensitive regenera-tive Ca21 waves. We suggest that Ca21 puffs caused byelectroporation functionally mimic hormone-evokedelementary events and can activate global Ca21 signalsif they provide a sufficient trigger.—Bobanovic, F.,Bootman, M. D., Berridge, M. J., Parkinson, N. A.,Lipp, P. Elementary [Ca21]i signals generated by elec-troporation functionally mimic those evoked by hor-monal stimulation. FASEB J. 13, 365–376 (1999)

Key Words: Ca21 puffs z Ca21 waves z HeLa cells z confocalmicroscopy z InsP3 receptors

Hormonal stimulation of many cell types triggersproduction of the Ca21-mobilizing messenger inosi-tol 1,4,5-trisphosphate (InsP3),3 which activates re-petitive intracellular Ca21 ([Ca21]i) oscillations (1,2). The subcellular correlate of [Ca21]i oscillations

are waves, where [Ca21]i is initially elevated in alocalized region of a cell and then propagates acrossthe entire cell in a regenerative manner. Recentevidence has suggested that such global [Ca21]i

oscillations and waves are generated by the recruit-ment of elementary Ca21 release events, such as‘Ca21 puffs’ (3, 4), and ‘Ca21 sparks’ (5–7; forreviews, see refs 8–11).

We previously observed that Ca21 puffs both initi-ated and propagated Ca21 waves in hormonallystimulated HeLa cells (4). The Ca21 puffs evokedduring agonist stimulation of HeLa cells are highlylocalized Ca21 signals (;6 mM in diameter), arisingfrom clusters of InsP3 receptors (InsP3Rs) spaced ;6mm apart (4). They occur stochastically during cellstimulation, and unless adjacent elementary Ca21

release sites become functionally coupled their Ca21

signals remain spatially confined (12). Essentially,each Ca21 puff contributes a small quantum (a fewattomoles) of Ca21 into the cytoplasm. When theCa21 puff sites operate at low frequencies or ampli-tudes, the released Ca21 can be effectively bufferedor resequestered by the cell. However, with sufficientactivity, the released Ca21 can overwhelm the cellu-lar buffering mechanisms, leading to a progressive[Ca21]i increase that eventually triggers a regenera-tive Ca21 wave (12, 13).

In the present study, we used an electroporationdevice to evoke spatially restricted, rapid and revers-ible influx of Ca21 into HeLa cells. Electroporationessentially involves the application of a high-intensityelectric field across a cell, leading to changes inplasma membrane potential that cause a localizedbreakdown of the lipid bilayer (14–16). The electro-

1 Correspondence: Laboratory of Molecular Signalling,Babraham Institute, Babraham, CB2 4AT, Cambridge, U.K.E-mail: martin.bootman@bbsrc.ac.uk

2 Present address: Department of Physiology, UniversityCollege London, Gower Street, London, WC1E 6BT.

3 Abbreviations: [Ca21]i, intracellular Ca21; EM, externalmedium; InsP3, inositol 1,4,5-trisphosphate; InsP3R, InsP3receptor; PLC, phospholipase C; RyR, ryanodine receptor.

3650892-6638/99/0013-0365/$02.25 © FASEB

poration effect is a localized phenomenon becausenot all parts of the membrane exposed to an electri-cal field are subject to the same change in mem-brane potential (17–20). In fact, the change inmembrane potential, and consequently the electro-porative effect, is most prominent at the restrictedparts of the cell facing the electrodes (21–23), par-ticularly the membrane facing the anode, where themembrane potential becomes hyperpolarized. Theextent and duration of permeabilization can bemodulated by changing parameters such as strength,duration, and repetition frequency of the electricfield. The use of electroporation to evoke spatiallydiscrete [Ca21]i changes allowed us to test thehypothesis that the transition from local to global[Ca21]i signals is dependent on the amplitude andfrequency of elementary Ca21 signals (13).

Electroporation of HeLa cells using low field in-tensities evoked Ca21 puffs with the same spatio-temporal characteristics as hormone-evoked elemen-tary events. These electroporation-induced signalswere a consequence of the minute influx of Ca21

that occurred during the rapid and transient plasmamembrane disruption. Electroporation could func-tionally mimic the effect of a hormone on the cells inthat it could trigger localized Ca21 puffs or globallypropagating regenerative Ca21 waves.

MATERIALS AND METHODS

Cell culture and fura-2 measurement

HeLa cells were cultured and prepared for imaging as de-scribed previously (24). For all experiments, cells were bathedin an external medium (EM) containing (in mM: NaCl 127,KCl 5, MgCl2 2, NaH2PO4 0.5, NaHCO3 5, HEPES 10, glucose10, CaCl2 1.8; pH 7.4). All experiments were carried out atroom temperature (20–22°C).

For video imaging studies, a coverslip of fura-2-loaded cellswas mounted in a homemade electropermeabilization cham-ber (see below). The chamber was placed on the stage of aNikon Diaphot 300 inverted microscope. Video images wereobtained using an intensified CCD camera (COHU, SanDiego, Calif.), stepping filter wheel (Rainbow; Life SciencesResources Ltd., U.K.), and Merlin acquisition system (LifeScience Resources Ltd., Cambridge, U.K.). The cells werealternately excited at 340 nm and 380 nm using a 75W Xenonlamp in conjunction with 7.5 HBW filters (Omega Optical,Dallas, Tex.); fluorescence emission was measured at 510 nm(1 image per second). [Ca21]i was estimated from the 340nm/380 nm fluorescence ratio according to the equation:[Ca21]i 5 Kd 3 b ((R-Rmin)/(Rmax-R)) (25). Rmin and Rmaxwere determined empirically by permeabilizing the cells toCa21 using 5 mM ionomycin (Sigma, Poole, Dorset, U.K.) andexposing the cells to extracellular solutions of high (10 mM)and low Ca21 (no added CaCl2 110 mM EGTA). Typicalvalues obtained were Rmin 5 0.15, Rmax 5 2. The dissociationconstant, Kd, was taken to be 225 nM (25). Experimentsinvestigating Mn21 quench of fura-2 were performed usingstandard EM supplemented with 250 mM MnCl2. The fluores-cent indicator was excited in Ca21-insensitive absorptionwavelength 360 nm (isosbestic point).

Confocal imaging

Confocal images were obtained using a Noran Oz laser-scanning confocal microscope (Noran, Milton Keynes, U.K.).Fluo-3-loaded cells mounted in the electroporation chamberwere placed on the stage of a Nikon Diaphot 300 microscope.Fluo-3 was excited using the 488 nm line of an argon-ionlaser, and the fluorescence emission was collected at wave-lengths .505 nm. The confocal slit was chosen to give a ‘z’resolution of ;1 mm. Images were acquired with a rate of 7.5images per second. Data processing was performed as de-scribed previously (4).

Measurements of relative membrane potential changeswere performed using di-8-ANEPPS-loaded cells as describedby Huser et al. (26). The cells were loaded with di-8-ANEPPSby incubation with a 5 mM concentration of the dye for 10min. Di-8-ANEPPS was excited at 514 nm and the fluores-cence emission was monitored at wavelengths of .525 nm. Tofollow the rapid membrane potential changes with sufficienttemporal resolution, we used the line-scan mode of theconfocal microscope (250 ns/line).

Electroporation

Electroporation was performed using a homemade systemcapable of generating mono- and bipolar square electricalpulses, with amplitudes ranging from 0 to 2000 V and aduration of 1 to 1000 ms. The amplitude and the waveform ofthe electric field exposure were monitored using a digitalstorage oscilloscope (Tektronix, Beaverton, Oreg.). The elec-troporation system essentially consists of a Perspex chamberwith two electrodes connected to a high-frequency voltagesource. The volume of the rectangular chamber (50 ml) waskept as small as possible to enable the application of auniform electric field simultaneously with a consistent rapidperfusion of EM. The cells used for the experiments werelocated in the central part of the narrow (3 mm) chamberfilled with EM that brings into contact two platinum elec-trodes. Perfusion of cells in the electroporation chamberavoids artifacts resulting from the generation of electrolysisby-products, such as O2 and H2. The temperature of theextracellular solution was unaffected by the electrical pulse(data not shown).

For video imaging studies, square monopolar pulses ofeither 10 or 50 ms duration were used. For the confocalstudies, a square bipolar pulse of 50 ms duration in bothdirections (100 ms total duration) was applied in order toincrease the chances of finding responsive cells at near-threshold levels of stimulation. For the confocal studies usingdi-8-ANEPPS, where changes in membrane potential weremonitored, longer pulses of 1 ms were used.

RESULTS

Electroporation-induced [Ca21]i signals

Initial experiments characterized the source andnature of the electroporation-induced [Ca21]i sig-nals in HeLa cells. Electroporation of cells reproduc-ibly evoked Ca21

i increases in the presence of extra-cellular Ca21 (Ca21

o) (Fig. 1A). After removal ofCa21

o such responses were absent, which suggeststhey were due, or at least triggered by, a Ca21 influx.In most cells the response consisted of a single

366 Vol. 13 February 1999 BOBANOVIC ET AL.The FASEB Journal

transient [Ca21]i increase, although in some cells(,10%; n.2000) a single electroporative pulse trig-gered a series of progressively diminishing [Ca21]i

oscillations (Fig. 1B).

The stepwise quench of fura-2 by Mn21 (Fig. 1C)confirmed that the cation permeability of the cellsincreased with each electroporative pulse. Over therange of 0.6–1 kV/cm, both the quench of fura-2 byMn21 and the corresponding increases in [Ca21]i

showed a supralinear relationship to the field inten-sity (Fig. 2A, B). These data indicate that increasingelectroporative potentials cause greater cation entryand that at a field intensity of ;700 V/cm, a thresh-old was reached where [Ca21]i responses weregreatly enhanced (Fig. 2A; filled squares). In cellswhere [Ca21]i oscillations were initiated by a single

Figure 1. Electroporation evokes transient increases in Ca21

resulting from inward cation flux. Panel A represents anaveraged response of fura-2-loaded HeLa cells to three elec-troporative pulses (denoted by vertical arrows marked EP) inthe presence or absence of Ca21

o. B) An oscillatory Ca21

signal arising from a cell in Ca21-containing medium stimu-lated with a single electroporative pulse. This pattern ofresponse was seen in a minority of cells. C) The extracellularmedium was supplemented with 250 mM MnCl2 and thefluorescence emission at 510 nm was monitored for 340, 360,and 380 nm excitation. The traces show that each [Ca21]i risecoincided with a transient increase in the rate of cation entry.The electroporative pulses (EP) (A, C, 1000 V/cm, 50 msduration, monopolar; B, 700 V/cm, 10 ms duration) wereapplied at the times shown by vertical arrows. The effect offield intensity on Mn21 quench of fura-2 was qualitativelysimilar in the presence or absence of Ca21

o (data not shown),confirming that neither the absence of Ca21 in the bathingmedia nor a [Ca21]i increase affected the sensitivity of thecells to electroporation. Panel A represents an average re-sponse (n537); panels B and C illustrate typical responsesfrom single cells.

Figure 2. Electroporation-induced Ca21 increases are deter-mined by the magnitude of the applied field and are inhib-ited by caffeine and thapsigargin. A) The amplitude of[Ca21]i increase triggered by application of different inten-sity electroporative fields in the absence (filled squares) orpresence (open circles) of 20 mM caffeine. The extent ofcation entry into the cells was also dependent on the fieldintensity, as illustrated in panel B. A, B) The data arepresented as mean 6sem (n between 6 and 26). C) Electro-poration-induced [Ca21]i increases were reduced to ;10% ofthe control response (n538) by treating the cells with thap-sigargin to remove functional Ca21 stores. The electropora-tion pulses used to obtain the data shown in this figure weremonopolar and 50 ms in duration.

367ELECTROPORATION-GENERATED [Ca21]i SIGNALS

electroporative pulse (e.g., Fig. 1B), the Mn21

quench of fura-2 revealed a single cation entry stepsimultaneous with the electroporation pulse, but nofurther discrete steps during the subsequent oscilla-tions (data not shown).

The supralinearity of the [Ca21]i signals (Fig. 2A)was not simply due to increasing Ca21 influx at highfield intensities, because the [Ca21]i response dis-played a . sixfold increase between 700 and 800V/cm (Fig. 2A), whereas the Mn21 quench responseincreased by only twofold (Fig. 2B). The enhance-ment of [Ca21]i increases with field intensities above700 V/cm was most likely due to amplification byCa21 release from intracellular stores, since thiseffect was absent during incubation of the cells in 20mM caffeine to inhibit InsP3R function. Caffeinereduced the amplitude of the [Ca21]i increases seenwith larger (.700 V/cm) electroporation pulses andproduced an almost linear relationship between fieldintensity and change in [Ca21]i (Fig. 2A; opencircles). Removal of functional intracellular Ca21

stores by prolonged treatment of cells with 200 nMthapsigargin substantially decreased electroporation-induced increases in [Ca21]i (Fig. 2C), also confirm-ing that the major portion of the [Ca21]i signalsarose from Ca21 release. Although HeLa cells ex-press ryanodine receptors (RyRs) (27, 28), theseintracellular Ca21 release channels did not seem tobe involved in the electroporation-evoked [Ca21]i

signals, since preincubation of cells with 10 mMryanodine for 30 min to block RyR opening did notalter the responses (data not shown).

As illustrated in Fig. 2A, low field intensities (500–700 V/cm) did not trigger large regenerative re-sponses. An example of the low-amplitude, non-regenerative responses obtained with such lowelectroporative potentials is shown in Fig. 3A. Repet-itive application (1/min) of the electroporativepulse evoked only modest (;20 nM) increases in[Ca21]i, which decayed back to basal levels in amatter of a few tens of seconds. But if such pulseswere applied at a twofold higher frequency, theyeventually triggered a regenerative response (Fig.3B). Similarly, low-frequency (1/min) electropora-tive pulses applied in the presence of a low (100 nM)histamine concentration also triggered a regenera-tive response (Fig. 3C). This subthreshold histamineconcentration was unable to evoke increases in[Ca21]i when applied on its own for periods of up to45 min (data not shown).

A consistent feature of [Ca21]i increases triggeredby electroporation was that they displayed desensiti-zation if large regenerative responses were induced,and the cells needed a period of recovery beforesimilar responses could be elicited again. For exam-ple, although pulses of 1000 V/cm evoked consistentamplitude responses when applied at low frequency

(1/6 min, Fig. 3D), the [Ca21]i increases triggeredby the same stimulus rapidly desensitized when thefrequency was increased to 1/min (Fig. 3E). Sinceelectroporation of cells invariably caused cation en-try, as measured by Mn21 quenching of fura-2 (e.g.,Fig. 1C), this desensitization did not reflect a pro-gressive failure to electroporate the cells but ratherthe progressive failure of regenerative Ca21 release.

These data suggest that electroporation caused aninflux of Ca21, which subsequently evoked regener-ative Ca21 release from InsP3Rs on intracellularstores. This temporal sequence of events is furthersupported by the observations that regenerative[Ca21]i changes could occur with a latency of tens ofseconds after the electroporative pulse (see below)and that electroporation-induced [Ca21]i oscilla-tions could persist for several minutes (Fig. 1B).

Figure 3. Low-frequency electroporative pulses evoke repro-ducible Ca21 signals. Panels A–C illustrate the effect of alow-intensity electroporation (filled arrows, 700 V/cm; 10 msduration; monopolar) applied at 1/min (A), 1 per 30 s (B), or1/min in the presence of a subthreshold histamine concen-tration (C). Panels D, E show the response to a high-intensityelectroporation (open arrows, 1000 V/cm; 50 ms duration;monopolar) applied at 1/6 min (D) or 1/min (E). A, D, E)Averaged responses from 17, 23, and 21 cells, respectively. B,C) Responses from multiple individual cells. The data in Abwere obtained from the section marked in Aa, and arepresented with an expanded y-axis scale.

368 Vol. 13 February 1999 BOBANOVIC ET AL.The FASEB Journal

Electroporation triggers regenerative [Ca21]i

waves, but does not stimulate InsP3 production

To confirm that electroporation triggered regenera-tive Ca21 release, we investigated the subcellularproperties of the electrically stimulated [Ca21]i tran-sients using rapid confocal imaging. Application ofsuprathreshold electric fields evoked [Ca21]i signalsthat began first in the regions of the cells closest tothe electrodes, then propagated throughout the cellsin a nondecremental manner (Fig. 4). Such regen-erative [Ca21]i waves could be reproducibly elicited,with the initiation points consistently being the partsof the cells most adjacent to the electrodes and withthe same direction of wave propagation. The veloci-ties of these electroporation-induced Ca21 waveswere only marginally less than those evoked byhistamine (Fig. 5A, C; ref 4). Preincubation of thecells with caffeine (20 mM) markedly reduced theamplitude and velocity of such [Ca21]i waves tomore slowly diffusing signals that penetrated

through the cytoplasm in a decremental manner(Fig. 5B, C).

The observation that electroporation above a cer-tain threshold led to a caffeine-inhibitable regener-ative [Ca21]i rise was surprising, since even thoughInsP3Rs are sensitive to Ca21, they usually requirecoactivation with InsP3. We therefore tested whetherelectroporation simply mimicked hormone actionon HeLa cells by evoking the production of InsP3,thereby triggering Ca21 release from intracellularstores. We used the phospholipase C (PLC) inhibi-tor, U-73122, to prevent production of InsP3 duringan electroporative pulse. U-73122 inhibited hista-mine-evoked [Ca21]i signals in a concentration-de-pendent manner (IC505273 nM; Fig 6A). The struc-tural analog U-73343, which is a much less potentPLC inhibitor, did not affect histamine-evoked Ca21

signals at concentrations of up to 10 mM (Fig. 6A).Electroporation-induced [Ca21]i signals were unaf-fected by U-73122 at concentrations that virtuallyabolished histamine responses (Fig. 6C).

Figure 4. Electroporation-induced Ca21

wave. Panels A, B illustrate the spatial andtemporal profile of an electroporation-evoked Ca21 wave in a fluo-3-loadedHeLa cell (500 V/cm; 100 ms duration;bipolar pulse). A) A sequence of confocalimages of the stimulated cell at the timesindicated in panel B. The [Ca21]i in thecircular regions (1 and 2) shown in Aa areplotted in panel B. C) A surface represen-tation of Ab demonstrating the initiallyinhomogeneous [Ca21]i signal triggeredby electroporation.

369ELECTROPORATION-GENERATED [Ca21]i SIGNALS

Electroporation caused rapid membranedisruption, followed by Ca21 influx and activationof Ca21 puffs

With low field intensities, the Ca21 influx occurringduring the electroporative pulse was small and usu-ally barely detectable, and it was the subsequentregenerative Ca21 signals that confirmed electropo-ration had actually taken place. However, withhigher field intensities ($750 V/cm), transient in-creases in [Ca21]i were occasionally observed thatwere far too rapid to be due to regenerative Ca21

release from InsP3Rs. Such [Ca21]i signals repre-sented the influx of Ca21 through the electropora-tion-induced membrane pores, and thus provided away of estimating the life-time of these pores. An

example of such a Ca21 influx signal is illustrated inFig. 7. Concomitant with the electroporative pulse, a[Ca21]i increase was observed at one pole of the cell(Fig. 7A, B). The influx signal (Fig. 7B) rapidlyreached a peak [Ca21]i of ;200 nM, and triggered amore slowly developing regenerative [Ca21]i in-crease throughout the cell (Fig. 7C). The expandedtime course of the influx signal (Fig. 7D) and its timederivative (Fig. 7E) revealed that the [Ca21]i in-crease resulting from membrane disruption lastedfor only ;1 s, with a rising phase of #4 ms. Such aresponse can only arise via near-instantaneous Ca21

entry after membrane disruption, followed by rapidmembrane resealing. We therefore conclude thatthe lifetime of the electroporation-induced mem-brane pores is in the range of ;1–2 ms.

Figure 5. Caffeine reduces electropo-ration-evoked Ca21 waves to a slowlydiffusing low-amplitude [Ca21]i in-crease. A, B) [Ca21]i signals triggeredby repetitive electroporative pulses(500 V/cm; 100 ms duration; bipolarpulse) applied to the same cell in theabsence (A) or presence (B) of 20 mMcaffeine. The electroporative pulseswere applied at the vertical arrows. Aaand Ba are line-scan images of theevoked [Ca21]i signals. Ab and Bbshow the temporal profile of [Ca21]ialong the regions indicated by thecolored horizontal arrows. For panelC, the [Ca21]i signals in the line-scanimages in panels A and B were maskedat a [Ca21]i value corresponding to150 nM (solid color lines; mask frompanel A is the red line and mask frompanel B is the green line). The velocityof [Ca21]i signal propagation was cal-culated from the slope of the derivedmask (indicated by the two dashedlines). In the presence of caffeine, theelectroporation-evoked Ca21 wavehad approximately one-third the ve-locity of the control response.

370 Vol. 13 February 1999 BOBANOVIC ET AL.The FASEB Journal

To more closely monitor the effect of the electricfield on the plasma membrane and its subsequentbreakdown, we used the rapid voltage-sensitive indi-cator di-8-ANEPPS. Upon application of an electro-porative pulse, a sudden change in membrane po-tential occurred in the regions of the cells mostclosely opposed to the electrodes (Fig. 8A). Thechange in membrane potential correlated with theposition of the cell relative to the electrodes; themembranes facing the anode or cathode displayedhyperpolarization or depolarization, respectively(Fig. 8A).

Two significant observations can be gleaned fromthe measurement of the di-8-ANEPPS signals. First,the electroporation-induced voltage change peaked

and then started to decline within 0.5 ms of the startof the pulse (Fig. 8B), even though the intensity ofthe field was constant during the exposure, whichlasted for 1 ms (data not shown). The recovery of themembrane potential during the application of theelectroporative pulse indicated that the resistance ofthe membrane rapidly decreased, and consequentlythat pores had been formed. Second, the change inmembrane potential can be repeatedly elicited at afrequency of up to 10 Hz (Fig. 8C), indicating rapidrecovery of the membrane resistance, i.e., that themembrane has resealed. This showed that electropo-ration with pulses of 10-fold longer duration than

Figure 6. Inhibition of hormone- but not electroporation-evoked [Ca21]i signals by U-73122. A) An example of inhibi-tion of the histamine-evoked [Ca21]i signal by 10 mM con-centration of U-73122 (averaged response; n525). B) Theconcentration-dependent inhibition of histamine-evoked[Ca21]i responses in HeLa cells (20 mM histamine; n521–37)by U-73122 (filled squares) but not U-73343 (open square).Panel C shows the lack of effect of U-73122 (10 mM) onelectroporation-evoked [Ca21]i signals (1000 V/cm; 50 msduration; monopolar pulse; averaged response; n527).

Figure 7. Rapid [Ca21]i increase arising directly from Ca21

influx. A–E) Spatial and temporal profile of [Ca21]i in anelectroporated HeLa cell (500 V/cm; 100 ms duration; bipo-lar pulse), which displayed a rapid Ca21 influx signal. Suchresponses were observed in a minority of cells (,5%; n590).The electroporative pulses were applied at the times indi-cated by the vertical arrows. A) Surface representation of theevoked [Ca21]i rise. The [Ca21]i signals along the regionsshown by the horizontal arrows are depicted in panels B andC. The rapid Ca21 influx signal is shown in panel B, and thesubsequent regenerative response is illustrated by the tracesin panel C. To more clearly illustrate the rapid Ca21 influxsignal in panel B, the signal and its derivative are shown on anexpanded time scale in panels D and E,, respectively.

371ELECTROPORATION-GENERATED [Ca21]i SIGNALS

those used to evoke Ca21 signals in this study did notimpose long-term membrane disruptions.

In the majority of cases, where electric fields ofmodest intensities were used, [Ca21]i increases aris-ing directly from Ca21 entry were difficult to detect.Instead, there was usually a variable latency of up toa few seconds between the electroporative pulse andthe onset of a [Ca21]i signal, suggesting that therapid undetectable Ca21 influx had activated Ca21

release (Fig. 9Aa). Although the exact nature of the[Ca21]i signal triggered by electroporation variedwith the magnitude of the electroporative pulse, the

most common response to low field intensities(#750 V/cm) was the activation of Ca21 puffs.

We earlier determined the nature of hormonallyinduced Ca21 puffs in HeLa cells (4, 13), and foundthem to be rapid (rise time, ,100 ms) and highlylocalized (full width at half maximum; FWHM up to6 mm). Electroporation evoked similar elementaryevents, usually occurring in the regions of the plasmamembrane most closely adjacent to the electrodes.These electroporation-induced Ca21 puffs tempo-rally and spatially resembled those induced by ahormone; they were visualized as spatially confined[Ca21]i increases, with an FWHM of 2–7 mm (Fig.9Ab). A single Ca21 puff was the most commonresponse to an individual electroporative pulse, al-though trains of Ca21 puffs were sometime evoked(Fig. 9B).

The transition between nonregenerative andregenerative responses was dependent onelectroporation intensity and frequency

Although the electroporation-induced Ca21 puffsoriginated beneath the regions of the plasma mem-brane most closely opposed to the electrodes, theirexact location was unpredictable. In addition, thethreshold field intensity required to evoke Ca21

puffs differed both between cells and between vari-ous parts of individual cells (data not shown). How-ever, once a threshold stimulus intensity had beenfound to elicit Ca21 puffs, pulsatile application ofthat voltage triggered repetitive [Ca21]i transients inthe same cellular location. The reproducible induc-tion of Ca21 puffs allowed us to investigate themodes of recruitment of such elementary events thatinduce globally regenerative [Ca21]i signals.

As illustrated in Fig. 2A, the intensity of theelectroporation pulse applied to the cells deter-mined the likelihood of triggering a regenerative[Ca21]i signal. Subthreshold field intensities gaverise to discrete Ca21 puffs, which at best triggeredonly a limited degree of local regeneration (Fig. 9Band Fig. 10). Despite repeated applications, suchlow-intensity fields did not trigger global [Ca21]i

signals. However, increasing the voltage applied tothe cells could switch the response from spatiallyrestricted Ca21 signals to a regenerative [Ca21]i wave(Fig. 10).

In addition to the effects of enhancing field inten-sity, increases in the frequency of electropora-tive pulses could also induce regenerative [Ca21]i

signals. The sequence of traces shown in Fig. 11A–Dillustrates the effects of increasing the frequency of athreshold electroporation pulse. At 0.05 or 0.1 Hz,the electroporative pulses evoked low-amplitude[Ca21]i signals, which were restricted to only acouple of cellular regions (Fig. 11A, B). Although

Figure 8. Electroporation-induced changes in membranepotential measured with di-8-ANEPPS. A, B) The effect of anelectroporation pulse (500 V/cm; monopolar pulse; 1 msduration) on the membrane potential of the single HeLa cell.The surface plot (A) and line traces (B) show the change ofdi-8-ANEPPS fluorescence, with downward or upward deflec-tion indicating depolarization (a) or hyperpolarization (b),respectively. The lines above the figures show the timing ofthe pulse. The region of the cell chosen for line scanning isshown in panel A (dashed line on inset cell image). The barchart (C) shows the ratio between peak fluorescence duringthe field stimulation at hyperpolarized side of the cell (Fh)and resting fluorescence of the same membrane site (Fo) foreach of 25 consecutive monopolar pulses (500 V/cm, 1 ms, 10Hz).

372 Vol. 13 February 1999 BOBANOVIC ET AL.The FASEB Journal

there was some variability, presumably due tochanges in the degree of local regenerativity, eachelectroporative pulse caused a localized [Ca21]i

increase. However, even after multiple pulses, aglobal [Ca21]i signal was not triggered. At 0.2 Hz,spatially restricted [Ca21]i signals were evoked,

Figure 9. Ca21 puffs and local Ca21 signalstriggered by electroporation. A, B) Individual(A) and multiple (B) Ca21 puffs evoked by asingle electroporative pulse (100 ms duration;bipolar pulse). The electroporative pulses wereapplied at the vertical arrows. For panels A, Bthe field intensities were 400 and 500 V/cm,respectively. Aa shows the time course of[Ca21]i in distinct regions of the electroporatedcell that gave no response (Aa1) or displayed aCa21 puff (Aa2). To more clearly illustrate thespatio-temporal profile of the electroporation-evoked Ca21 puff, the event is presented as asurface plot in panel Ab. B) A train of Ca21

puffs resulting from a single electroporativepulse. As shown in the line-scan plot in panelBa, these Ca21 puffs triggered only a limitedamount of local regeneration. The [Ca21]i re-sponses along the regions shown by the hori-zontal arrows are depicted in panel Bb. Forclarity, the more obvious puffs are marked withan asterisk.

Figure 10. Induction of regenerative Ca21

signals depends on electroporation inten-sity. Panels A, B depict the spatial and tem-poral profile of [Ca21]i in an individualHeLa cell stimulated repeatedly (0.3 Hz;vertical arrows in panel B) with a subthresh-old intensity field (400 V/cm; 100 ms dura-tion; bipolar pulse), followed by a single(open arrowhead) suprathreshold stimula-tion (500 V/cm; 100 ms duration; bipolarpulse). [Ca21]i in the two cellular regionsdepicted in Aa is plotted in panel B. Thetimes at which the cell images in panel Awere obtained are shown by the vertical linesin panel B.

373ELECTROPORATION-GENERATED [Ca21]i SIGNALS

which were not able to recover back to basal levelsbefore the onset of the next [Ca21]i rise. Eventu-ally, the elevated [Ca21]i in the responsive regionsdiffused throughout the rest of the cell, causing auniform [Ca21]i elevation (Fig. 11C), however, aregenerative [Ca21]i signal was not triggered. In-creasing the frequency of electroporation pulses to0.5 Hz initially triggered discrete Ca21 puffs andeventually led to a regenerative globally propagat-ing [Ca21]i wave (Fig. 11D).

DISCUSSION

We have used electroporation to evoke small andbrief Ca21 influx events into restricted subplasma-

lemmal regions of HeLa cells, and investigated theirsubsequent ability to trigger local and global re-sponses. The main reason for using electroporationin this way is that it is substantially more reproduc-ible and controllable than hormone stimulation. Inaddition, the level of permeabilization at particularparts of the cell can be varied in several ways: fieldintensity, duration, frequency, and direction. Elec-troporation acts almost instantaneously, is rapidlyreversible, and at low field intensities causes minimalmembrane disruption (Fig. 8). An important featureof electroporation is that it allows permeabilizationof a small, restricted part of the membrane (Fig. 10),providing control over the timing, amount, and thepart of the membrane through which the transmem-brane Ca21 fluxes are generated.

In HeLa cells, electroporation caused a Ca21o-

dependent [Ca21]i increase (Fig. 1) that was propor-tional to the intensity of the electric field applied tothe cells (Figs. 2 and 10). Although the electropora-tion-evoked signals required Ca21

o, the [Ca21]i in-crease appeared to arise largely through activation ofregenerative Ca21 release by intracellular Ca21 chan-nels, since it was substantially inhibited by caffeine(Fig. 2A and Fig. 5), required functional Ca21 stores(Fig. 2C), and could occur some time after theelectroporative pulse (Fig. 4 and Fig. 9A). In addi-tion, whereas electroporation invariably causedMn21 quench of fura-2 (Fig. 1C), [Ca21]i increaseswere not always recorded (e.g., third electroporationpulse in Fig. 11A). Furthermore, application of high-frequency electroporation pulses desensitized the[Ca21]i response (Fig. 3). Together, these observa-tions suggest that a small, usually imperceptible Ca21

influx activated regenerative Ca21 release from in-ternal stores.

Changes in the concentration of ions other thanCa21—e.g., Na1 and K1—probably occurred duringthe formation of membrane pores by electropora-tion. Exact quantitation of changes in these ions wasnot undertaken during our study. However, we thinkit is reasonable to assume that the largest flux of ionswill be Ca21, as it has the largest electrochemicalgradient across the cell. Since we are unable tomeasure the very small flux of Ca21 during weakelectroporation events, it is unlikely that we couldmeasure changes in Na1 or K1. Furthermore, theseions are much more highly diffusible than Ca21 inthe cytoplasm, and therefore we would expect anychanges in Na1 or K1 to be rapidly reequilibrated.

It was surprising to observe that electroporationon its own could trigger regenerative [Ca21]i signals,since the activation of InsP3Rs is believed to requireboth InsP3 and Ca21 (29–35). However, the electro-poration-evoked [Ca21]i responses were not sensitiveto the PLC inhibitor U-73122 (Fig. 6), suggestingthat it did not look to the cells as if they had been

Figure 11. Induction of regenerative Ca21 signals depends onelectroporation frequency. Panels A–D illustrate the responseof a single HeLa cell to pulsatile electroporation with alow-intensity (400 V/cm; 100 ms duration; bipolar pulse) at: A,0.05 Hz; B, 0.1 Hz; C, 0.2 Hz; D, 0.5 Hz). The traces in panelsA–D represent the [Ca21]i response in the four subcellularareas depicted in the cell image, with the time course of the[Ca21]i signal in each area shown by the correspondingcolored line.

374 Vol. 13 February 1999 BOBANOVIC ET AL.The FASEB Journal

treated with an agonist. It seems that in the absenceof hormonal stimulation, the basal intracellularInsP3 concentration is sufficient to allow regenera-tivity from InsP3Rs, providing a sufficient triggerCa21 is given. The effect of a subthreshold histamineconcentration in sensitizing the cells to low-intensityelectroporative pulses (Fig. 3) is consistent with thenotion that electroporation-induced Ca21 influx andmodest intracellular InsP3 concentrations can elicit asynergistic effect.

Electroporation of HeLa cells with low (,700V/cm) field intensities frequently generated Ca21

puffs with a similar spatial spreading and time courseto hormone-evoked elementary events (Fig. 9). Theability to evoke such localized Ca21 signals essentiallyresults from the fact that electroporation is an asym-metric process that disrupts the membrane regionsmost closely opposed to the electrodes, particularlythe membrane portion facing the anode. At low fieldintensities, only a small area of the plasma mem-brane may be disrupted, giving a highly localizedCa21 influx. We suggest that this small Ca21 influxtriggers Ca21 release from adjacent InsP3R, thusproducing the observed Ca21 puffs.

It should be noted, however, that although Ca21

puffs were commonly observed, not all electropora-tion-evoked Ca21 signals resembled hormonal Ca21

puffs. Application of high-intensity voltages (.750V/cm) sometimes caused subplasmalemmal [Ca21]i

rises that were too fast to be accounted for byregenerative Ca21 release (Fig. 5). These rapid sig-nals most likely arose due to Ca21 entry through theelectroporation-induced membrane pores. The spa-tial spread of these signals was also distinct fromCa21 puffs, with a semicircular diffusion shell imme-diately beneath the membrane (data not shown).

Hormone-evoked regenerative [Ca21]i waves inHeLa cells result from the spatio-temporal recruit-ment of elementary Ca21 events (4, 13). In a previ-ous study, we suggested that the transition fromlocalized elementary events to global Ca21 waves isdetermined by a change in the activity of the elemen-tary signals themselves. Increases in frequencyand/or amplitude of the Ca21 puffs are required toevoke regenerative Ca21 signals (13). ElementaryCa21 signals that do not advance in frequency oramplitude fail to trigger global signals. Cytoplasmicintegration of the Ca21 released during each puff iscrucial to allow the ambient level of Ca21 to reachthe threshold for regenerativity to occur. With sub-threshold levels of activity, the Ca21 signals arisingfrom individual Ca21 puffs have a relatively shortcytoplasmic lifetime before being buffered or seques-tered. With sufficient activity, the Ca21 releasedduring each event overwhelms the cellular bufferingmechanisms and is integrated by the cytoplasm tocause a progressive [Ca21]i increase until the thresh-

old for regnerativity is reached. The reproducibilityof electroporation-induced Ca21 signals allowed usto overcome the stochastic nature of hormone-evoked Ca21 puffs and directly test this scheme.

Increasing the frequency at which low-intensityelectroporative pulses were applied enhanced theirability to trigger regenerative Ca21 waves (Fig. 11).In addition, the effectiveness of single electropora-tive pulses in evoking Ca21 waves was greater withhigher voltages (Fig. 2A and Fig. 10). These datasuggest that the electroporation-induced Ca21 sig-nals functionally mimic hormone-evoked Ca21 puffsin generating regenerative Ca21 waves when theyprovide a suitable trigger.

One significant difference between electropora-tion-induced events and hormonally stimulated ele-mentary Ca21 signals is that whereas the former areevoked beneath the plasma membrane in the pe-riphery of the cells (Figs. 9 and 11), the latter mostfrequently occur adjacent to the nucleus (36). Thereason why agonist-stimulated Ca21 puffs are ob-served largely around the nucleus is unclear. How-ever, the consequence is that a majority of Ca21

waves are initiated from perinuclear regions. Incontrast, electroporation-induced Ca21 waves wereinitiated far from the nucleus (Fig. 5). These datarevealed that different parts of the cytoplasm can befunctionally equivalent, and that the location ofelementary signals determine initiation sites of re-generative responses.

The qualitative effect of triggering regenerativeCa21 waves with particular electroporative frequen-cies or field intensities was observed whether theelectroporation triggered Ca21 puffs or simply adirect rapid Ca21 entry signal (see above). Thissuggests that the important criterion for causingregenerative Ca21 release is not the location ornature of the activating signal, but simply that atriggering [Ca21]i threshold has to be attained.

The work was supported by Babraham Institute’s DEBSInitiative, the BBSRC, and the Royal Society CooperationGrant Scheme. The authors would also like to acknowledgethe generous support of the EMF Biological Research Trust.M.D.B. is a Royal Society University Fellow.

REFERENCES

1. Berridge, M. J. (1993) Inositol trisphosphate and calcium sig-nalling. Nature (London) 361, 315–325

2. Petersen, O. H., Petersen, C. C. H., and Kasai, H. (1994)Calcium and hormone action. Annu. Rev. Physiol. 56, 297–319

3. Yao, Y., Choi, J., and Parker, I. (1995) Quantal puffs of intracel-lular Ca21 evoked by inositol trisphosphate in Xenopus oocytes.J. Physiol. (London) 482, 533–553

4. Bootman, M. D., Niggli, E., Berridge, M. J., and Lipp, P. (1997)Imaging the hierarchical Ca21 signalling system in HeLa cells.J. Physiol. (London) 499, 307–314

5. Cheng, H., Lederer, W. J., and Cannell, M. B. (1993) Calciumsparks—elementary events underlying excitation-contractioncoupling in heart-muscle. Science 262, 740–744

375ELECTROPORATION-GENERATED [Ca21]i SIGNALS

6. Lipp, P., and Niggli, E. (1994) Modulation of Ca21 release incultured neonatal rat cardiac myocytes—insight from subcellu-lar release patterns revealed by confocal microscopy. Circ. Res.74, 979–990

7. Lopez-Lopez, J. R., Shacklock, P. S., Balke, C. W., and Wier,W. G. (1995) Local calcium transients triggered by single L-typecalcium channel currents in cardiac cells. Science 268, 1042–1045

8. Bootman, M. D., and Berridge, M. J. (1995) The elementalprinciples of calcium signaling. Cell 83, 675–678

9. Parker, I., Choi, J., and Yao, Y. (1996) Elementary events ofInsP3-induced Ca21 liberation in Xenopus oocytes—hot-spots,puffs and blips. Cell Calcium 20, 105–121

10. Lipp, P., and Niggli, E. (1996) A hierarchical concept of cellularand subcellular Ca21 signalling. Prog. Biophys. Mol.. Biol. 65,265–296

11. Berridge, M. J. (1997) Elementary and global aspects of calciumsignalling. J. Physiol. (London) 499, 291–306

12. Parker, I., and Yao, Y. (1992) Regenerative release of calciumfrom functionally discrete subcellular stores by inositol trisphos-phate. Proc R. Soc. London B 246, 269–274

13. Bootman, M. D., Berridge, M. J., and Lipp, P. (1997) Cookingwith calcium: the recipes for composing global signals fromelementary events.Cell 91, 367–373

14. Weaver, J. C. (1993) Electroporation: a general phenomenonfor manipulating cell and tissues. J. Cell Biochem. 51, 426–435

15. Weaver, J. C., and Chizmadzhev, Y. A. (1996) Theory of electro-poration: a review. Bioelectrochem. Bioenerg. 41, 135–160

16. Tsong, T. Y. (1991) Electroporation of cell membranes. Bio-phys. J. 60, 297–306

17. Kinosita, K., Jr., Ashikawa, I., Saita, N., Yoshimura, H., Hiro-yashu, I., Nagayama, K., and Ikegami, A. (1988) Electroporationof cell membrane visualized under a pulsed-laser fluorescencemicroscope. Biophys. J. 53,1015–1019

18. Hibino, M., Itoh, H., and Kinosita, K., Jr. (1993) Time courses ofcell electroporation as revealed by submicrosecond imaging oftransmembrane potential. Biophys. J. 64,1789–1800

19. Kotnik, T., Bobanovic, F., and Miklavi, D. (1997) Sensitivity oftransmembrane voltage induced by applied electric fields—atheoretical analysis. Bioelectrochem. Bioenerg. 43, 285–291

20. Bobanovic, F. (1998) Electroporation: a method for introduc-tion of non-permeable molecular probes. In: Fluorescent andLuminescent Probes, 2nd Ed (Mason, W. T., ed) Academic PressLtd., London In press

21. Tekle, E., Astumian, R. D., and Chock, P. B. (1994) Selectiveand asymmetric molecular transport across electroporated cellmembranes. Proc. Natl. Acad. Sci. USA 91, 11512–11516

22. Sun, F. Z., Hoyland, J., Huang, X., Mason W. T., and Moor R. M.(1992) A comparison of intracellular changes in porcine eggsafter fertilization and electroactivation. Development 115, 947–956

23. Teruel, M. N., and Meyer, T. (1997) Electroporation-inducedformation of individual calcium entry sites in the cell body andprocesses of adherent cells. Biophys. J. 73, 1785–1796

24. Bootman, M. D., Taylor, C. W., and Berridge, M. J. (1992) Thethiol reagent, thimerosal, evokes Ca21 spikes by sensitizing theinositol 1,4,5-trisphosphate receptor. J. Biol. Chem. 25113–25119

25. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985) A newgeneration of Ca21 indicators with greatly improved fluores-cence properties J. Biol. Chem. 260, 3440–3450

26. Huser, J., Lipp, P., and Niggli, E. (1996) Confocal microscopicdetection of potential-sensitive dyes used to reveal loss of voltagecontrol during patch-clamp experiments. Pfluegers Arch. 433,194–199

27. Giannini, G., Conti, A., Mammarella, S., Scrobogna, M., andSorrentino, V. (1996) The ryanodine receptor calcium-channelgenes are widely and differentially expressed in murine brainand peripheral tissues. J. Cell Biol. 128, 893–904

28. Bennett, D. L., Cheek, T. R., Berridge, M. J., De Smedt, H.,Parys, J. B., Missiaen, L., and Bootman, M. D. (1996) Expressionand function of ryanodine receptors in non-excitable cells.J. Biol. Chem. 271, 6356–6362

29. Iino, M. (1990) Biphasic Ca21 dependence of inositol 1,4,5-trisphosphate-induced calcium release in smooth-muscle cells ofthe guinea-pig Taenia caeci. J. Gen. Physiol. 95, 1103–1122

30. Finch, E. A., Turner, T. J., and Goldin, S. M. (1991) Calcium asa coagonist of inostiol 1,4,5-trisphosphate-induced calcium re-lease. Science 252, 443–446

31. Bezprozvanny, I., Watras, J., and Ehrlich, B. E. (1991) Bell-shaped calcium response curves of Ins(1,4,5)P3-gated and calci-um-gated channels from endoplasmic-reticulum. Nature (Lon-don) 351, 751–754

32. Iino, M., and Endo, M. (1992) Calcium-dependent immediatefeedback-control of inositol 1,4,5-trisphosphate-induced Ca21

release. J. Gen. Physiol. 360, 76–7833. Missiaen, L., De Smedt, H., Parys, J. B., and Casteel, R. (1994)

Co-activation of inositol trisphosphate-induced Ca21 release bycytosolic Ca21 is loading dependent. J. Biol. Chem. 269, 7238–7242

34. Bootman M. D., Missiaen L., Parys J. B., De Smedt H., andCasteels R. (1995) Control of inositol 1,4,5-trisphosphate-in-duced Ca21 release by cytosolic Ca21. Biochem J. 306, 445–451

35. Missiaen, L., De Smedt, H., Parys, J. B., and Casteels, R. (1996)Threshold for inositol 1,4,5-trisphosphate action. J. Biol. Chem.271, 12287–12293

36. Lipp, P., Thomas, D., Berridge, M. J., and Bootman, M. D.(1997) Nucleoplasmic calcium signalling by individual cytoplas-mic calcium puffs. EMBO J. 16, 7166–7173

Received for publication June 9, 1998.Revised for publication October 19, 1998.

376 Vol. 13 February 1999 BOBANOVIC ET AL.The FASEB Journal