cooperativeandstochasticcalciumreleasesfrommultiple ... › content › 287 › 29 ›...

Cooperative and Stochastic Calcium Releases from MultipleCalcium Puff Sites Generate Calcium Microdomains in IntactHeLa Cells*□S

Received for publication, October 7, 2011, and in revised form, April 8, 2012 Published, JBC Papers in Press, May 25, 2012, DOI 10.1074/jbc.M111.311399

Hideki Nakamura‡§¶, Hiroko Bannai‡, Takafumi Inoue¶�, Takayuki Michikawa‡�**1, Masaki Sano§,and Katsuhiko Mikoshiba‡�2

From the ‡Laboratory for Developmental Neurobiology, RIKEN, Brain Science Institute, 2-1 Hirosawa, Wako, Saitama 351-0198,Japan, the §Department of Physics, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan, the �CalciumOscillation Project, ICORP, Japan Science and Technology Corporation (JST), 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan,the ¶Department of Life Science and Medical Bioscience, Waseda University, 2-2 Wakamatsu-cho, Shinjuku-ku, Tokyo, 162-8480,Japan, and the **Brain Science Institute, Saitama University, 255 Shimo-Okubo, Sakura-ku, Saitama 338-8570, Japan

Background: Ca2� microdomains regulate many essential physiological events.Results: IP3-evoked Ca2� microdomains in histamine-stimulated intact HeLa cells were visualized using a total internal reflec-tion fluorescence microscope.Conclusion: Ca2� microdomains in intact cells are generated from spatially distributed multiple channel clusters, rather thana single tight cluster.Significance: The results provide a basic understanding of the spatiotemporal signal regulations of Ca2� microdomain forma-tion in living cells.

Ca2� microdomains or locally restricted Ca2� increases in thecell have recently been reported to regulatemany essential physio-logical events. Ca2� increases through the inositol 1,4,5-trisphos-phate (IP3) receptor (IP3R)/Ca2� releasechannels contribute to theformation of a class of suchCa2� microdomains, whichwere oftenobserved and referred to as Ca2� puffs in their isolated states. Inthis report, we visualized IP3-evokedCa2� microdomains in hista-mine-stimulated intact HeLa cells using a total internal reflectionfluorescencemicroscope, andquantitatively characterized the spa-tial profile by fitting recorded images to a two-dimensional Gauss-ian distribution. Ca2� concentration profiles weremarginally spa-tially anisotropic, with the size increasing linearly even after theamplitude began to decline. We found the event centroid driftedwith an apparent diffusion coefficient of 4.20� 0.50�m2/s, whichis significantly larger than those estimated for IP3Rs. The sites ofmaximal Ca2� increase, rather than initiation or termination sites,were detected repeatedly at the same location. These results indi-cate that Ca2� microdomains in intact HeLa cell are generatedfrom spatially distributedmultiple IP3R clusters orCa2�puff sites,rather than a single IP3R cluster reported in cells loadedwithCa2�

buffers.

The calcium ion is one of the most essential and universalsignalingmolecules inmany organisms, regulating awide rangeof physiological events, including development, gene expres-sion, and neuronal plasticity (1). The rich variety of both tem-poral and spatial dynamics of Ca2� signals facilitates the mul-tifunctionality of Ca2� (2, 3) because the vast amount ofinformation required to regulate various downstream signalsshould be encoded in the spatiotemporal profiles of the signals.Ca2� microdomains, which refer to Ca2� signaling eventslocalized to a specific part of the cell (4, 5), are an example ofsuch versatility of Ca2� signal dynamics. Ca2� microdomainshave recently attracted significant attention, as they have beenreported to regulate many essential biological processes, andthe molecular entities underlying the local signaling eventshave been elucidated (4).Several classes of Ca2� ion channels have been demonstrated

to form Ca2� microdomains (5); voltage-operated channels,ryanodine receptors and inositol 1,4,5-trisphosphate (IP3)3receptors (IP3Rs). These ion channels have distinct localizationpatterns in the cell, forming different classes of Ca2� microdo-mains with different biological roles. Voltage-operated chan-nel-dependent microdomains (Ca2� sparklets) function bytriggering excitation-contraction coupling in cardiac muscleand neurotransmitter release in neurons. Ryanodine receptor-dependent microdomains (Ca2� sparks) function in the regu-lation of excitation-contraction coupling, whereas IP3R-depen-dent microdomains (Ca2� blips, Ca2� puffs) are proposed toserve as an elementary building block of IP3R-dependent Ca2�

signaling, which is particularly well known for its versatile spa-tiotemporal patterns, includingCa2�waves andoscillations (2).

* This work was supported by grants from the Ministry of Education, Science,Sports and Culture of Japan 23650197 (to H. N.), 16770124 (to H. B.),14380364 (to T. I.), 20370054 and 24500476 (to T. M.) and 2022007 (toK. M.), and from JST ICORP-SORST (to K. M.).

□S This article contains supplemental Figs. 1–3.1 To whom correspondence may be addressed: Brain Science Institute,

Saitama University, 255 Shimo-Okubo, Sakura-ku, Saitama 338-8570,Japan. Tel.: 81-48-858-9278; Fax: 81-48-858-9278; E-mail: [email protected].

2 To whom correspondence may be addressed: Laboratory for Developmen-tal Neurobiology, RIKEN Brain Science Institute, 2-1 Hirosawa, Wako,Saitama 351-0198, Japan. Tel.: 81-48-467-9745; Fax: 81-48-467-9744;E-mail: [email protected].

3 The abbreviations used are: IP3, inositol 1,4,5-trisphosphate; inositol 1,4,5-trisphosphate receptor, IP3R; TIRFM, total internal reflection fluorescencemicroscope; NP-EGTA, o-nitrophenyl EGTA acetoxymethyl ester.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 287, NO. 29, pp. 24563–24572, July 13, 2012© 2012 by The American Society for Biochemistry and Molecular Biology, Inc. Published in the U.S.A.

JULY 13, 2012 • VOLUME 287 • NUMBER 29 JOURNAL OF BIOLOGICAL CHEMISTRY 24563

by guest on June 14, 2020http://w

ww

.jbc.org/D

ownloaded from

http://www.jbc.org/cgi/content/full/M111.311399/DC1

http://www.jbc.org/

Among the Ca2� ion channels underlying Ca2� microdo-mains, IP3Rs mediate IP3-induced Ca2� release from the ER ina wide range of cell types, regulating many essential biologicalphenomena, including gene expression, development, and neu-ronal plasticity (6). In parallel with the channel regulation by itsligand IP3, cytoplasmic Ca2� regulates the open probability ofIP3Rs in a biphasicmanner; Ca2� release is potentiated at lowerCa2� concentrations but inhibited at higher Ca2� concentra-tions (7). The stimulatory effect byCa2� suggests that the chan-nels display the process of Ca2�-induced release, which under-lies the versatile patterns of IP3-evoked Ca2� signals such asCa2� spike generation and wave propagation (3).The IP3R-dependent Ca2� signaling system exhibits a hier-

archical organization that ranges from Ca2� microdomains(Ca2� blips, Ca2� puffs) to global Ca2� rises that include Ca2�

waves or oscillations (2). IP3R-dependent Ca2� microdomainshave recently drawn considerable attention (4, 5) as theselocally restricted signals are implied to have significant physio-logical roles such as neuronal transmitter release (8, 9), cellmetabolism (10), and cell survival under stress (11). The richvariety of physiological functions is probably dependent on thevariety of both spatial and temporal profiles of Ca2� concentra-tions facilitated by the formation of microdomains (5). Conse-quently, it is essential to investigate the physiological roles ofthe IP3R-dependent Ca2� microdomains in intact cells, focus-ing on their spatial and temporal organizations.IP3-evoked Ca2� microdomains were first observed and

termed Ca2� puffs in Xenopus oocytes (12, 13), and then inHeLa cells (14), PC12 cells (15), and hippocampal neurons (16).Despite the large differences among cell types, e.g. origin of theorganism, expression profiles, and size or morphology, theevents observed in those diverse cell types often showed similaramplitudes (17), spatial widths (18), and durations (17), sug-gesting the phenomenon as an elementary building block ofIP3-evoked Ca2� signaling (14). Recent experiments using fastfluorescence imaging and a cytosolic Ca2� buffer EGTA, whichefficiently isolated each Ca2� puff, identified the number ofIP3R channels involved in an individual event (19), and con-cluded that each Ca2� puff is generated by a pre-establishedand practically immobile IP3R cluster (20), which is called theCa2� puff site. The term Ca2� puff originally referred to anylocalized Ca2� signal evoked by the opening of multiple IP3Rs,as explained above. However, recent identification of the puffsites as single IP3R clusters in their isolated states (i.e. withcytosolic EGTA) indicates a stricter redefining of the term asCa2� signaling caused by the opening of a single IP3R cluster.We therefore term the localized Ca2� increase in the cytosolinduced by the activation of IP3 production as an IP3-evokedCa2� microdomain.Although a vast amount of knowledge has been accumulated

on isolated Ca2� puffs in EGTA-loaded cells, the spatiotempo-ral characters of the IP3-evoked Ca2� microdomains in intactcells have not been determined. Provided that IP3R gating isregulated both by IP3 and biphasically by Ca2� as describedabove, buffering of the cytosolic Ca2� can cause a significanteffect on the gating of each IP3R involved in themicrodomain instudies involving EGTA-loaded cells. Moreover, elucidatingthe spatiotemporal profiles of intact IP3-evokedCa2�microdo-

mains should contribute to the understanding of the physiolog-ical function of the microdomain.To address the spatial and temporal Ca2� profiles of IP3-

evoked Ca2� microdomains, we visualized IP3-evoked Ca2�

microdomains inHeLa cell using a total internal reflection fluo-rescence microscope (TIRFM) (21) and the fluorescent Ca2�

indicator fluo-4 to monitor the two-dimensional spatial profileof the local Ca2� signals. Ca2� microdomains in the basalregion of the cytoplasm were visualized using a relatively highsampling rate (20 ms/frame), consistent with previous reports(22). Such an approach enabled the analysis of the two-dimen-sional spatial profile of the Ca2� concentrations generated byCa2� microdomains. The underlying organization of Ca2� puffsites and the mechanism of cooperative activation of the chan-nels are discussed based on these results. These lead to theconclusion that each Ca2� microdomain involves multipleCa2� release sites or Ca2� puff sites that are widely distributed,rather than a single tight IP3R cluster observed in isolated Ca2�

puff studies with exogenous cytosolic Ca2� buffers (20).

EXPERIMENTAL PROCEDURES

Ca2� Imaging—HeLa cells were grown in DMEM (NacalaiTesque) supplemented with 10% fetal bovine serum and a 1%penicillin-streptomycin solution (Nacalai Tesque). Approxi-mately 3.0 � 104 cells were transferred to a 35-mm glass-bot-tomed dish (Mattek) �48 h before imaging. Cells were loadedby incubation for 40min at 1 �Mwith the fluo-4 acetoxymethylester (AM) (Molecular Probes) in Ca2� imaging buffer contain-ing 150 mM NaCl, 40 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 3 mM

HEPES, and 5.6 mM glucose, pH 7.4, at room temperature. Thecells were then washed with the Ca2� imaging buffer and incu-bated for 15 min at room temperature to allow de-esterifica-tion. In experiments with cytosolic EGTA-AM (MolecularProbes), 5 �M EGTA-AM was loaded simultaneously withfluo-4 AM. Ca2� imaging was carried out with an OlympusIX70-based TIRFM and an oil immersion 60� objective lens(numerical aperture (NA), 1.45). Fluorescence images wereobtained by excitation at 488 nmwith an Ar laser at room tem-perature (20–22 °C). The laser intensity was reduced by usingneutral density filters (6, 10, and 25%) to minimize cellular tox-icity. The emission signal was collected with a 505 nm dichroicmirror and a 510–550 nm band-pass barrier filter (Olympus),and the signals were captured with a cooled CCD, ORCA-ER(Hamamatsu Photonics) at 50 Hz with 8 � 8 pixel binning. Anose piece stage IX2-NPS (Olympus) was attached to themicroscope for stable focusing. The flow rate of the extracellu-lar solution during recordings was �1.5 ml/min. The Ca2�

imaging buffer containing 1 mM EGTA instead of CaCl2 wasused when extracellular Ca2� was quenched.Caged Ca2� Spot Uncaging—HeLa cells were loaded simul-

taneouslywith cagedCa2� o-nitrophenyl EGTAacetoxymethylester (NP-EGTA) (Molecular Probes) at 1 �M and fluo-4 for 40min. Cells were then observedwith aNikonA1 confocalmicro-scope systemwith the resonantmode of 488-nm laser scanning.A 403-nm laser was used to uncage NP-EGTA at a stationaryfocused spot for a predetermined period (typically 100–200ms) with minimal amplitude. To prevent IP3R openings duringthe experiment, 100 �M 2-aminoethyldiphenyl borate contain-

Multiple Release Sites Involved in Ca2� Microdomain

24564 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 287 • NUMBER 29 • JULY 13, 2012


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

ing buffer was used. Fluorescence images were acquired withthe NIS elements software (Nikon). Ca2� ionophore BrA23187(Molecular Probes) was added manually to the extracellularmedium at a final concentration of 5 �M when used.Data Acquisition and Analysis—Fluorescence images were

acquired and analyzed with a custom-made application TIWorkbench software (23) and the Igor Pro software (Wavem-etrics). The fluorescence signals F at each pixel were normal-ized against the baseline fluorescence F0 of the same pixel. F0was calculated by averaging the fluorescence intensity over fiveframes prior to the onset of the Ca2� microdomain signaldetection. Relative changes in fluorescence signals in each pixelwere calculated by using the formula r � F/F0. Part of eachframe (21 � 21 pixels, 12.3 � 12.3 �m) containing Ca2�

microdomains detected by visual inspectionwas used for fittingthe R values to the two-dimensional Gaussian distributionfunction,

R� x,y� � 1.0 � A � exp� � 1

2 � �1 � C2��x � x0

�x�2

� �y � y0

�y�2

�2 � C � �x � x0�� y � y0�

�x � �y�� (Eq. 1)

where x and y are two-dimensional pixel coordinates, x0 and y0are the x and y positions, respectively, of the centroid of thefluorescent signals in each frame; A is the amplitude; theparameters �x and �x are the S.D. of the pixel coordinates x andy, respectively; and C is the correlation coefficient (�1 � C �1). A smaller region (11 � 11 pixels) was occasionally used forfitting when the events occurred near the edges of the cells,where the noise level was high. Spatial widths along themajor axis and minor axis, Rl and Rs, respectively (supple-mental Fig. 1), of Ca2� profiles at height A�e�1⁄2 were calcu-lated by Equation 2.

Rl �1

� 1

2 � �1 � C2�� 1

�x2 �

1

�y2 � � 1

�x4 �

1

�y4 �

4 � C2 � 2

�x2 � �y

2 �Rs �

1

� 1

2 � �1 � C2�� 1

�x2 �

1

�y2 � � 1

�x4 �

1

�y4 �

4 � C2 � 2

�x2 � �y

2 �(Eq. 2)

The angle � (��/2 � � � �/2) between the major axis of theCa2� profiles at height A�e�1⁄2 and the x axis (supplemental Fig.1) was calculated by using Equations 3–6,

� � Arc tan (Eq. 3)

where cor 0, Arc tan 0;

� � Arc tan ��

2(Eq. 4)

where cor 0, Arc tan 0;

� � Arc tan ��

2(Eq. 5)

where cor � 0, Arc tan 0;

� � Arc tan (Eq. 6)

where cor� 0,Arc tan � 0; where � �2�C/(�y/�x � �x/�y).The statistical analyses were performed using Excel software(Microsoft).Estimation of ApparentDiffusionCoefficient of Ca2�Concen-

tration Profile Centroids—Assuming that the position of thecentroid (x0, y0) of the two-dimensional Gaussian profiledirectly reflects the position of the point source of Ca2� in eachframe, we estimated the apparent diffusion coefficient,D, of thepoint source based on the following relationship betweenmeansquared displacement and the interval t, as described previously(24).

MSD � 4 � D � t (Eq. 7)

Computational Modeling—To evaluate the ability of ouranalysis to extract the spatial properties of the IP3R distributionunderlying the event, we carried out a simple simulation of thedistribution of the Ca2�-bound indicator concentrations. Thesimulation space was a two-dimensional 20 � 20 �m space.The model involves fluo-4 as a mobile Ca2� buffer (diffusioncoefficient DFluo � 30 �m2/s, kon � 150 �M�1 s�1, koff � 450s�1), immobile endogenous Ca2� buffers (kon � 25 �M�1 s�1,koff � 50 s�1), and immobile IP3Rs as stationary Ca2� sources.All of the IP3R channels were assumed to be identical and toopen throughout the entire simulation period (20 ms). The dif-fusion coefficient of Ca2� was set to DCa � 300 �m2/s. Weignored the Ca2� pump activity and let excessive Ca2� diffuseout across the boundary toward the outside according to a pre-vious study (25). The Ca2� concentration outside the simula-tion spacewas set to the equilibriumvalue [Ca2�]equil� 100nM.The simulation was carried out in C language with the Xcodeapplication on a Macintosh computer.

RESULTS

Detection of Histamine-induced Ca2� Microdomains withTotal Internal Reflection Fluorescence Microscope—Fig. 1shows a local, transient Ca2� increase observed at 50 Hz withTIRFM in fluo-4-loaded HeLa cells. Similar local Ca2� tran-sients were detected in HeLa cells stimulated with 0.5–20 �M

histamine (Table 1), and none were detected in the absence ofhistamine (data not shown). Spatially restricted local Ca2�

increases were observed in the absence of extracellular Ca2�

(supplemental Fig. 2), indicating that Ca2� influx from theextracellular space was unnecessary to generate local Ca2�

transients. No histamine-induced local Ca2� transients weredetected in the presence of the IP3R- and capacitative Ca2�

entry-blocker 2-APB (100 �M) or the Ca2� pump inhibitorthapsigargin (1 �M) (data not shown). Because these pharma-cological characteristics and their spatial size and durationwere comparable with those of the IP3R-mediated Ca2� puffsor microdomains reported in HeLa cells (17) and Xenopusoocytes (26), we then analyzed detailed spatiotemporal profiles




ww

.jbc.org/D

ownloaded from






http://www.jbc.org/

of the local Ca2� transients detected with the TIRFM to inves-tigate the mechanism underlying the generation of Ca2�

microdomains.Quantitative Analysis of Individual Ca2�Microdomains by Fit-

ting with Two-dimensional Gaussian Distribution Function—We selected frames (512 � 512 pixels) that contained localCa2� transients to quantitatively analyze the spatiotemporalprofiles of the Ca2� microdomains and positioned a squareregion of interest measuring 18� 18 pixels (10.6� 10.6 �m) ineach frame to maximize the total fluorescence intensity withinthe region of interest. The fluorescence intensity of each pixelwas normalized against its intensity before the onset of Ca2�

transients (F/F0), and the events whose F/F0 value within theregion of interest was 1.15 were used to quantitatively evalu-ate Ca2� transients by fitting to a two-dimensional Gaussiandistribution function (Equation 1) (regions of 21 � 21 pixelswere used for the fitting; see “Experimental Procedures”).Images that apparently contained spatially overlapping Ca2�

transients were excluded from the analysis. The numbers ofCa2� transients analyzed are shown in Table 1. Repeated Ca2�

transients were observed at 16 sites in 13 cells during therecordings.The spatial profiles of the F/F0 values (Fig. 2,A–E) were then

fitted to a two-dimensional Gaussian distribution function (Fig.2, F–J). The signals remaining after subtracting the fitted two-dimensional Gaussian function from the original experimentaldata (Fig. 2, K–O) showed no obvious spatial bias, indicatingthat the two-dimensionalGaussian distribution function can beused to quantify the spatiotemporal properties of IP3-evoked

Ca2�microdomains. Equation 1 contains six parameters:A, thepeak height of the distribution; x0, y0, the coordinates of thecentroid in the two-dimensional space; �x and �y, the x-radiusand y-radius, respectively, of the elliptic intersection of the two-dimensional Gaussian distribution at the height A�e�1⁄2; and C,the cross-correlation term (�1 � C � 1). To quantitativelyanalyze the directional anisotropy of the Ca2� microdomains,the radii along the major axis and minor axis, Rl and Rs, respec-tively, of the elliptic horizontal intersection at height A�e�1⁄2

were calculated by substituting the values of parameters �x, �y,andC in Equation 2 (see “Experimental Procedures”). The valueof angle � between the major axis of the ellipse and the x axiswas calculated from the same parameters by using Equations3–6 (see “Experimental Procedures”).Spatiotemporal Evolution of Ca2� Microdomains—A typical

spatiotemporal evolution of histamine-evoked Ca2� microdo-main is shown in Fig. 3. The color-coded horizontal intersec-tions of the fitted two-dimensional Gaussian distribution atheight A�e�1⁄2 in the seven consecutive frames were superim-posed (Fig. 3A). Most of the horizontal intersections were ellip-tic rather than circular. Fig. 3B shows the temporal changes inamplitude of the two-dimensional Gaussian distributionsshown in Fig. 3A. The direction of the major axis seemed tochange randomly (Fig. 3C). It was obvious that the horizontalintersections were not concentric and that the centroid exhib-ited a considerable drift during both the rising phase and decayphase of the Ca2� transient. We analyzed these characteristicsof Ca2� microdomains in detail as described below.Ca2� Microdomain Size; Its Temporal Evolution, Variability,

and Relationship to Amplitude—Figs. 4,A–C, show the tempo-ral evolutions of the amplitude and spatial width. The ampli-tude of the Ca2� microdomains of a representative exampleobserved in HeLa cells peaked�50ms after crossing the detec-tion threshold and then gradually decreased with a time con-stant of 155 ms, when fitted to an exponentially decaying func-tion (Fig. 4A). The length of the major and minor axes (Fig. 4B)and the area (Fig. 4C) of the elliptic intersection at heightA�e�1⁄2

increased almost linearly, irrespective of the rising phase ordecay phase of the Ca2� puff amplitude. Both the peak ampli-tude and the spatial width at the peak frame, in which theamplitude of the profilewasmaximal, exhibited large variabilityin theCa2�microdomains detected (supplemental Fig. 3,A–C).Themean length of Rl and Rs in the peak frame was 2.50 0.08�m and 1.88 0.06 �m, respectively (n � 65). There was noclear relationship between peak amplitude and the area in thepeak frame (r � �0.216) (Fig. 4D).Isotropy of Ca2� Microdomains and Deviation from Two-di-

mensional Gaussian Distribution after Peak Frame—The tem-poral changes in the aspect ratio, Rs/Rl, of a typical Ca2�

FIGURE 1. A typical local Ca2� transient observed with a total internalreflection fluorescence microscope. Pseudo-color images normalizedfluo-4 signals (F/F0) in HeLa cells stimulated with 0.5 �M histamine. The inter-vals between the images are 40 ms. The outline of the cell is shown in the firstframe. Scale bar, 10 �m.

TABLE 1Statistics of the IP3-evoked Ca2� microdomain signals used in spatial profile analysis

Histamine(�M)

Cellsobserved (cells)

Cells withevents (cells)

Cells with repetitive Ca2� microdomainsignals at the same sites (cells)

Ca2� microdomainsignals (events)

0.5 33 3 1 71 33 2 2 52 21 1 1 45 50 8 5 3210 21 3 2 1020 22 4 2 8




ww

.jbc.org/D

ownloaded from


http://www.jbc.org/

microdomain are shown in Fig. 5A. The raw images and two-dimensional Gaussian distributions fitted to the acquiredimages for the frames indicated by the arrowheads in Fig. 5A arealso shown in Fig. 5B. The aspect ratio was not constant duringthe evolution of the signal, and its maximum value wasobserved in the peak frame (time � 0) (Fig. 5A). This tendencyis clearly seen in the plot of the average of all of the Ca2�

microdomain data (n� 65); the aspect ratio wasmaximal in thepeak frame (Fig. 5C). The relationship between Rl and Rs in thepeak frame is shown in Fig. 5D. They could be fitted to a straightline with a slope of 0.74 0.016 (r � 0.754) that contained theorigin (Fig. 5D), indicating that the anisotropy of the spatialdistribution of Ca2� in the peak frame is almost constant irre-spective of Ca2� microdomain size.To examine whether the anisotropic nature of the spatial

profile is significant, we prepared a series of computer-gener-ated isotropic Gaussian signal intensity profiles with the actualnoise data we observed in our experiments. The spatial size(2.19 �m) and the height (0.45) of the profile were set to be

equal to the experimental results at the peak frames, and thenoise data (a background cell image without Ca2� microdo-main signals) were collected arbitrarily from a region of 21� 21pixels in a quiescent frame before Ca2� microdomain signalswere observed. Two-dimensional Gaussian fitting analysis wascarried out with the computer-generated data, and the aspectratio was calculated for each profile. The value obtained for theisotropic profiles was 0.891 0.041 (mean S.D., n � 10),whereas the result of the experiment was 0.762 0.111(mean S.D.,n� 65). The experimental valuewas significantly(p � 0.01, Student’s t test) smaller than that expected for theisotropic profiles, implying that the observed anisotropy wassignificant, rather than an artifact of our analysis methods.Spatial Drift of Ca2� Distribution Centroids—Fig. 6A shows

the trajectory of the centroid (x0, y0) of the two-dimensionalGaussian distribution fitted to the Ca2� microdomainsobserved in a HeLa cell stimulated with 2 �M histamine. Thetrajectory seemed to undergo random drift resembling Brown-ian diffusion, and the position of the centroid must somewhatreflect the distribution of Ca2� supply through IP3Rs involvedin the event. To evaluate the possibility that individual Ca2�

FIGURE 2. Two-dimensional Gaussian fitting of an IP3-evoked Ca2�

microdomain signaling spatial profile. Representative pseudo-colorimages of normalized signals (F/F0) observed in HeLa cells stimulated with 0.5�M histamine (A–E) and two-dimensional (2D) Gaussian profiles fitted to thenormalized fluo-4 signals (F and G) are presented. Residual signals after sub-tracting the fitted two-dimensional Gaussian function (F–J) from the corre-sponding raw data (A–E) are shown in a different pseudo-color scale (K–O)that covers the same range as for (A–J). The intervals between consecutiveimages are 20 ms.

FIGURE 3. Temporal evolution of the two-dimensional profile of a typicalCa2� microdomain signal. A, color-coded horizontal intersections at heightA�e�1⁄2 of the two-dimensional Gaussian distribution fitted to the normalizedfluo-4 signals in seven consecutive frames are superimposed. The fluo-4 sig-nals were recorded in a cell stimulated with 0.5 �M histamine. The framenumber of each intersection is indicated by the same color code as in B.B, time course of the amplitude of the puff event shown in A. C, time course ofangle � of the event shown in A. rad, radian.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

microdomains are generated from a single IP3R cluster underBrownian diffusion and cause the centroid drift in the distribu-tion of Ca2�, the apparent diffusion coefficient of the Ca2�

distribution centroids was estimated as described under“Experimental Procedures.” The relationship between meansquared displacement of the centroid shown in the left panel(EGTA-free) in Fig. 6A, and time t is shown in Fig. 6B. In thisexample, the apparent diffusion coefficient was estimated to be3.20 �m2/s. Fig. 6C shows the histogram of the apparent diffu-sion coefficients estimated in a similar manner for 42 events.The apparent diffusion coefficient estimated from the averagemean squared displacement value was D � 4.20 0.50 �m2/s(mean S.D., n� 42), which is�10-fold larger than the appar-ent diffusion coefficient estimated for type 1 IP3Rs (0.26 �m2/s(27)) and type 3 IP3Rs (0.45�m2/s (27) or 0.04�m2/s (28)), bothof which are expressed inHeLa cells (29). These results stronglysuggest that the centroid drift of the two-dimensional Gaussiandistribution cannot be explained by a point source of Ca2� thatcorresponds to a single IP3R cluster.

In previous studies adopting intracellular slow Ca2� buffers,such as EGTA, to facilitate the observation of microdomains,the sites of the local Ca2� increase were reported to beextremely stable (20).We therefore speculated that the discrep-ancy between the spatially mobile nature of the Ca2� microdo-mains we observed and the previous reports is due to theinvolvement of the excessive cytosolic Ca2� buffers. To con-firm this speculation, we carried out the spatial drift detectionof Ca2� microdomain centroids in our system in the existenceof cytosolic EGTA-AM (Fig. 6A, right panel). The spatial size ofeach Ca2� microdomain signal was considerably narrowed bythe cytosolic buffers (major radius Rl EGTA � 0.742 0.12 �m,minor radius Rs EGTA � 0.606 0.11 �m for peak frames,respectively. n � 10 events from five cells) in consistent withother reports (22). The centroids were spatially stable with sig-

nificantly lower extent of the spatial drift (DEGTA � 0.611 0.413 �m2/s, n � 6 events from five cells).

The above experiments suggest that the rapid centroid spa-tial drift observed in the current studywas because of the lack ofexcessive exogenous Ca2� buffers in the cytosol. However, itshould be noted that the spatially restricted profiles of themicrodomains in the presence of cytosolic EGTAcan inevitablychange the scale factor. Consequently, the addition of EGTAreduces the potential range of centroid spatial excursionswithin the overall calcium microdomain that could be artifi-cially generated by noise in the raw image data when processedwith the two-dimensional Gaussian fit methodology. To con-firm that the centroid drift is not due to experimental noise, wecarried out the spot uncaging of caged Ca2� NP-EGTA in thecytosol, which caused a local Ca2� increase similar to the Ca2�

FIGURE 4. Relationship between the temporal changes in amplitude andspatial width. Shown are the time course of the amplitude A (A), the length ofthe major axis and minor axis, Rl (filled circles) and Rs (empty circles), respec-tively (B), and the area of the horizontal intersection at height A�e�1⁄2 (C) of aCa2� microdomain signal observed in a cell stimulated with 0.5 �M histamine.D, relationship between the amplitude and area in the peak frame. The hista-mine concentrations applied are indicated by different symbols.

FIGURE 5. Spatial isotropy of IP3-evoked Ca2� microdomain signals.A, temporal evolution of the aspect ratio, Rs/Rl, of a typical IP3-evoked Ca2�

microdomain signal. The event was observed in a cell stimulated with 10 �M

histamine. B, pseudo-colored images of normalized fluo-4 signals and fittedtwo-dimensional Gaussian distributions in the frames indicated by the arrow-heads in A. x indicates the position of the centroids of the fitted distributions.C, mean values of the aspect ratio averaged over all of the events (n � 65)were plotted against time after the peak frame. D, relationship between radiialong the major and minor axes, Rl and Rs, respectively in the peak frames. Thehistamine concentrations applied are indicated by different colors. 2D,two-dimensional.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

microdomains in its amplitude and spatial size (Fig. 7A). Two-dimensional Gaussian fitting was similarly performed for thelocal Ca2� increases, although the noise level was significantlyhigher than that of the TIRFM data. This was primarily due todifferences in optical systems (Fig. 7B). This synthetic localCa2� increase reached the steady state shortly after the onset ofthe photo-uncaging, and the amplitude started to decay directlyafter the photo-stimulation ended (Fig. 7A). To evaluate thepossibility that Fluo-4 was saturated with Ca2� during uncag-ing, we measured the fluorescence ratio increase in the pres-ence of an excess amount of Ca2� evoked by Ca2� ionophoreBrA23187. After the addition of 5 �M BrA23187, the fluores-cent ratio reached 3.01 0.37, which was significantly higher(p � 0.01, Student’s t test) than that evoked by NP-EGTAuncaging under the condition used (1.24 0.09, n � 22 fromthree cells). These results indicate that the steady state increasein Fluo-4 signals during uncaging treatments was not due to thesaturation of the dye. The spatial size (major radius, 1.92 0.195�m;minor radius, 1.67 0.140�mat peak frames,n� 22from three cells) and the amplitude of the signal profile wascomparable with those obtained for Ca2� microdomain signals(supplemental Fig. 3); however, the underlying spatiotemporaldistribution of Ca2� release fluxes should be quite different.The time constant of the decay in the amplitude was obtainedby fitting the decay phase to an exponential function, ( � 14.4ms for the data shown in Fig. 7A). This decay time constant was

significantly smaller than time constant of the real Ca2�

microdomains, which was 223 214 ms (n � 64, Fig. 7C). Therapid kinetics of the signal decay may be due to the rapid Ca2�

buffering effect caused by the Ca2�-unbound fraction of NP-EGTA. An alternative interpretation is that the decay phase ofthe native Ca2� microdomain signals is caused by the gradualdecrease in the number of active Ca2� release channels. Thecentroid of the Ca2� distribution caused by spot uncagingshowed practically no drift over 200 ms of photo-stimulation(Fig. 7D). Although we could not completely exclude the possi-bility that a different distribution of Ca2� sources (Ca2� releasechannels on the ER andNP-EGTA in cytosol) causes the differ-ence in the movement of the centroid, these results demon-strate that the Ca2� release from an immobile or stationaryrelease site does not lead to the spatial drift of the centroid, evenwith experiments that have higher noise levels.To estimate the possible underlying distributions of theCa2�

release sites, we performed a numerical simulation of the Ca2�-bound fluo-4 concentration in two-dimensional space. Wefound that the uniform distributions across the �m-scale spaceof Ca2� release sites (Fig. 8,A–D), which should be IP3Rs in thecurrent study, can generate various quasi-Gaussian fluores-cence intensity profiles similar to the ones we observed (Fig. 8,E–H). To further evaluate the two-dimensionalGaussian fittinganalysis, we pixelated the results to simulate the pixels of theCCD (Fig. 8, I–L) and added distinct actual noise profiles from

FIGURE 6. Apparent diffusion coefficients of Ca2� microdomain signalcentroids. A, trajectories of the centroid of typical Ca2� microdomain signalsevoked with 2 �M histamine in a cell without cytosolic EGTA (left panel) and ina cell loaded with 5 �M EGTA-AM (right panel) are shown in the same spatialscale. The position of the centroid in the termination frame (see text) is indi-cated by the x in the left panel. B, the mean squared displacement (MSD)-t plotfor the centroid shown in the left panel in A. The apparent diffusion coefficientwas estimated to be 3.2 �m2/s. C, histogram of the apparent diffusion coeffi-cient values estimated from the Ca2� microdomain signals observed (n � 42).

FIGURE 7. Comparison between IP3-evoked Ca2� microdomain signalsand the Ca2� signals caused by the spot uncaging of NP-EGTA. A, localCa2� increases evoked by the spot uncaging of NP-EGTA by confocal micro-scope. The transient of the peak amplitude is shown in the lower panel. Thesignal rapidly declined after the termination of the photo-stimulation ( �69.7 ms). The spatial profile of F/F0 at the frame shown by the open arrowheadin the plot is shown in the upper panel. B, noise level analysis of the opticalsystems. S.D. of the fluorescence intensities of 21 � 21 pixels in frames beforethe Ca2� increase are shown for TIRFM used for Ca2� microdomain imaging(n � 71 images) and the confocal microscope used for NP-EGTA uncaging(n � 4 images). *, p � 0.01; Student’s t test. C, histogram of the characteristictime of the declining phase of the IP3-evoked Ca2� microdomain signal (n �64 events). The value obtained for the synthetic signals evoked by NP-EGTA(17.2 3.70 ms; mean S.D., n � 4) is shown by the filled arrowhead.D, spatial drift of the signal centroid evoked by NP-EGTA uncaging. The signalshowed little spatial drift (D � 0.44 �m2/s for the data shown (solid line))during photo-stimulation, compared with the result for IP3-evoked Ca2�

microdomain signals shown by the dashed line (D � 4.20 �m2/s, n � 42).




ww

.jbc.org/D

ownloaded from


http://www.jbc.org/

quiescent frames at the Ca2� microdomain sites obtained inour measurements (Fig. 8,M–P). The two-dimensional Gauss-ian fitting was carried out with the simulated data and theresults clearly reflect the spatial profiles of the underlying Ca2�

release site distributions, e.g. spatial isotropy (Fig. 8,Q–T). Thesimulation thus demonstrated the ability of our analysis toextract the spatial nature of the underlying distribution of Ca2�

release sites.Because in some cells, repeated local Ca2� increases that

originated at the same sites were detected, the trajectories ofthe centroid drifts of different events observed at the same siteswere compared. Fig. 9 shows representative examples observedin two different cells. The trajectories are superimposed in Fig.9, A and B. When the positions of the centroids in the initialframewere compared, inwhich the amplitude of the signals hadjust crossed the detection threshold (Fig. 9, C and D), the peakframe (Fig. 9, E and F), and the termination frame, i.e. the lastframes in which the fluorescent signals were successively fittedwith the two-dimensional Gaussian distribution function (Fig.9, G and H), we found that the centroids were located within arelatively small area in the peak frames (Fig. 9, E and F) but werewidely distributed in the initiation frames (Fig. 9, C and D) andtermination frames (Fig. 9, G and H). We also found that thecentroids moved randomly, rather than following stable orbits(Fig. 9, I and J).

DISCUSSION

Intracellular Ca2� concentrations are buffered strongly bydiverse Ca2� binding proteins in living cells. Buffered diffusionof Ca2� near a point source in the presence of stationary andmobile Ca2� buffers has been studied (30–33). Complex shap-ing of Ca2� signal profiles by Ca2� buffers described by previ-ous studies implies the importance of observing the event withminimal exogenous Ca2� buffers, e.g. cytosolic EGTA has beenused often in previous studies (33). These assumptions led us toadopt a two-dimensional spatial profile analysis by Gaussianfitting in Ca2�-chelator-free, or intact HeLa cells.In this study, we visualized histamine-evokedCa2�microdo-

mains in intact HeLa cells using TIRFM and the Ca2� indicatorfluo-4. The results showed that the two-dimensional fluores-cence signals during IP3-evoked Ca2� microdomains in hista-

FIGURE 8. Computational modeling of the spatial profile of Ca2�-boundfluo-4. Simple computational simulation in two-dimensional space was car-ried out with four different distribution patterns of the open IP3Rs depicted asred filled circles in A–D. Spatial profiles of Ca2�-bound fluo-4 at 20 ms after thechannel opening with the four IP3R distribution patterns shown as F/F0 pseu-do-colored images in E–H. To simulate the resultant fluorescence images, theresults were pixelated by spatial averaging within the region covered by asingle CCD pixel (I--L). To further demonstrate the efficiency of the analysis,actual noise profiles were added to the results (M–P), and the two-dimen-sional Gaussian distributions were fitted to the data (Q–T).

FIGURE 9. Trajectories of the centroids of repeated Ca2� microdomainsignals evoked at the same site. Trajectories of Ca2� microdomain signalcentroids detected in two different cells stimulated with histamine 5 �M (Cell1) or 10 �M (Cell 2) are superimposed (A, B, I, and J),. Distinct events are indi-cated by different colors. The positions of the centroid in the initiation frame,peak frame and termination frame are indicated by squares (A–D), circles (A, B,E, and F), and x (A, B, E and F), and x (A, B, G, and H), respectively. Scale bars, 1�m.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

mine-stimulatedHeLa cellswere closely approximated by a sin-gle two-dimensional Gaussian distribution function (Fig. 2),and the approximation enabled us to analyze the amplitude,spatial width, isotropy, and the position of the centroid of theevents quantitatively in intact HeLa cells. A simple simulationstudy with the same Gaussian fitting analysis demonstratedthat the results of the fitting would actually reflect the proper-ties of the underlying IP3R distributions. (Fig. 8).If a spatially fixed single point source alone accounted for the

generation of Ca2� microdomains, maximal fluorescence sig-nals were expected to be observed at the same site during localCa2� transients and fluorescence signals should evolve in theform of concentric circles from the point source. Actually, alocal Ca2� increase induced by the spot NP-EGTA uncagingrevealed very little spatial drift of the centroid (Fig. 7). As shownin Fig. 3A, however, the elliptic horizontal intersections of thefitted two-dimensional Gaussian distributionwere not concen-tric and the centroid exhibited considerable drift, with a meanapparent diffusion coefficient value of D � 4.20 0.50 �m2/s(mean S.D., n� 42) (Fig. 6), which is�10-fold larger than thevalues estimated for type 1 IP3Rs (0.26 �m2/s (27)) and type 3IP3Rs (0.45 �m2/s (27), 0.04 �m2/s (28)). More than 90% of thecentroids analyzed (38 of 42 events) had an apparent diffusioncoefficient 1 �m2/s (Fig. 6C). Although the diffusion coeffi-cients of IP3Rs should be affected bymany factors, including thegeometry of the cell and the ER, and binding proteins, theresults clearly indicate that almost all of the Ca2� microdo-mains detected in histamine-stimulated intact HeLa cells werenot generated from spatially fixed or stable single IP3R clusters,nor from single IP3R clusters under Brownian diffusion.The alternative mechanism for the centroid drift of Ca2�

microdomains would be the cooperative employment of multi-ple release sites to a single event. We found that the horizontalintersections of the fitted two-dimensional Gaussian distribu-tion were not the concentric circles that would be expectedfrom the distribution of Ca2� released from a single pointsource. Most of the horizontal intersections were elliptic, andthe direction of the major axis rotated randomly (Fig. 3, A andC). We also found that the centroids of repeated local Ca2�

increases were located within a relatively small area in the peakframes (Fig. 9, E and F) but were distributed widely in the initi-ation and termination frames (Fig. 9,C,D,G, andH).Moreover,the trajectories of the centroids of repeated events at the samesites did not coincide (Fig. 9, I and J). Based on these findings,we conclude that individual IP3-evoked Ca2� microdomainsare generated by non-coordinated, stochastic Ca2� releasesfrommultiple Ca2� puff sites. Ca2� microdomains may be ini-tiated by stochastic release at a single site and gradually growsimultaneously with non-coordinated Ca2� releases fromnearby sites. Delayed releases of Ca2� from nearby sites mayresult in a rapid drift of Ca2� microdomain centroids (Fig. 6)and random rotations of the direction of the major axes of thesignal profiles (Fig. 3). Our conclusion is consistent with anearlier observation of microscopic Ca2� waves within Ca2�

puffs, or microdomains in our terms, by confocal line-scanimaging of histamine-stimulated HeLa cells (14).Demuro and Parker (34) analyzed IP3-evoked Ca2� puffs in

Xenopus oocytes and found that the centroid of some puffs

exhibited submicron jumps. Although the authors injected aslow Ca2� buffer EGTA (35) into the cytosol before the imag-ing, the behavior of the puff centroids in Xenopus oocytes wasreminiscent of the behavior observed in HeLa cells in the cur-rent study (Fig. 6A). However, the distance of the jump theyobserved was basically limited to several hundred nanometers,whereas the centroid drift we observed sometimes exceeded 1�m.Moreover, they observed the jump in a subset of the events,which contrasts with the 90% of the events showing the cen-troid drift in our study. The same group has recently adoptedTIRFM to image Ca2� puffs in human neuroblastomaSH-SY5Y cells (20, 22) and several other cell lines, includingHeLa cells (20), also in the presence of cytosolic EGTA. Theyreported that the nature of the Ca2� puff site locations are evenmore stable than reported in Xenopus oocytes; this observationcontrasts the results presented herein.These differences are probably caused by the existence of

excessive Ca2� buffers in the cytosol in previous studies, asconfirmed by our observation with cytosolic loading of Ca2�

indicator EGTA-AM (Fig. 6A). IP3Rs are, by their nature, Ca2�-induced Ca2�-release channels; they are activated by low (� �M)concentrations of Ca2�, whereas at high Ca2� concentrations,the IP3Rs are inhibited by the Ca2� (7). The positive feedbackmechanism at lower Ca2� concentrations via Ca2� releasedthrough IP3Rs enables the cooperative opening of spatially iso-lated neighboring IP3Rs. Taking into account the positive feed-back effect of Ca2� and the larger drift of the event centroid inHeLa cells without cytosolic EGTA, the behavior of IP3-evokedCa2� microdomains under physiological conditions should beaffected by the existence of cytosolic Ca2� buffers. The exis-tence of Ca2� indicators, such as fluo-4 in the current study,which cannot be avoided for Ca2� imaging studies, might thusaffect the physiological Ca2� signaling system, and requirescareful consideration in the context of Ca2� microdomainstudies.We found that the length of the major and minor axes (Fig.

4B) and the area (Fig. 4C) of the elliptic intersection of the Ca2�

distribution profiles continued to increase throughout theevent, irrespective of the rising or decaying phase of the peakamplitudes. There was no clear correlation between the peakamplitudes and the spatial areas in the peak frames (Fig. 4D).The mechanism responsible for this linear increase in the areaover time is unknown, but the release of Ca2� may bemediatedby the same process during both the rising phase and the fallingphase of Ca2� microdomains. In this respect, stochastic prop-agation of the cooperative IP3R openings across multiple Ca2�

release sites seems to offer a reasonable explanation for thephenomenon, as the declining phase could simply be attributedto fewer channels being open, rather than passive diffusion andbuffering of Ca2�. This explanation was further supported bythe computational modeling we carried out, in which spatialdistributions of Ca2� release sites directly reflect fluo-4 signalprofiles (Fig. 8). Although more detailed computational analy-sis involving precise evaluations of Ca2� homeostatic toolkitsand gating models of IP3Rs will be desirable and eventually berequired, these results suggest the cooperative and stochasticmultiple IP3R cluster openings as amechanism for the develop-ment of each Ca2� microdomain.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

CONCLUSIONS

Although a large body of knowledge about the behavior oftightly clustered IP3Rs has been accumulated by a series ofCa2�

puff imaging studies using cytosolic EGTA (20, 34, 35), actualactivity of IP3Rs in intact cells remains to be clarified. To eluci-date the spatial and temporal organization of each IP3-inducedCa2� microdomain in intact cells, we investigated Ca2�

microdomains in histamine-stimulated HeLa cells withoutcytosolic loading with EGTA. Quantitative analysis on the spa-tial drift and other spatiotemporal features of the event lead tothe conclusion that the functional unit underlying each IP3-de-pendent Ca2� microdomain is composed of multiple Ca2� puffsites in intact cells, which was also confirmed by a numericalsimulation.

Acknowledgments—We thank Drs. Haruka Yamazaki, Toru Matsu-ura,Masahiro Enomoto, Sachiko Ishida, and all of the othermembersof the Developmental Neurobiology Laboratory for stimulating dis-cussions. We thank Nikon Instruments, Inc. for technical support.

REFERENCES1. Clapham, D. E. (1995) Calcium signaling. Cell 80, 259–2682. Berridge, M. J., Lipp, P., and Bootman, M. D. (2000) The versatility and

universality of calcium signaling. Nat. Rev. Mol. Cell Biol. 1, 11–213. Berridge, M. J., Bootman, M. D., and Roderick, H. L. (2003) Calcium sig-

naling: Dynamics, homeostasis, and remodeling. Nat. Rev. Mol. Cell Biol.4, 517–529

4. Rizzuto, R., and Pozzan, T. (2006) Microdomains of intracellular Ca2�:Molecular determinants and functional consequences. Physiol. Rev. 86,369–408

5. Berridge,M. J. (2006) Calciummicrodomains: Organization and function.Cell Calcium 40, 405–412

6. Berridge, M. J. (2009) Inositol trisphosphate and calcium signaling mech-anisms. Biochim. Biophys. Acta 1793, 933–940

7. Bezprozvanny, I.,Watras, J., and Ehrlich, B. E. (1991) Bell-shaped calcium-response curves of Ins(1,4,5)P3- and calcium-gated channels from endo-plasmic reticulum of cerebellum. Nature 351, 751–754

8. Denk, W., Sugimori, M., and Llinas, R. (1995) Two types of calcium re-sponse limited to single spines in cerebellar Purkinje cells. Proc. Natl.Acad. Sci. U.S.A. 92, 8279–8282

9. Simkus, C. R., and Stricker, C. (2002) The contribution of intracellularcalcium stores to mEPSCs recorded in layer II neurones of rat barrel cor-tex. J. Physiol. 545, 521–535

10. Cardenas, C.,Miller, R. A., Smith, I., Bui, T.,Molgo, J.,Muller,M., Vais, H.,Cheung, K. H., Yang, J., Parker, I., Thompson, C. B., Birnbaum, M. J.,Hallows, K. R., and Foskett, J. K. (2010) Essential regulation of cell bioen-ergetics by constitutive InsP3 receptor Ca2� transfer to mitochondria.Cell 142, 270–283

11. Hayashi, T., and Su, T. P. (2007) �-1 receptor chaperones at the ER-mitochondrion interface regulate Ca2� signaling and cell survival. Cell131, 596–610

12. Parker, I., and Ivorra, I. (1990) Localized all-or-none calcium liberation byinositol trisphosphate. Science 250, 977–979

13. Parker, I., and Yao, Y. (1991) Regenerative release of calcium from func-tionally discrete subcellular stores by inositol trisphosphate. Proc. Biol. Sci.246, 269–274

14. Bootman, M., Niggli, E., Berridge, M., and Lipp, P. (1997) Imaging thehierarchical Ca2� signaling system in HeLa cells. J. Physiol. 499, 307–314

15. Reber, B. F., and Schindelholz, B. (1996) Detection of a trigger zone ofbradykinin-induced fast calcium waves in PC12 neurites. Pflugers Arch432, 893–903

16. Koizumi, S., Bootman,M.D., Bobanovic, L. K., Schell,M. J., Berridge,M. J.,and Lipp, P. (1999) Characterization of elementary Ca2� release signals inNGF-differentiated PC12 cells and hippocampal neurons. Neuron 22,125–137

17. Thomas, D., Lipp, P., Berridge, M. J., and Bootman, M. D. (1998) Hor-mone-evoked elementary Ca2� signals are not stereotypic, but reflect ac-tivation of different size channel clusters and variable recruitment ofchannels within a cluster. J. Biol. Chem. 273, 27130–27136

18. Bootman, M. D., and Berridge, M. J. (1996) Subcellular Ca2� signals un-derlying waves and graded responses in HeLa cells.Curr. Biol. 6, 855–865

19. Smith, I. F., and Parker, I. (2009) Imaging the quantal substructure ofsingle IP3R channel activity during Ca2� puffs in intact mammalian cells.Proc. Natl. Acad. Sci. U.S.A. 106, 6404–6409

20. Smith, I. F., Wiltgen, S. M., Shuai, J., and Parker, I. (2009) Ca2� puffsoriginate from preestablished stable clusters of inositol trisphosphate re-ceptors. Sci. Signal 2, ra77

21. Axelrod, D. (2003) Total internal reflection fluorescence microscopy incell biology.Methods Enzymol. 361, 1–33

22. Smith, I. F., Wiltgen, S. M., and Parker, I. (2009) Localization of puff sitesadjacent to the plasmamembrane: Functional and spatial characterizationof Ca2� signaling in SH-SY5Y cells utilizing membrane-permeant cagedIP3. Cell Calcium 45, 65–76

23. Matsu-ura, T., Michikawa, T., Inoue, T., Miyawaki, A., Yoshida, M., andMikoshiba, K. (2006) Cytosolic inositol 1,4,5-trisphosphate dynamics dur-ing intracellular calcium oscillations in living cells. J. Cell Biol. 173,755–765

24. Kusumi, A., Sako, Y., and Yamamoto,M. (1993) Confined lateral diffusionof membrane receptors as studied by single particle tracking (nanovidmicroscopy). Effects of calcium-induced differentiation in cultured epi-thelial cells. Biophys. J. 65, 2021–2040

25. Shuai, J., and Parker, I. (2005)Optical single-channel recording by imagingCa2� flux through individual ion channels: Theoretical considerationsand limits to resolution. Cell Calcium 37, 283–299

26. Yao, Y., Choi, J., and Parker, I. (1995) Quantal puffs of intracellular Ca2�

evoked by inositol trisphosphate in Xenopus oocytes. J. Physiol. 482,533–553

27. Fukatsu, K., Bannai, H., Zhang, S., Nakamura, H., Inoue, T., andMikoshiba, K. (2004) Lateral diffusion of inositol 1,4,5-trisphosphatereceptor type 1 is regulated by actin filaments and 4.1 N in neuronaldendrites. J. Biol. Chem. 279, 48976–48982

28. Ferreri-Jacobia, M., Mak, D. O., and Foskett, J. K. (2005) Translationalmobility of the type 3 inositol 1,4,5-trisphosphate receptor Ca2� releasechannel in endoplasmic reticulum membrane. J. Biol. Chem. 280,3824–3831

29. Tovey, S. C., de Smet, P., Lipp, P., Thomas, D., Young, K.W., Missiaen, L.,De Smedt, H., Parys, J. B., Berridge, M. J., Thuring, J., Holmes, A., andBootman,M. D. (2001) Calcium puffs are generic InsP3-activated elemen-tary calcium signals and are down-regulated by prolonged hormonal stim-ulation to inhibit cellular calcium responses. J. Cell Sci. 114, 3979–3989

30. Stern, M. D. (1992) Buffering of calcium in the vicinity of a channel pore.Cell Calcium 13, 183–192

31. Wagner, J., and Keizer, J. (1994) Effects of rapid buffers on Ca2� diffusionand Ca2� oscillations. Biophys. J. 67, 447–456

32. Smith, G. D., Wagner, J., and Keizer, J. (1996) Validity of the rapid buffer-ing approximation near a point source of calcium ions. Biophys. J. 70,2527–2539

33. Smith,G.D., Keizer, J. E., Stern,M.D., Lederer,W. J., andCheng,H. (1998)A simple numerical model of calcium spark formation and detection incardiac myocytes. Biophys. J. 75, 15–32

34. Demuro, A., and Parker, I. (2008)Multidimensional resolution of elemen-tary Ca2� signals by simultaneous multifocal imaging. Cell Calcium 43,367–374

35. Callamaras, N., and Parker, I. (2000) Phasic characteristic of elementaryCa2� release sites underlies quantal responses to IP3. EMBO J. 19,3608–3617




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

and Katsuhiko MikoshibaHideki Nakamura, Hiroko Bannai, Takafumi Inoue, Takayuki Michikawa, Masaki Sano

Generate Calcium Microdomains in Intact HeLa CellsCooperative and Stochastic Calcium Releases from Multiple Calcium Puff Sites

doi: 10.1074/jbc.M111.311399 originally published online May 25, 20122012, 287:24563-24572.J. Biol. Chem.

10.1074/jbc.M111.311399Access 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/2012/05/25/M111.311399.DC1

http://www.jbc.org/content/287/29/24563.full.html#ref-list-1

This article cites 35 references, 9 of which can be accessed free at


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/lookup/doi/10.1074/jbc.M111.311399

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;287/29/24563&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/287/29/24563

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=287/29/24563&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/287/29/24563

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/suppl/2012/05/25/M111.311399.DC1

http://www.jbc.org/content/287/29/24563.full.html#ref-list-1

http://www.jbc.org/

cooperativeandstochasticcalciumreleasesfrommultiple ... › content › 287 › 29 ›...

Documents