Ca2� Homeostasis during Mitochondrial Fragmentation andPerinuclear Clustering Induced by hFis1*□S

Received for publication, November 12, 2003, and in revised form, March 8, 2004Published, JBC Papers in Press, March 15, 2004, DOI 10.1074/jbc.M312366200

Maud Frieden‡, Dominic James§, Cyril Castelbou‡, Anne Danckaert¶, Jean-Claude Martinou§,and Nicolas Demaurex‡�

From the ‡Department of Physiology and the ¶Bioimaging Core Facility, University of Geneva Medical Center,1 Michel-Servet, CH-1211 Geneva 4, Switzerland and the §Department of Cell Biology, University of Geneva,Quai Ernest-Ansermet 30, 1211 Geneva, Switzerland

Mitochondria modulate Ca2� signals by taking up,buffering, and releasing Ca2� at key locations nearCa2� release or influx channels. The role of such localinteractions between channels and organelles is diffi-cult to establish in living cells because mitochondriaform an interconnected network constantly remodeledby coordinated fusion and fission reactions. To studythe effect of a controlled disruption of the mitochon-drial network on Ca2� homeostasis, we took advantageof hFis1, a protein that promotes mitochondrial fissionby recruiting the dynamin-related protein, Drp1. hFis1expression in HeLa cells induced a rapid and completefragmentation of mitochondria, which redistributedaway from the plasma membrane and clusteredaround the nucleus. Despite the dramatic morpholog-ical alteration, hFis1-fragmented mitochondria main-tained a normal transmembrane potential and pH andtook up normally the Ca2� released from intracellularstores upon agonist stimulation, as measured with atargeted ratiometric pericam probe. In contrast, hFis1-fragmented mitochondria took up more slowly theCa2� entering across plasma membrane channels, be-cause the Ca2� ions reaching mitochondria propagatedfaster and in a more coordinated manner in intercon-nected than in fragmented mitochondria. In parallelcytosolic fura-2 measurements, the capacitative Ca2�

entry (CCE) elicited by store depletion was only mar-ginally reduced by hFis1 expression. Regardless of mi-tochondria shape and location, disruption of mito-chondrial potential with uncouplers or oligomycin/rotenone reduced CCE by �35%. These observationsindicate that close contact to Ca2� influx channels isnot required for CCE modulation and that the forma-tion of a mitochondrial network facilitates Ca2� prop-agation within interconnected mitochondria.

Mitochondria actively participate to the cellular Ca2� home-ostasis and modulate the pattern of agonist-induced Ca2� sig-nals by their ability to sequester and release Ca2� (1). Becauseof the low Ca2� affinity of the uniporter that constitutes themain mechanism of Ca2� entry into mitochondria, it was pro-posed that the ability of these organelles to accumulate Ca2�

relies on their close location to Ca2� release channels on theendoplasmic reticulum (ER)1 (2, 3). Mitochondria also interactwith plasma membrane channels and thereby modulate theso-called capacitative Ca2� entry (CCE) pathway, the ubiqui-tous Ca2� entry mechanism triggered by emptying of the ERCa2� store (4, 5). Although the molecular identity of the chan-nel(s) responsible for CCE as well as its mechanism of activa-tion are still debated, recent evidence indicates that mitochon-dria represent a key organelle in CCE activity and/oractivation. Indeed, CCE is inhibited by intracellular Ca2� ele-vations, and mitochondria were shown to act as local buffers toprevent Ca2�-mediated inhibition of the CCE pathway (6–9).

Local interactions between mitochondria and other subcellu-lar structures are difficult to establish in living cells becausemitochondria display a complex architecture that varies con-siderably between cell types. This ranges from a largely inter-connected tubular network in COS-7, endothelial, or HeLa cellsto round punctuated structures in hepatocytes (10). Moreover,mitochondria are highly dynamic organelles that move in thecytosol and that constantly undergo fusion and fission. Bothprocesses are under the control of certain GTPases and theirassociated proteins (11). hFis1, the human orthologue of theyeast Fis1p (12), is a 17-kDa transmembrane protein located inthe outer membrane of the mitochondria that is involved in themachinery of mitochondria fission, and overexpression of thisprotein enhances the fission process in HeLa cells (13). In thisstudy, we overexpressed the protein hFis1 in HeLa cells toinduce a controlled fragmentation of mitochondria and meas-ured the impact of these structural changes on cytosolic andmitochondria Ca2� signals with fura-2 and with a targetedratiometric pericam probe, respectively. This approach allowedus to investigate the role of mitochondria interconnection oncytosolic and mitochondrial Ca2� homeostasis and to distin-guish the local and global effects of mitochondria on the Ca2�

entry process.

EXPERIMENTAL PROCEDURES

Materials—Minimum essential medium, fetal calf serum, penicillin,and streptomycin were obtained from Invitrogen. Histamine, thapsi-gargin, oligomycin, and rotenone were obtained from Sigma. Acetoxym-ethyl ester form of fura-2 (fura-2/AM) and Mitotracker Red were ob-tained from Molecular Probes Europe (Leiden, Netherlands).Carbonylcyanide m-chlorophenylhydrazone (CCCP) was obtained from

* This work was supported by Grant 31–068317.02 from the SwissNational Science Foundation. The costs of publication of this articlewere defrayed in part by the payment of page charges. This article musttherefore be hereby marked “advertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.

□S The on-line version of this article (available at http://www.jbc.org)contains two supplemental figures and a film.

� To whom correspondence should be addressed. Tel.: 41-22-379-5399;Fax: 41-22-379-5402; E-mail: Nicolas.Demaurex@medicine.unige.ch.

1 The abbreviations and trivial names used are: ER, endoplasmicreticulum; CCE, capacitative Ca2� entry; CCCP, carbonylcyanide m-chlorophenylhydrazone; [Ca2�]mit, mitochondrial [Ca2�]; [Ca2�]cyt, cy-tosolic [Ca2�]; RP3.1mit, ratiometric pericam targeted to the mitochon-drial matrix; SERCA, sarco/endoplasmic reticulum Ca2� ATPase;ICRAC, Ca2� release-activated Ca2� current; ��m, mitochondrial mem-brane potential; TMRM, tetramethylrhodamine methyl ester; GFP,green fluorescent protein.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 279, No. 21, Issue of May 21, pp. 22704–22714, 2004© 2004 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org22704

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Fluka (Buchs, Switzerland). Transfast transfection reagent was pur-chased from Promega.

Cell Culture and Transfection—HeLa cells were grown in minimumessential medium containing 10% heat-inactivated fetal calf serum, 2mM L-glutamine, 50 units/ml penicillin, 50 �g/ml streptomycin and weremaintained at 37 °C under 5% CO2. For experiments, cells were platedon 25-mm diameter glass coverslips 2–3 days before use. After reaching40–60% of confluence, cells were transiently transfected with the dif-ferent plasmids using the Transfast reagent according to the protocolsupplied by the manufacturer. For measurements of cytosolic Ca2�

concentration, [Ca2�]cyt, hFis1 was co-transfected with a GFP targetedto the nucleus to identify cells expressing hFis1. All experiments wereperformed 16–20 h after transfection with hFis1. To measure mitochon-drial Ca2�, [Ca2�]mit, cells were transfected with the ratiometric peri-cam targeted to the mitochondrial matrix (RP3.1mit, a gift from Dr.Atsushi Miyawaki, RIKEN Brain Science Institute, Wako-city, Japan)and 24 h later with hFis1.

Organelle Imaging and Morphometric Analysis—Organelle morphol-ogy was imaged on an Axiovert 200M equipped with an array laserconfocal spinning wheel (Nipkow disc; Visitech, Sunderland, UK) usinga �63, 1.4 NA oil-immersion objective (Carl Zeiss AG, Feldbach,Switzerland). Images were acquired on a cooled, 16-bit CCD camera(CoolSnap HQ; Roper Scientific, Trenton, NJ) operated by the Meta-morph 5.0 software (Universal Imaging, West Chester, PA). Imagesshown in Figs. 1A and 3 were deconvolved with the Huygens algorithm(Scientific Volumetric Imaging, Ilversum, The Netherlands) using theImaris software (Bitplane AG, Zurich, Switzerland). To determine thesurface of the cytosol occupied by mitochondria, HeLa cells stablyexpressing a cytosolic pericam probe were loaded with Mitotracker Red(500 nM for 90 s) to label mitochondria. The cytosolic and mitochondrialstainings were imaged using 488 nm excitation, 535 nm emission, and546 nm excitation, �580 nm emission, respectively. Optical slices of 200nm step size in z section were acquired. Of the stack, the five mostinformative images were visually selected. This corresponded for allcells (control and transfected) to five adjacent slices located at thebottom of the cells. Following this selection, a series of filters wasapplied to each image before performing mitochondria and membranesegmentation. First, an “autodensity filter” was applied to increase thecontrast, followed by an “inverse video” of the images. An automaticthreshold corresponding to the histogram average was applied to gen-erate binary images, and a “median filter” was used to smooth therelevant information. The borders of mitochondria and the membranewere segmented from the filtered signal automatically. For the mito-chondrial segmentation, objects smaller than a perimeter of 20 pixelswere not taken into account. The border coordinates were exported inan Excel file (points x, y), and the impact points between the membraneand the mitochondria borders were calculated. The impact points cor-responded to a distance of 0 nm (superimposed pixels) between mito-chondrial and membrane borders or �200 nm (neighboring pixels) tomatch the optical resolution of the confocal images.

Cytosolic Ca2� Measurements—Experiments were performed inHepes-buffered solution containing (in mM): 140 NaCl, 5 KCl, 1 MgCl2,2 CaCl2, 20 Hepes, 10 glucose, pH 7.4 with NaOH. Ca2�-free solutioncontained 1 mM EGTA instead of CaCl2, except for the Mn2� quenchexperiments, where no EGTA was added. Glass coverslips weremounted in a thermostatic chamber (Harvard Apparatus, Holliston,MA) equipped with gravity feed inlets and vacuum outlet for solutionchanges. Cells were imaged on a Axiovert s100 TV using a �100, 1.3 NAoil-immersion objective (Carl Zeiss AG, Feldbach, Switzerland). HeLacells were loaded for 30 min with 2 �M fura-2/AM at room temperaturein the dark, washed twice, and equilibrated for 15–20 min to allowde-esterification. To monitor [Ca2�]cyt, cells were alternatively excitedat 340 and 380 nm with a monochromator (DeltaRam; Photon Technol-ogy International Inc., Monmouth Junction, NJ) through a 430 DCLPdichroic mirror. Emission was monitored through a 510WB40 filter(Omega Optical, Brattleboro, VT). Prior to the experiments, cells wereloaded with 500 nM Mitotracker Red for 90 s and washed two to threetimes with experimental buffer. Transfected cells were recognized bythe fluorescence of the nuclear-targeted GFP (480 nm excitation, 535nm emission), and the characteristic morphology of mitochondria (frag-mented versus tubular) was verified by imaging the mitotracker label-ing (577 nm excitation, 590 nm emission). The fluorescence of thenuclear GFP was also observed following excitation at 380 nm duringthe measurement of [Ca2�]cyt, especially if the labeling was strong. Forthis reason, we only selected cells with a moderate nuclear fluorescence,and the region of the cytosol used to estimate [Ca2�]cyt did not includethe nucleus. Fluorescence emission was imaged using a cooled, 16-bitCCD back-illuminated frame transfer MicroMax camera (Princeton

Instruments, Ropper Scientific, Trenton, NJ). Image acquisition andanalysis were performed with the Metafluor 4.6 software (UniversalImaging, West Chester, PA).

Measurements of Store-operated Ca2� Entry—Mn2� (100 �M) wassubstituted for Ca2� to estimate the ion flux through store-operatedCa2� channels, according to the Mn2� quench technique. Cells wereexcited at 356–358 nm, which corresponded to the isosbestic point offura-2. The rate of fluorescence decrease reflects the rate at which Mn2�

enters the cells, and the slope during the first 1–2 min was used as anindicator for CCE activity.

Mitochondrial Membrane Potential (��m) Measurements—To moni-tor changes in ��m, cells were loaded for 20 min with 50 nM tetrameth-ylrhodamine methyl ester (TMRM) in Hepes-buffered solution, andexperiments were carried out in the same buffer. Cells were excited at545 nm and emission collected through an LP 590 long pass filter.Changes in ��m were expressed as R/Ro, where R is the ratio of thefluorescence in the mitochondria divided by the cytosolic fluorescence ata given time and Ro is the initial ratio of the mitochondrial overcytosolic fluorescence.

Mitochondrial Ca2� and pH Measurements—We took advantage ofthe properties of RP3.1mit, whose fluorescence is Ca2�-sensitive whenexcited at 410 nm and pH-sensitive when excited at 480 nm (Refs. 14,15, see also Fig. 3). The cells were excited alternatively at 410 and 480nm, and emission was collected at 535 nm (535RDF45; Omega Optical)through a 505DCXR (Omega Optical) dichroic mirror. Changes in pHwere expressed as F/Fo, where F is the fluorescence (480 nm excitation)at a given time and Fo is the mean fluorescence of 5–10 individualmeasurements collected at the beginning of the recording. Changes inmitochondrial Ca2� are shown as 1-F/Fo, because RP3.1mit fluorescenceat �exc � 410 nm decreases with increasing Ca2� concentrations.

RESULTS

Effects of hFis1 Expression on Mitochondria Morphology andLocation—Mitochondria in HeLa cells form a largely intercon-nected network constantly remodeled by fusion and fissionreactions. To disrupt this balance toward fission, we overex-pressed hFis1, the human orthologue of the yeast protein Fis1pknown to participate in mitochondrial division. As shown inFig. 1A, expression of hFis1 rapidly fragmented the mitochon-drial network into punctuate organelles that clustered aroundthe nucleus. The fragmentation process occurred immediatelyupon hFis1 expression and, once initiated, was complete within4 h as documented by time-lapse video microscopy of cellsco-transfected with a nuclear-targeted GFP (See supplemen-tary movie S1). Within 1 h, all the cells expressing the hFis1cDNA had a punctiform mitochondria phenotype, consistentwith a previous study (13). To quantify the extent of mitochon-drial redistribution, we took several confocal optical sections ofmitochondria (labeled with Mitotracker Red) and of the cellcytosol (labeled with the cytosolic protein ratiometric pericam).Using the mitochondrial image as a mask, we determined thesurface of the cell occupied by mitochondria on the cytosolicimages (the nucleus was included in the cell surface, see “Ex-perimental Procedures”). As shown in Fig. 1B, mitochondriaspread out to the periphery and covered a larger area of thecytosol in control cells. On average, the percentage of the cellsurface “lacking” mitochondria was �2-fold larger in hFis1-expressing cells compared with control (Fig. 1C, n � 15 and 32cells, respectively). The same images were used to measure thenumber of contact points between mitochondria and the cellmembrane, defined as the outline of the cytosolic staining. Asshown in Fig. 1D, �18% of the cell membrane was apposed tomitochondria in control cells (at the resolution of our confocalsystem of �250 nm), a proportion that was reduced by �3 timesupon hFis1 expression. Thus, hFis1 not only induces mitochon-drial fragmentation but also redistributes mitochondria awayfrom the plasma membrane, leaving large regions of the cellperiphery devoid of mitochondria and fewer contacts betweenmitochondria and the cell surface.

Effects of hFis1 Expression on Mitochondrial Membrane Po-tential, pH, and Ca2� Homeostasis—To assess the effects of

Mitochondrial Fragmentation and Ca2� Signaling 22705

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from

hFis1 expression on mitochondrial function, we first testedwhether the mitochondrial membrane potential (��m) was al-tered after fragmentation. For this purpose, cells were loadedwith TMRM and challenged sequentially with 5 �g/ml oligo-mycin (to prevent ATP consumption) and 25 �M rotenone (toblock the complex I of the respiratory chain). As shown in Fig.2, the drugs elicited similar changes in ��m in hFis1-overex-pressing and control cells. In both cases, neither the application

of oligomycin nor rotenone alone had a significant effect on��m, whereas their combined application dissipated ��m to asimilar extent (Fig. 2B). The subsequent addition of the prot-onophore CCCP induced a rapid and complete dissipation ofthe membrane potential. The lack of depolarization in thepresence of oligomycin indicates that the respiratory chain wasfunctional in hFis1 cells and that the mitochondrial membranepotential was not maintained by the mitochondrial H� ATPase

FIG. 1. Effect of hFis1 expression onthe architecture of the mitochondrialnetwork. A, HeLa cells stably expressingRP3.1mit were transfected with hFis1 anda nuclear GFP. Typical pattern is shownof tubular mitochondria in untransfectedcells (right and left) and a fragmentedmitochondria hFis1-overexpressing cell(middle) imaged at 488/535 nm. The cellwith fragmented mitochondria also ex-pressed the nuclear GFP. Images areshadow projection of 22 adjacent, 200-nmwide z sections deconvoluted with the it-erative constrained Tikhonov-Miller res-toration algorithm. Scale bar, 10 �m. B,HeLa cells stably expressing the cytosolicprotein ratiometric pericam were loadedwith Mitotracker Red and imaged at 488/535 nm (left images) and 514/580 nm(middle images) on a spinning wheel con-focal system. Images of a control (top) andof a hFis1-overexpressing cell (bottom)are shown with the perimeter of the cyto-solic and mitochondrial staining outlinedin green and red, respectively. The regionsdefined by these borders are superim-posed in the right panels, with the cytoso-lic surface devoid of mitochondria repre-sented in gray. Scale bars, 10 �m. C,surface of the cytosol devoid of mitochon-dria as a percentage of the total cytosolicsurface. Statistics were performed on 150optical sections from 32 untransfectedcells and 70 of 15 hFis1-overexpressingcells, using the “mask” approach illus-trated in Fig. 1B. D, percentage of the cellperimeter “in contact” with the mitochon-dria (n � 120 sections from 25 untrans-fected cells and 69 of 14 hFis1-overex-pressing cells). Bars are mean S.E. *, p�0.05.

Mitochondrial Fragmentation and Ca2� Signaling22706

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from

functioning in reverse mode and consuming glycolytic ATP.Next, we measured the H� and Ca2� activities inside the

mitochondrial matrix of live HeLa cells. For these measure-ments, we took advantage of the dual sensitivity to both Ca2�

and pH of a ratiometric pericam probe targeted to mitochon-dria, RP3.1mit (kindly provided by Dr. A. Miyawaki, Tokyo).RP3.1mit fluorescence is highly sensitive to pH at an excitationof 480 nm, but not at 410 nm (14, 15). Conversely, RP3.1mit

fluorescence decreases with increasing concentrations of Ca2�

at 410 nm but is largely insensitive to Ca2� at 480 nm. Wecould verify this dual sensitivity by exposing RP3.1mit-labeledHeLa cells to the mitochondrial uncoupler CCCP or to thecalcium-mobilizing agonist histamine. As expected, CCCPcaused a selective drop in RP3.1mit fluorescence at 480 nm asthe mitochondria acidified to equilibrate its pH with the pH ofthe cytosol (Fig. 3A). In contrast, addition of histamine pro-duced a transient drop only at 410 nm (Fig. 3B), confirmingthat changes in mitochondrial Ca2� concentration, [Ca2�]mit,could be monitored selectively at this wavelength. We used thisapproach to evaluate the effect of hFis1 expression on mito-chondrial Ca2� and pH homeostasis. As shown in Fig. 3, C andE, addition of 1 �M CCCP elicited a drop in fluorescence at 480nm that was of similar magnitude in control and in hFis1-expressing cells. The drop in fluorescence corresponded to asimilar �pH, because the RP3.1mit calibration curves weresimilar in hFis1-overexpressing and -untransfected cells in thepH range 7.4–8.4 (see Supplementary Fig. 2). These data con-firm the TMRM measurements and indicate that the pH of themitochondrial matrix was not altered by hFis1 expression.

In concurrent Ca2� measurements at 410 nm, addition of 50�M histamine evoked identical [Ca2�]mit transients regardlessof the induction of hFis1 (Fig. 3D). In the absence of extracel-lular Ca2�, neither the maximal [Ca2�]mit elevation nor theduration of the response was different in control and hFis1-overexpressing cells (Fig. 3F). The [Ca2�]mit transients meas-ured in the presence of extracellular Ca2� were also not signif-icantly different; the maximal amplitude of the signal averaged0.167 0.015 (n � 33) in untransfected cells and 0.141 0.015(n � 13) in hFis1 overexpressors. Taken together, these dataindicate that mitochondrial pH and Ca2� homeostasis is notaffected by hFis1 expression despite the complete fragmenta-tion of the mitochondrial network.

Effects of hFis1 Expression on ER Structure—The presence ofnormal [Ca2�]mit transients in HeLa cells with fragmented andclustered mitochondria is surprising, because the proximity ofmitochondria to the Ca2� source (i.e. the inositol 1,4,5-trisphos-phate Ca2� release channels) was shown to be crucial for aproper mitochondrial Ca2� uptake. Mitochondria have beenproposed to form stable, long-term interactions with ER Ca2�

release channels to account for the efficient transfer of Ca2�

between the two organelles (15). Because hFis1 induced dra-matic alterations in mitochondrial architecture, we investi-gated whether the ER was also affected. As shown in Fig. 4, thestaining pattern of the ER-targeted yellow cameleon probe(YC4.1ER) was not grossly altered upon hFis1 expression, indi-cating that the mitochondrial remodeling was not accompaniedby visible changes in ER architecture.

Effects of hFis1 on Ca2� Transfer from Plasma MembraneChannels to Mitochondria—hFis1-fragmented mitochondriaappear to handle normally the Ca2� released from the ER, asindicated by the normal [Ca2�]mit transient elicited by hista-mine. Because the main effect of hFis1, apart from fragmenta-tion, is to move mitochondria away from the plasma membrane(Fig. 1), we assessed whether hFis1 altered the ability of mito-chondria to take up Ca2� originating from the plasma mem-brane. For this purpose, Ca2� was readmitted to cells previ-ously stimulated with 50 �M histamine in the nominal absenceof Ca2�. As shown in Fig. 5A, the amplitudes of the [Ca2�]mit

elevations were similar in control (0.109 0.019; n � 12) andin hFis1-overexpressing cells (0.110 0.014; n � 15). Interest-ingly however, the time needed to reach this level was signifi-cantly prolonged by hFis1 expression. The Ca2� enteringacross the plasma membrane took, on average, 31 s longer tocause a maximal response in hFis1-fragmented mitochondria.Similar results were obtained in cells stimulated with theSERCA pump inhibitor, thapsigargin (1 �M), instead of hista-mine (Fig. 5, C and D), indicating that ER Ca2� pumps werenot involved in the transfer of Ca2� from the plasma membraneto mitochondria.

To understand the structural basis of this slower [Ca2�]mit

increase, we analyzed the spatio-temporal pattern of the[Ca2�]mit signal during Ca2� readdition to control and hFis1cells. As shown in Fig. 6, [Ca2�]mit increased rapidly in large,contiguous regions of control tubular mitochondria. In contrast,[Ca2�]mit increased sequentially in discrete regions of hFis1-fragmented mitochondria. The slower response of hFis1-frag-mented mitochondria was not because of a delay in the transferof Ca2� from the plasma membrane to mitochondria, becausethe [Ca2�]mit signal initiated at the same time or even earlier inindividual mitochondria from hFis1-transfected cells (Fig. 6A).Rather, Ca2� spread faster and in a more coordinated mannerwithin tubular mitochondria (Fig. 6B), indicating that thepropagation of the [Ca2�]mit signal was impaired by the frag-mentation of the mitochondrial network.

The delayed [Ca2�]mit increase in hFis1 cells might possibly

FIG. 2. ��m is not altered by hFis1 overexpression. Cells wereloaded with 50 nM TMRM, and the ratio of the mitochondrial overcytosolic TMRM fluorescence (Fmito/Fcyto) was measured by fluorescenceimaging (10). A, changes in Fmito/Fcyto following successive applicationof oligomycin, rotenone, and CCCP. Ratio values are normalized to theinitial Fmito/Fcyto value. B, changes in TMRM Fmito/Fcyto ratio fluores-cence induced by oligomycin and rotenone, applied alone and in combi-nation. Bars are mean S.E. (n � 9–18 for hFis1-overexpressing cellsand 9–21 for untransfected cells).

Mitochondrial Fragmentation and Ca2� Signaling 22707

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from

reflect a reduced or slower influx of Ca2� across the plasmamembrane. To test this possibility, we measured Ca2� influxwith fura-2. As shown in Fig. 7, the cytosolic Ca2� changesupon Ca2� readdition to cells stimulated with thapsigarginwere of similar amplitude and kinetics in control and hFis1-overexpressing cells. To confirm this observation, CCE activitywas measured by following the rate of Mn2� influx (Fig. 7C).The rates of Mn2� quench were not significantly different incontrol and hFis1-expressing cells (Fig. 7D), indicating thatCCE was largely unaffected by the fragmentation and subcel-lular redistribution of the mitochondrial network.

Effects of hFis1 on CCE Modulation by Mitochondria—Func-tional mitochondria are required to sustain CCE, but it is not

clear whether mitochondria act locally, i.e. as Ca2� buffers thatremove Ca2�-dependent channel inhibition, or globally, i.e. bymodulating the filling state of the ER or by releasing a diffus-ible messenger. Because mitochondria in hFis1 cells wereclearly located farther away from the plasma membrane thanin untransfected cells, they provided a convenient model toseparate the local and global effects of mitochondria on CCE.For this purpose, cells were stimulated with thapsigargin toactivate CCE and mitochondria function was inhibited by ei-ther 1 �M CCCP or by a combination of 25 �M rotenone and 5�g/ml oligomycin. The effects of the mitochondria inhibitors onCCE were then assessed by the Ca2� readdition protocol or bythe Mn2� quench technique. As shown in Fig. 8, Ca2� entry

FIG. 3. Mitochondrial pH and Ca2� uptake is not altered by the fragmentation of the mitochondrial network. HeLa cells weretransiently transfected with the RP3.1mit, and the pH and Ca2� sensitivity of the probe are demonstrated in panels A and B. A, 1 �M CCCP wasapplied to acidify the mitochondrial matrix. The change in pH is reflected by a drop of the fluorescence following excitation at 480 nm. The 410-nmwavelength showed almost no change. B, the cell was stimulated with histamine that induced a mitochondrial Ca2� uptake, which is reflected bya decrease of the 410-nm excitation wavelength, whereas the 480 nm only marginally changed. C–F, HeLa cells were transfected with the RP3.1mitand 24 h later with hFis1. C, the application of 1 �M CCCP, which dissipated the H� gradient across the mitochondrial matrix, produced a similarchange of the fluorescence recorded after excitation at 480 nm. E, statistical evaluation of the effect of CCCP. Bars are mean S.E. (n � 12 forthe untransfected cells and 6 for the hFis1-overexpressing cells). D, representative recording of the mitochondrial Ca2� increase followingstimulation with 50 �M histamine in Ca2�-free medium. F, statistics of the amplitude of Ca2� increase induced by histamine. Bars are mean S.E.(n � 33 for the untransfected cells and 13 for the hFis1-overexpressing cells).

Mitochondrial Fragmentation and Ca2� Signaling22708

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from

was reduced by about one third in the presence of CCCP or ofoligomycin/rotenone, regardless of hFis1 expression. Mn2� en-try was reduced to a similar extent in the presence of 1 �M

CCCP, both in control (from 8.503 1.442, n � 18 to 4.630 0.464, n � 25; p �0.05) and in hFis1-overexpressing cells (from6.876 1.196, n � 14 to 3.276 0.225, n � 10; p �0.05). Theseresults indicate that functional mitochondria are required foroptimal activation of CCE in HeLa cells, although the modula-tion of CCE by mitochondria (30–40%) is less pronounced thanin other cell types.

DISCUSSION

In this study we investigated the effect of a controlled dis-ruption of the mitochondrial network on the Ca2� homeostasisof mitochondria. For this purpose, we expressed the proteinhFis1 in HeLa cells to induce a rapid fragmentation and pe-rinuclear clustering of their mitochondria. Surprisingly, thesedramatic morphological alterations had little impact on theorganelle function because mitochondria were still able tomaintain a normal membrane potential and pH and to take upand release Ca2�. This experimental paradigm allowed us tostudy Ca2� handling by mitochondria located close or far from

the plasma membrane, to define the role of mitochondria in-terconnection in the propagation of Ca2� signals, and to assessthe local and global effects of mitochondria on plasma mem-brane Ca2� channels.

hFis1 Initiates Mitochondria Fragmentation without Alter-ing the Function of the Organelle—Mitochondria are dynamicorganelles that often form an extensive tubular network re-flecting the balance of ongoing fusion and fission processes.Among the proteins regulating the fusion and fission processes,hFis1 was recently shown to induce mitochondrial fission inmammalian cells by recruiting the dynamin-related GTPasesDrp1 from the cytosol to the outer mitochondrial membrane(13, 16). We could confirm that expression of hFis1 in HeLacells induces a complete fragmentation of mitochondria within16–20 h. This effect was selective for mitochondria becauseexpression of hFis1 did not modify the ER architecture. Uponfragmentation, the mitochondria clustered around the nucleus,leaving large parts of the cytosol devoid of these organelles.Morphometric analysis revealed that 45% of the cellular areawas lacking mitochondria in cells expressing hFis1 comparedwith 25% in control cells and that the fragmented mitochondria

FIG. 4. hFis1 overexpression doesnot alter the organization of the ER.Typical fluorescence of intact HeLa cellstransfected with the ER-targeted yellowcameleon probe (YC4.1ER; left images)and labeled with the Mitotracker Red(right images). Images are shadow projec-tion of 14 and 16 (top and bottom, respec-tively) adjacent, 200-nm wide z sectionsdeconvoluted with the iterative con-strained Tikhonov-Miller restoration al-gorithm. Scale bars are 10 �m (top) and 5�m (bottom).

Mitochondrial Fragmentation and Ca2� Signaling 22709

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from

were located farther away from the plasma membrane. Wecould not precisely evaluate the location of mitochondria rela-tive to the upper portion of the plasma membrane (i.e. the “roof”of the cell) given the limited optical resolution of our confocalsystem in the z-axis. Nevertheless, functional data (see below)strongly suggest that upon hFis1 expression mitochondriamoved toward deeper regions of the cytosol.

hFis1 overexpression did not affect the mitochondrial mem-brane potential (��m) as measured in situ with low concentra-tions of the rhodamine probe TMRM. Furthermore, oligomycinand rotenone, alone or in combination, had similar effects on��m. This latter experiment rules out the possibility that frag-mented mitochondria maintained a normal membrane poten-tial by the hydrolysis of glycolytic ATP, because the ATP syn-thase inhibitor oligomycin did not dissipate ��m. Consistentwith the maintenance of a normal H� electrochemical gradient,the pH of the mitochondrial matrix was not altered 16 h aftertransfection of hFis1 (Fig. 3). We did not try later time points,because expression of hFis1 for �48 h has been shown to inducecytochrome c release and apoptosis (13). Thus, during acutefragmentation mitochondria maintained a normal membranepotential and pH, in agreement with a recent report on hip-

pocampal neurons showing that mitochondrial redistributionand aggregation did not modify their energy status (17).

Mitochondria Fragmentation and Ca2� Handling—Giventhe preserved function but altered shape and location of mito-chondria, the question arises as to whether they were still ableto accumulate Ca2� during cell stimulation. This question is ofparticular interest because the mitochondrial Ca2� uniporter,which drives the entry of Ca2� into mitochondria, has a lowaffinity for Ca2� (18). It was thus postulated that mitochondriamust be located in close vicinity to Ca2� release sites on the ER(i.e. near inositol 1,4,5-trisphosphate-gated Ca2� release chan-nels) to rapidly and efficiently accumulate Ca2� (2, 3). This“high microdomain” model is widely accepted to account for therapid [Ca2�]mit increase that occurs during cell stimulation. Asan extension of this postulate, it was proposed recently that thecontact points between the ER and mitochondria are highlystable over time, suggesting that specific structural interac-tions exist between the two organelles (15). This conclusion wasbased on 1) the similar distribution of highly responsive mito-chondria inside cells during successive histamine stimulations,and 2) the larger than expected reduction in aequorin re-sponses to repetitive histamine challenges, which indicates

FIG. 5. hFis1 overexpression delays the transfer of Ca2� from the plasma membrane to mitochondria. A and C, after stimulation with50 �M histamine (A) or 1 �M thapsigargin (C) in Ca2�-free medium, 2 mM Ca2� was readded and the Ca2� uptake by mitochondria was evaluated.The fragmented mitochondria needed significantly more time to accumulate Ca2� maximally. B and D, statistical evaluation of the delayedmitochondrial Ca2� uptake. Bars are mean S.E. (n � 11 for untransfected cells and 15 for hFis1-overexpressing cells in panel B; n � 18 foruntransfected cells and 10 for hFis1-overexpressing cells in panel D. *, p �0.05).

Mitochondrial Fragmentation and Ca2� Signaling22710

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from

FIG. 6. Disruption of the mitochondrial network increases the heterogeneity of the [Ca2�]mit signal. Cells were transiently transfectedwith RP3.1mit and [Ca2�]mit measured during Ca2� readdition by ratio imaging. A threshold corresponding to 80% of the spatially averaged[Ca2�]mit response was applied to delineate regions of high [Ca2�]mit. A, fluorescence images taken every 15 s during the Ca2� readdition are shownwith pixels exceeding the 80% threshold highlighted in red. Bar, 10 �m, applies to all images. B, the percentage of the mitochondrial area exceedingthe threshold is plotted over time. Arrow indicates start of the Ca2� readdition.

Mitochondrial Fragmentation and Ca2� Signaling 22711

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from

that mitochondria that capture large amounts of Ca2� (andthus consume aequorin) are not replaced by other mitochondriafrom the remaining population. In our imaging measurements,we did not detect any significant differences in the ability offragmented or tubular mitochondria to take up Ca2� duringhistamine stimulation, neither in the amplitude nor in thekinetic of the response. Although this result is not contradic-tory to the concept of a close vicinity between certain parts ofthe ER and mitochondria, it is hard to reconcile with theexistence of permanent physical contacts. Our data showedthat the ER structure was not affected by the fragmentation ofthe mitochondrial network, indicating that hFis1 specificallyaltered the structure and location of one organelle (the mito-chondria) while leaving the other intact (the ER). Although it isconceivable that mitochondria can move and reform specificinteractions with other ER Ca2� channels at a new location, itis difficult to envisage that stable ER-mitochondria complexescan move inside cells without altering the ER structure. Thus,to account for the rapid uptake of Ca2� into fragmented mito-chondria, the most likely hypothesis is that close contacts be-

tween the ER and mitochondria occur stochastically but at arelatively high frequency given the density of the two or-ganelles in perinuclear regions.

Although fragmented mitochondria captured normally theCa2� released from the ER, they accumulated Ca2� with asignificant delay compared with tubular mitochondria whenthe Ca2� source was the extracellular space. The kinetic wasabout two times slower, whereas the maximal Ca2� increasewas not affected. Because morphometric analysis indicatedthat fragmented mitochondria are located deeper in the cytosol,Ca2� ions must, on average, travel a longer distance beforereaching a fragmented than a tubular mitochondria. However,this longer distance is unlikely to account for the delayedtransfer of Ca2� to fragmented mitochondria, because duringfura-2 measurements Ca2� equilibrated within seconds in thecytosol. The increased distance from the plasma membranemight, however, translate into a slightly reduced Ca2� con-centration around perinuclear mitochondria. In this case,Ca2� would enter at a lower rate through the mitochondrial Ca2�

uniporter without altering its capacity to accumulate Ca2�. Re-

FIG. 7. Impact of mitochondrial fragmentation on CCE. HeLa cells were loaded with 2 �M fura-2 to measure cytosolic Ca2� changes. Aftertreating the cells with 1 �M thapsigargin in Ca2�-free, the CCE was evaluated either by adding 2 mM Ca2� (A) or by adding 100 �M Mn2� to quenchthe fura-2 fluorescence (C). B, maximal Ca2� increase after Ca2� readdition in untransfected cells (n � 39) and hFis1-overexpressing cells (n � 15).Bars are mean S.E. D, statistics of the Mn2� quench, represented as the initial slope of the fluorescence decrease in untransfected cells (n � 18)and hFis1-overexpressing cells (n � 14). Bars are mean S.E. In both cases the CCE was smaller in hFis1-overexpresing cells, although thereduction did not reach significance.

Mitochondrial Fragmentation and Ca2� Signaling22712

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from

gardless of the mechanism, the ability of perinuclear mitochon-dria to take up Ca2�, albeit at slower rates, indicated that closecontacts to plasma membrane channels were not required forCa2� uptake by mitochondria during capacitative Ca2� entry.Moreover, our observations indicated that the transfer of Ca2�

from the extracellular space to mitochondria did not involve theER, because mitochondria located deep in the cell were stillable to take up Ca2� when ER SERCA ATPases were inhibitedby thapsigargin (Fig. 5). These data indicate that Ca2� does nottransit through the ER to reach mitochondria and that highCa2� microdomains are not required for the slow uptake ofCa2� that occurs during CCE.

Our observations also indicated that the formation of a tu-bular network facilitates the propagation of Ca2� along mito-chondria. As shown in Fig. 6, Ca2� uptake was not only slowerduring Ca2� readdition but also more heterogeneous in frag-mented than in tubular mitochondria. In fragmented mitochon-dria, Ca2� increased sequentially in small regions that accu-mulated Ca2� in an uncoordinated manner. In contrast, Ca2�

increased in an explosive manner within large, contiguous re-gions of the tubular network. This indicates that the Ca2�

entering mitochondria can tunnel within the matrix and equil-ibrate rapidly along tubular, but not fragmented, mitochon-dria. Lumenal connectivity between mitochondria has beentested in HeLa cells using the fluorescence recovery after pho-tobleaching (FRAP) technique (10). The fluorescence of DsRed

recovered within 90 s after bleaching when a long (�25 �m),filamentous mitochondrial cluster was irradiated but failed torecover for up to 1 h when larger regions of the cell werebleached (10). This indicates that small molecules such as Ca2�

can tunnel within the matrix of fused mitochondria but notbetween mitochondrial clusters that are not interconnected.Our observations are consistent with these findings, becauseduring Ca2� readdition [Ca2�]mit did not increase at the sametime in all mitochondria, even in cells with an extensive tubu-lar network. The response was relatively homogenous, how-ever, because the [Ca2�]mit signal initiated simultaneously atseveral sites at the cell periphery and then rapidly propagatedtoward deeper cellular regions in a synchronous manner. Sucha coordinated [Ca2�]mit increase was not observed in cells withfragmented mitochondria, indicating that the tubular organi-zation of mitochondria facilitates Ca2� propagation betweenindividual organelles.

Local and Global Effects of Mitochondria on Plasma Mem-brane Ca2� Channels—Because mitochondria remainedlargely functional upon hFis1 expression, we investigatedwhether these fragmented mitochondria located far from theplasma membrane were still able to modulate CCE. Depletionof the ER Ca2� stores triggers an influx of extracellular Ca2�

called CCE (4). Several studies have shown that mitochondriaare involved in the maintenance and/or activation of CCE indifferent cells, but their exact contribution as well as the mech-

FIG. 8. Functional mitochondria are required for optimal Ca2� entry regardless of hFis1 expression. A and B, effect of mitochondrialdepolarization due to 1 �M CCCP on the Ca2� entry following thapsigargin treatment in untransfected cells (A) and hFis1-overexpressing cells (B).C, statistical evaluation of the effect of CCCP on Ca2� entry. Bars are mean S.E. (n � 13 for untransfected cells and 6 for hFis1-overexpressingcells). *, p �0.05 versus control. D, similar experiments were performed with a mixture of 25 �M rotenone and 5 �g/ml oligomycin to poisonmitochondria. Bars are mean S.E. (n � 25 for untransfected cells and 20 for hFis1-overexpressing cells). *, p �0.05 versus control.

Mitochondrial Fragmentation and Ca2� Signaling 22713

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from

anism of action are still debated. In electrophysiological stud-ies, the best characterized current supporting CCE is ICRAC, acurrent that is highly selective for Ca2� and carried by chan-nels of small unitary conductance (19–21). This current isinhibited at high intracellular Ca2� concentrations, and mito-chondria were shown to prevent Ca2�-dependent channel inac-tivation by their ability to accumulate Ca2� near the cytosolicmouth of the channel (6). In agreement with this hypothesis, arecent cell-attached patch-clamp study on endothelial cellsshowed that mitochondria can maintain low Ca2� concentra-tions under single plasma membrane channels (9). In additionto their buffering effects, mitochondria were recently proposedto release an as yet unidentified diffusible factor(s) that regu-lates the activity of ICRAC (22).

Our experiments using either thapsigargin or histamine todeplete ER Ca2� stores showed that upon Ca2� readdition thebulk cytosolic Ca2� elevation was not significantly different incontrol or hFis1-overexpressing cells. This was also confirmedby Mn2� quench experiments, although with both assays asmall but not significant reduction in Ca2�/Mn2� entry wasobserved in hFis1-overexpressing cells. Because in hFis1 cellslarge parts of the plasma membrane are devoid of underlyingmitochondria, the presence of mitochondria near membranechannels is clearly not essential for CCE. This does not implythat mitochondria do not exert local effects on CCE, becausereduced subplasmalemmal Ca2� buffering could have oppositeeffects on membrane channels. Ca2�-dependent K� channelsare more active when located far from underlying mitochondria(9), leading to a larger hyperpolarization and enhanced drivingforce for Ca2� entry. Because Ca2�-dependent K� channels arealso present in HeLa cells, (23), lack of subplasmalemmal mi-tochondria might elicit opposite mechanisms, the reduction inlocal Ca2� buffering enhancing the Ca2� feedback inhibition onCa2� entry channels while increasing the driving force for Ca2�

entry. Thus, a local role of mitochondria on Ca2� entry chan-nels cannot be formally excluded but is apparently not thedominant mechanism by which mitochondria modulate CCE inHeLa cells.

Regardless of their location, functional mitochondria wererequired for optimal CCE, because poisoning mitochondria ei-ther with CCCP or oligomycin/rotenone significantly reducedCCE both in control and hFis1-overexpressing cells. It shouldbe noted, however, that in other cellular systems such as RBL-1(22), T lymphocytes (7), or endothelial cells (9) such inhibitionof mitochondrial function resulted in a more pronounced reduc-tion of CCE (�80–90%). Thus, the proportion of CCE under the

influence of mitochondria is relatively modest in HeLa cells(�30–40%). Nonetheless, the observation that optimal CCEactivity requires functional mitochondria even if the organellesare located far from the plasma membrane suggests that CCEmodulation is not a local effect but rather a global effect thatmight involve a diffusible factor, as proposed recently (22).Thus, function, but not location, of mitochondria is critical forCCE modulation.

In conclusion, our data show that mitochondrial fragmenta-tion and perinuclear clustering did not alter the ability ofmitochondria to take up Ca2� ions released by the ER or tomodulate CCE but significantly decreased the speed of Ca2�

propagation between these organelles during Ca2� influx.Thus, mitochondria remain functional and able to modulateCCE regardless of their shape and location, although the for-mation of a mitochondria network might facilitate the propa-gation of specific Ca2� signals within cells.

Acknowledgment—We thank Dr. Wolgang Graier for critical readingof the manuscript.

REFERENCES

1. Duchen, M. R. (1999) J. Physiol. 516(Pt 1), 1–172. Rizzuto, R., Brini, M., Murgia, M., and Pozzan, T. (1993) Science 262, 744–7473. Rizzuto, R., Pinton, P., Carrington, W., Fay, F. S., Fogarty, K. E., Lifshitz,

L. M., Tuft, R. A., and Pozzan, T. (1998) Science 280, 1763–17664. Putney, J. J. (1986) Cell Calcium 7, 1–125. Putney, J. J. (1990) Cell Calcium 11, 611–6246. Hoth, M., Fanger, C. M., and Lewis, R. S. (1997) J. Cell Biol. 137, 633–6487. Hoth, M., Button, D. C., and Lewis, R. S. (2000) Proc. Natl. Acad. Sci. U. S. A.

97, 10607–106128. Parekh, A. B. (1998) J. Biol. Chem. 273, 14925–149329. Malli, R., Frieden, M., Osibow, K., and Graier, W. F. (2003) J. Biol. Chem. 278,

10807–1081510. Collins, T. J., Berridge, M. J., Lipp, P., and Bootman, M. D. (2002) EMBO J.

21, 1616–162711. Shaw, J. M., and Nunnari, J. (2002) Trends Cell Biol. 12, 178–18412. Mozdy, A. D., McCaffery, J. M., and Shaw, J. M. (2000) J. Cell Biol. 151,

367–38013. James, D. I., Parone, P. A., Mattenberger, Y., and Martinou, J. C. (2003)

J. Biol. Chem. 278, 36373–3637914. Nagai, T., Sawano, A., Park, E. S., and Miyawaki, A. (2001) Proc. Natl. Acad.

Sci. U. S. A. 98, 3197–320215. Filippin, L., Magalhaes, P. J., Di Benedetto, G., Colella, M., and Pozzan, T.

(2003) J. Biol. Chem. 278, 39224–3923416. Yoon, Y., Krueger, E. W., Oswald, B. J., and McNiven, M. A. (2003) Mol. Cell.

Biol. 23, 5409–542017. Wang, G. J., Jackson, J. G., and Thayer, S. A. (2003) J. Neurochem. 87, 85–9418. Rizzuto, R., Bernardi, P., and Pozzan, T. (2000) J. Physiol. 529 (Pt. 1), 37–4719. Hoth, M., and Penner, R. (1992) Nature 355, 353–35620. Parekh, A. B., and Penner, R. (1997) Physiol. Rev. 77, 901–93021. Prakriya, M., and Lewis, R. S. (2003) Cell Calcium 33, 311–32122. Glitsch, M. D., Bakowski, D., and Parekh, A. B. (2002) EMBO J. 21,

6744–675423. Sauve, R., Simoneau, C., Parent, L., Monette, R., and Roy, G. (1987) J. Membr.

Biol. 96, 199–208

Mitochondrial Fragmentation and Ca2� Signaling22714

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from

and Nicolas DemaurexMaud Frieden, Dominic James, Cyril Castelbou, Anne Danckaert, Jean-Claude Martinou

Induced by hFis1 Homeostasis during Mitochondrial Fragmentation and Perinuclear Clustering2+Ca

doi: 10.1074/jbc.M312366200 originally published online March 15, 20042004, 279:22704-22714.J. Biol. Chem. 

  10.1074/jbc.M312366200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

Supplemental material:

  http://www.jbc.org/content/suppl/2004/03/25/M312366200.DC1

  http://www.jbc.org/content/279/21/22704.full.html#ref-list-1

This article cites 23 references, 13 of which can be accessed free at

by guest on March 22, 2020

http://ww

w.jbc.org/

Dow

nloaded from