Journal of Cell Science 102, 417-425 (1992)Printed in Great Britain © The Company of Biologists Limited 1992

417

Subcellular imaging of calcium exchange in cultured cells with ion

microscopy

SUBHASH CHANDRA, WALTER A. AUSSERER* and GEORGE H. MORRISON

Department of Chemistry, Baker Laboratory, Cornell University, Ithaca, NY 14853, USA

*Present address: Life Sciences Division, SRI International, 333 Ravenwood Avenue, Menlo Park, CA 94025, USA

Summary

Calcium to calcium exchange between intracellular andextracellular pools has been imaged directly in indi-vidual cells using stable 44Ca (98.78% enrichment) in thenutrient medium and an isotopic imaging technique, ionmicroscopy. Observations were made by imaging mass40 to determine the native intracellular calcium (40Ca),and mass 44 to localize the 44Ca from the extracellularmedium that exchanged with cellular calcium. LLC-PK,porcine kidney epithelial cells were exposed to thenutrient medium with 1.87 mM stable 44Ca for 0,1, 2, 5,8, 20, 60 and 90 min, and cryogenically prepared priorto ion microscopic analysis. The cell nucleus, the Golgiregion and the remaining cell cytoplasm could bespatially resolved to within about 0.5 fim using the ionmicroscope. On the basis of the kinetics of 40Ca to 44Caexchange it was observed that all three compartmentshad a rapidly exchanging pool of calcium, which took

about 2 min to exchange. A moderately rapidlyexchanging pool of calcium was identified between 2 and20 min of calcium exchange. The cells had exchangedabout 50% of their total internal calcium with theexternal calcium in less than 20 min. The remaining 50%of the cellular calcium could be classified as a slowlyexchanging pool. Isotopic images of 39K and 23Na wererecorded along with 40Ca and 44Ca images to assess thehealth status of the cell. Isotopic imaging has the uniqueability to distinguish intracellular calcium from theextracellular calcium that enters cells and has enormouspotential for studies of calcium transport under physio-logical and pathological conditions.

Key words: calcium to calcium exchange, calcium imaging,LLC-PKj cells, ion microscopy, SIMS, ^Ca-stable isotope,laser scanning confocal microscopy, Golgi apparatus.

Introduction

Intracellular calcium exchanges rapidly with extracell-ular calcium; its measurement at a subcellular level hasbeen a major challenge. The rapid exchange of internalcalcium with external calcium has been shown using theradionuclide 45Ca2+ as a tracer (Borle, 1972; Kuwataand Langer, 1989; Langer et al. 1976). This radiotracerhas also been used in autoradiographic studies (McGuf-fee et al. 1981; Sampson et al. 1970), and on isolated cellfractions, in order to study calcium transport indifferent cellular compartments (Finch et al. 1991;Krause and Lew, 1987; Thevenod et al. 1989). With45Ca2+, however, one can only monitor a unidirectionalinflux or efflux of the radioactive species, and theinformation is generally based on a population of cells.For understanding the role of intracellular calciumstores in the release and uptake of free ionized [Ca2+]junder physiological and pathological conditions, it iscrucial to make these observations at a subcellularlevel. Although intracellular free Ca2+ concentrations[Ca2+]j can now be imaged and quantitated at subcellu-lar levels with fluorescent indicators (Brundage et al.

1991; Hernandez-Cruz et al. 1990; Tsien and Poenie,1986), it has been difficult to estimate the magnitude ofintracellular calcium stores, and calcium exchangebetween stores and the extracellular pool of calcium.We have been able to study calcium exchange at asubcellular level in individual cells by using stable ^Ca(98.78% enrichment) in the nutrient medium and thetechnique of ion microscopy. Ion microscopy, based onsecondary ion mass spectrometry (SIMS), providesvisual images of isotopic gradients in relation to cellmorphology with a spatial resolution of about 0.5 [im(Burns, 1982; Castaing and Slodzian, 1962; Chandraand Morrison, 1988; Galle, 1986; Linton, 1986; Morri-son and Slodzian, 1975). Since the technique distinguishesbetween isotopes on the basis of their mass-to-chargeratio and the analysis is made by continuously eroding(sputtering) the cell surface, it is possible to imageintracellular distributions of several different isotopesfrom the same cell. This unique isotopic imagingcapability of ion microscopy was utilized in studyingcalcium to calcium exchange between spatially resolvedintracellular compartments of LLC-PKi porcine kidneyepithelial cells and the external nutrient medium.

418 S. Chandra and others

Materials and methods

Cell culture and cryogenic sample preparationLLC-PK] cells (from American Type Culture Collection)were seeded at a density of 5x 105 cells per 100 mm in Falconplastic dishes and incubated at 37°C in a 5% CO2 atmosphere.Each dish contained 16 small wafer chips (approximately 1cm2) of semiconductor grade high-purity germanium (Gen-eral Diode) in order to provide sister cells growing on thedifferent chips, to be subjected to separate treatments. Anelectrically conducting cell growth substratum is necessary forSIMS analysis, since the sample is held at 4500 volts in the ionmicroscope, and germanium does not interfere in the analysisof 40Ca, ^Ca, 39K and BNa isotopes. The nutrient mediumwas Medium-199 supplemented with 10% fetal bovine serum(GIBCO). This cell growth medium contained 1.87 mMcalcium as determined by atomic emission spectroscopy usingan inductively coupled plasma system (Fassel and Kniseley,1974). Within 48 h of the cell seeding the cells were about 60%confluent on germanium. At this time the cells wereequilibrated in fresh growth medium for 2 h, after which thecells growing on different germanium chips under identicalconditions (sister cells) were subjected to eight treatments.The cells were exposed for 0 min (the cells sampled directlyafter equilibration in the fresh growth medium), 1 min, 2 min,5 min, 8 min, 20 min, 60 min and 90 min in the nutrientMedium-199, which contained 1.87 mM stable ^Ca (^CaC^,98.78% ^Ca enrichment; Oak Ridge National Laboratory,TN). The medium with stable Ca was prepared bysupplementing 1.87 mM ^CaC^ in calcium-free Medium-199(GIBCO). After the desired treatments, the cells wereprepared for ion microscopic analysis by a cryogenic method(Chandra et al. 1986). Briefly, this method involved flashfreezing of the cells in Freon-22 slush, freeze-fracturing by asandwich technique, and freeze-drying at —90°C for 24 h priorto analysis in the ion microscope. Such a fracture produces afew randomly scattered areas containing a group of cellstogether where the fracture plane has split the apical side ofthe plasma membrane, thereby exposing the cell interiors fordirect analysis. The fractured cells, therefore, consist ofessentially the whole cells except for the outer leaflet of theapical membrane. This method preserves the native elementaland morphological integrity (Ausserer et al. 1988), has beenchecked for SIMS matrix effects (Ausserer et al. 1988;Chandra et al. 1987), and has been used for imaging Na+/K+

transport in cell cultures after the inhibition of Na+, K+-ATPase with ouabain (Chandra and Morrison, 1985) andelemental quantitation in cell cultures (Ausserer et al. 1989).

Ion microscopy and image processingA CAMECA IMS-3f stigmatic ion microscope, operating witha 5.5 keV mass-filtered 100 nA O2

+ primary ion beam (spotsize of about 70 fim), was used to monitor positive secondaryions. The primary ion beam was directed onto a 250 fim2

raster. The secondary ions were extracted from a circularregion centered within the raster. A 60 fim contrast aperturewas employed in the imaging mode. Seven successive imagesof the isotopes with masses of 12, 39, 23, 40, 44, 40 and 44(recorded in this order) revealed the distribution of 12C, 39K,23Na, ^Ca, ^Ca, '"'Ca and ^Ca, respectively, in freeze-fractured freeze-dried LLC-PK, cells. By recording twoimages each of ^Ca and ^Ca, a clean sampling of the cellnucleus (or Golgi) was ensured. After the cell nucleus wasseen in the first '"'Ca image, the following '"'Ca and ^Caimages were used for quantitative measurements. Thisapproach eliminated any possibility of sampling the overlay-

ing cell cytoplasm along with the cell nucleus or the Golgicomplex. An approximately 0.3 fim thin slice of the cellularmaterial was eroded for integration of the individual ^Ca (or^Ca) image. Isotope images were digitized directly from themicrochannel plate/fluorescent screen detector assembly ofthe ion microscope using a charge-coupled device (CCD)imager (Photometries), and digitized to 14 bits per pixel by aPhotometries camera controller (Mantus and Morrison,1990). Image processing was performed using a PDP 11/34-based system and a secondary ion mass image processingprogram (Ling et al. 1987).

High mass-resolution analysisHigh mass-resolution analyses (m/Aw=5000) were performedwith the CAMECA IMS-3f ion microanalyzer to check formass-interference components in the positive secondary ionsignals from ^Ca, Ca, 39K and 23Na. A 2 fiA primary ionbeam of O2

+ (spot size of about 90 fim) was directed onto a250 fim square raster, and a 750 fim field aperture was used inthe spectrometer for high-mass-resolution analyses. Freeze-fractured freeze-dried cells grown on germanium were usedfor these analyses. The purity of the " ^ a positive secondaryion signal was 94 ± 6% (mean ± s.d., n=6, 2 samples fromcontrol cells and 4 samples from 20 min of calcium exchangetreatment). The purity of ^Ca positive secondary ion signalwas 94 ± 3% (n=6 samples, 3 samples from control cells and 3from 20 min of calcium exchange treatment). Over 99%purity was observed for 39K+ and 23Na+ signals. Theseobservations imply that low-mass-resolution imaging reliablyreveals intracellular distribution of these isotopes. Silicon, atypical substratum used in ion microscopy, could not be usedin this study because it interfered at mass 44, due to theformation of 44(SiO)+ species under the oxygen bombard-ment. In a previous study stable ^Ca could not be used as atracer for calcium, most probably because of a siliconsubstratum or the plastic impregnation used in preparing thesamples (Burns, 1984). Recently, stable ""Ca has beensuccessfully used as a tracer for calcium transport inphysiological studies (Chandra et al. 1990; Levi-Setti et al.1985).

Results

Subcellular imaging of calcium exchange in LLC-PKjcells is shown in Fig. 1. The top row represents cellsgrown in the nutrient medium that contained 1.87 mMnaturally occurring calcium. In naturally occurringcalcium the major isotope is '"'Ca (96.94%) and theminor isotopes are ^Ca (2.09%), and 41Ca, 4iCa, 43Ca,^Ca and ^Ca, which together comprise less than 1.0%of naturally occurring calcium. Sister cells (cells grownunder identical conditions) were exposed to the nutri-ent medium that contained 1.87 mM stable '"Ca(98.78% enrichment) for 1 min (second row), 8 min(third row), and 60 min (fourth row, the bottom row).

Ion microscopic images revealing intracellular distri-bution of '"'Ca, ^Ca, K and 23Na isotopes from thecells grown in the nutrient medium with naturallyoccurring calcium (top row, Fig. 1) illustrate a spatialheterogeneity for Ca distribution (three cells are inthe field of view). A perinuclear region with higher ^Caintensities has recently been identified as the Golgiregion (G) by analyzing the same rapidly frozen freeze-fractured freeze-dried cells, first with laser scanning

Subcellular imaging of calcium-calcium exchange 419

confocal fluorescence microscopy (LSCFM) for Golgilocalization, and then with ion microscopy for intra-cellular calcium distribution; a high calcium perinuclearregion was identified as the Golgi apparatus in LLC-PK! cells, 3T3 fibroblast and L5 rat myoblast cells(Chandra et al. 1991). It was also observed that thecalcium sequestered in the Golgi could be released bycalcium ionophore A23187 (Chandra et al. 1991). Amuch higher spatial resolution in the X, Y and Zdirections is required to resolve the calcium content ofindividual membranes within the Golgi complex.Therefore, our measurements underestimate the actuallocal calcium concentration in the Golgi. However, theestimated distributions of Ca amongst the three speci-fied cellular regions are not significantly degraded bythe resolution limitations.

The cell nucleus (N) in each cell has less 40Ca than theGolgi or the remaining cell cytoplasm (C) (the top rowin Fig. 1). It should be noted that the observations madeon the cell cytoplasm (excluding Golgi) are representedby a homogenized ion microscopic sampling of smallercytoplasmic structures such as mitochondria, endoplas-mic reticulum, etc. and cytoskeleton and cytosol.Calcium concentration in this region, therefore, is notdirectly comparable to these distinguishable smallercytoplasmic structures, which are below the spatialresolution of the ion microscope.

The amount of 40Ca in naturally occurring calcium is46.38 times higher than ^Ca. The ^Ca image wasenhanced for visualization. As expected, the naturalabundance of ^Ca is reflected in the image; forinstance, higher perinuclear 40Ca and ^Ca intensitiesare in the same spatial correspondence (the top row inFig. 1). The isotope image of 39K reveals a relativelyhomogeneous distribution of intracellular K. The 23Naand K images were processed for a direct visualcomparison of intracellular intensities of these elementsand, therefore, the 23Na image appears dark (Chandraet al. 1986). The K/Na ratio was about 9 in these cells.The cell peripheries have high levels of 23Na, an artifactdue to the adherence of the extracellular Na from thenutrient medium (Chandra et al. 1986). Imaging of Kand Na along with other isotopes of interest provides apowerful measure of injured/damaged or dead cells, sothat these cells can be discarded from the study(Chandra and Morrison, 1992; Chandra et al. 1986).

High-mass-resolution analyses indicated that positivesecondary-ion signals of ^Ca, ^Ca, 39K and 23Na werefree from any significant mass interference, thereforelow-mass-resolution imaging data reliably reveal intra-cellular distributions of these isotopes. Quantitativeanalysis using a method for calculating the totalconcentration of an element from the isotope images(Ausserer et al. 1989) indicated that in LLC-PKj cellsthe cell nucleus contained 161±44, the Golgi apparatus362±97, and the remaining cell cytoplasm 304±99 jug/gdry weight (mean ± s.d., n=20 cells) total calcium(which accounts for both free and bound, and mem-brane sequestered calcium). These measurements arein close agreement with our recent observations on thiscell line (Chandra et al. 1991) and are typical of the

majority of cells. Assuming 85% cell water, the dryweight total calcium concentrations could be used toestimate the wet weight molar concentrations. The cellnucleus contained 0.60 mM, the Golgi apparatus 1.36mM, and the remaining cell cytoplasm 1.14 mM totalcalcium. From these concentrations it can be estimatedthat the total Ca concentration in the cytoplasm is about10,000 times higher than the free ionized [Ca2+]; inresting cells, which has been measured to be about 10~7

M in LLC-PK, cells (Bonventre and Cheung, 1986). Inorder to estimate the contribution of the individualcompartment to the total cell calcium, the volumefractions of the nucleus, the Golgi region and theremaining cell cytoplasm were measured by laserscanning confocal fluorescence microscopy in LLC-PK!cells in a recent study (Chandra et al. 1991). Thenucleus occupied 18.5%, the Golgi region 3.3% and theremaining cytoplasm 78.2% of the total cell volume. Onthe basis of the volume fractions and the total Caconcentration in each compartment, it can be estimatedthat the nucleus contributes about 11%, the Golgiregion about 4% and the remaining cytoplasm about85% to the total cell calcium.

The cells exposed to the nutrient medium containing1.87 mM ^Ca for 1 min (second row, Fig. 1), 8 min(third row) and 60 min (the bottom row) reveal agradual increase in ^Ca and a decrease in Ca in thecells with increase in time. The exchange in the Golgiregion up to 1 and 8 min is slower than the cytoplasm as40Ca in this region takes longer to exchange with ^Ca.At 60 min the perinuclear Golgi region is marked byhigh levels of Ca. This pattern of calcium exchangewas consistent in three separate trials on LLC-PKj cells.

For an unequivocal confirmation that the perinuclearregion is the Golgi apparatus, the LLC-PK4 cells werestained with C6-NBD-ceramide (Lipsky and Pagano,1985; Pagano and Sleight, 1985) for labeling the Golgiapparatus and then allowed to exchange ^Ca for 45 minby exposing the cells to the nutrient medium with stable

Ca. The cells were then cryogenically prepared andanalyzed first with LSCFM for Golgi localization andthen with ion microscopy for isotopic localization(Chandra et al. 1991). Fig. 2 shows correlative optical,fluorescence, and ion microscopy of the same cells. Thecomposite image shows a reflected light Nomarskimicrograph of several fast-frozen freeze-fracturedfreeze-dried LLC-PKi cells (Fig. 2a), a LSCFM imageof the C6-NBD-ceramide fluorescence that indicates thelocation of the Golgi in the same individual cells (Fig.2b), and ion microscopic isotopic images of ^Ca (Fig.2c) and ^Ca from identical cells (Fig. 2d). Thenumbered cells can be easily matched between theoptical, fluorescence and ion microscopic images. It isalso evident that after 45 min of calcium exchange, thelocation of the Golgi region in the numbered cells canbe clearly matched between the Golgi region revealedby LSCFM (Fig. 2b) and a high-intensity perinuclearregion in ^Ca ion microscopy image (Fig. 2d).

Quantitative measurements from the isotope imagedata are shown in Table 1. Table 1 contains additionaltime points in the calcium exchange that are not shown

420 S. Chandra and others

in Fig. 1. The ratio of ^Ca/^Ca was calculated from thecell nucleus, the Golgi region, and the remaining cellcytoplasm, and the percentage calcium exchange wasthen calculated, on the basis of this ratio, and is shown

in Fig. 3. In cells exposed only to the medium thatcontained 1.87 mM naturally occurring calcium (0 minof stable ^Ca exposure), the ratio of ^Ca to ^Ca is46.38. The calculated ^Ca/^Ca ratios from the isotope

Subcellular imaging of calcium-calcium exchange 421

Fig. 1. SIMS images revealing intracellular distribution of40Ca, ^Ca, 39K and 23Na isotopes in cryogenically preparedfreeze-fractured freeze-dried LLC-PKt porcine kidneyepithelial cells. The first row represents cells grown inmedium that contained 1.87 mM naturally occurringcalcium. The cells in the second, the third and the fourthrows were exposed to the nutrient medium that contained1.87 mM stable ^Ca for 1, 8 and 60 min, respectively.Brightness indicates relative ion intensity within an image.Isotope images were recorded on a slow scan CCD cameradirectly from the fluorescent screen of the ion microscope,and digitized to 14 bits per pixel. The raw images werescaled down (divided) by a factor to display the imagebetween 0 to 255 gray levels. The natural abundance of^Ca is 2.09%. This contribution was corrected from themass 44 image by digitally registering (overlaying) andsubtracting the 2.09% fraction of a mass 40 image from themass 44 image. The isotopic composition of stable ^Ca wasas follows: 44=98.78%; 40=1.12%; 42, 43, 46 and 48=< 0.06% each. The 1.12% contribution from the mass 40image was also corrected as described above. Because ofvery high enrichment of stable MCa and low naturalabundance of this isotope, these corrections were minorand made no significant change in images or calciumexchange rate. The images of ^Ca and '"'Ca wereprocessed for optimal image quality for the visualization ofcalcium exchange. Quantitative measurements for thisexchange are presented in Table 1. Within a treatment, the^Na image was normalized to the 39K image by keepinginstrumental, image integration time and photographicconditions identical between these isotopes for a directvisual comparison of the intensities of sodium andpotassium. ^Ca and ^Ca images were integrated on theCCD camera for 120 s, and 39K and 23Na images for 0.2 s.Bar, 15 fim.

image data in three cellular compartments are inexcellent agreement with this value (Table 1). At 1 minof exposure to ^Ca, the three compartments aresignificantly different (P < 0.01) in their ^Ca/^Caratio. At 2, 5, 8 and 20 min the ^Ca/^Ca in thecytoplasm was significantly different (P < 0.01) fromthe Golgi or the nucleus, which were not different. At

Table 1. The ratio of40Ca/44Ca isotope imageintensities (mean+s.d.) in the cell nucleus, the Golgiregion and the remaining cell cytoplasm in LLC-PKj

cells

Time*(min)

01258

206090

Nucleus

44.7±9.04.07±0.862.34±0.752.04+0.371.56±0.330.94±0.150.75±0.060.73±0.06

""Ca/^Ca

Golgi region

43.0±7.65.53+1.562.63+0.982.29±0.401.73±0.320.84±0.190.47±0.070.42±0.05

Cytoplasm

49.2±7.82.68±0.591.78±0.671.29±0.300.95±0.160.68±0.090.55±0.080.49±0.04

No. ofcells

2021202114232524

*The time of exposure of the cells to the nutrient medium with1.87 mM stable '"Ca. Two-way analysis of variance with blockingon cells (randomized complete block design) was used for thestatistical analysis (Snedecor and Cochran, 1989). At time 0, thethree cellular compartments were not significantly different.

60 min of Ca exposure of the cells, all threecompartments were significantly different (P < 0.01) in^Ca/^Ca. At 90 min of ^Ca exposure of the cells theratio of ^Ca/^Ca in the nucleus was significantlydifferent (P < 0.01) from the Golgi and the cytoplasm,which were not different from each other.

Fig. 3 graphically depicts the percentage calciumexchange in the three cellular compartments withincreased time of exposure of cells to ^Ca; the insetmagnifies the early time points of exchange. On thebasis of the average rate of calcium exchange %exchange per min), it is estimated that for up to 2 min ofcalcium exchange the exchange rates in the nucleus, theGolgi region and the cytoplasm were 15.0, 13.8 and18.0% per min, respectively. Between 2 and 20 min ofcalcium exchange, the exchange rate in each compart-ment dropped about an order of magnitude as com-pared to the exchange rates up to 2 min. The calciumexchange rates between 20 and 90 min of exchange ineach compartment were also about an order ofmagnitude less than the exchange rates between 2 and20 min. These observations suggests that all threecompartments have a rapidly exchanging pool ofcalcium, which takes about 2 min to exchange. Amoderately rapidly exchanging pool of calcium existsbetween 2 and 20 min of calcium exchange. The cellsexchanged about 50% of their total calcium in less than20 min. A slowly exchanging pool of calcium isindicated at time points after 20 min of calciumexchange. Even after 90 min of ^Ca exposure of thecells, approximately 30% calcium in the cytoplasm andthe Golgi, and 40% in the nucleus, remained unex-changed. It is plausible that after 20 min of calciumexchange the main contribution to the exchange is fromrelatively tightly bound intracellular calcium.

Estimates of the magnitude of calcium exchange inthe three intracellular compartments were calculated onthe basis of percentage calcium exchange and total Caconcentration in individual compartments (Table 2).Only the mean values from both measurements wereused for these calculations (i.e. s.d. of the mean valuewas ignored), and concentrations are expressed in wetweight (micromolar). The calcium exchange in indi-vidual compartments is also expressed as percentagetotal cell calcium that accounts for volume fractions ofthe individual compartment.

Table 2. Estimates of intracellular calcium exchangein the nucleus, the Golgi region and the remaining

cytoplasm expressed in wet weight micromolar (fiM)concentrations and % of the total cell calcium

Nucleus Golgi Cytoplasm

Calcium exchange up to 2 min (fiM) 180 376 410% of the total cell calcium 3.2 1.2 30.6

Calcium exchange up to 20 min (/iM) 309 734 678% of the total cell calcium 5.4 2.3 50.6

Calcium exchange up to 90 min (,uM) 347 957 765% of the total cell calcium 6.1 3.0 73.0

422 S. Chandra and others

Fig. 2. Correlative optical,laser scanning confocalfluorescence and ionmicroscopy on the samefrozen-fractured freeze-driedcells. Live cells grown on thegermanium substratum werestained with Q-NBD-ceramide(Chandra et al. 1991). Thecells were then exposed to thenutrient medium thatcontained 1.87 mM ^Ca.Following the treatment thecells were fast frozen, freeze-fractured and freeze-dried asdiscussed in Materials andmethods. An Olympus lightmicroscope was used in areflected light Nomarski modefor photographing the cells. ABioRad LSCFM instrument(Wells et al. 1990) was usedfor imaging the Q-NBD-ceramide fluorescence infreeze-dried cells prior toanalysis with the ionmicroscope. Three cells havebeen numbered for the ease ofmatching the same cellbetween a reflected lightNomarski (a); Q-NBD-ceramide fluorescence imagethat indicates the Golgiapparatus in same cells (b);

and 40Ca (c) and ^Ca (d) ion microscopic images. These observations were confirmed on at least 20 cells. 40Ca and ^Caimages were integrated on the CCD camera from the fluorescent screen of the ion microscope for 120 s. The K/Na rato inthese cells was about 9 (images not shown). Isotope images were processed by digital image processing and printed foroptimum image quality. Bar, 15 /zm.

<DCDC

oX

a>E3

lei

S5?

o•CO

a>a.oou>

Eok—

20

BU

60

40

20

-

-,C

•cOl

3

A

;

o

a] A

DA

o

o

aA

1

Nuclear RegionGolgi regionCytoplasm

0

AD

1

60

50

40

30

20

10

°Ci

_

o

- DA

1

O

s

1

2

i •

A

D

1

i4

i

0

i ,

I

1

6

1 i

o

2

8i

AO

a

-

-

-

1

-

•

-

-

-

0 .

40 60

Time (min)

80 100

Fig. 3. % Calcium exchange inthe nucleus, the Golgi regionand the remaining cytoplasmcalculated from the meanisotope ratio of 40Ca/MCa ateach time point. In the insetthe early time points aremagnified. The % standarddeviation observed atindividual time points of•^Ca/^Ca ratio also apply tothe respective time points of% calcium exchange.

Subcellular imaging of calcium-calcium exchange 423

Discussion

The use of stable ^Ca in conjunction with ionmicroscopy provides a powerful approach for calciumexchange and transport studies. At present no othertechnique offers imaging capabilities for endogenouscalcium separately from the calcium that enters the cellfrom extracellular stores. One can also use other stableisotopes of calcium such as 42Ca, ^Ca, 48Ca, etc. inaddition to ^Ca, in order to image the differentresponses of calcium to the separate treatments. ^Ca isthe most cost effective (about $8 per mg) among all thestable calcium isotopes. A recent study has utilizedstable 42Ca as a tracer to study calcium exchange in theretina after 30 min and 6 h of 42Ca exposure (Burns,1991). Although subcellular details could not beresolved in this study, it was observed that by 30 min theretinal layers had exchanged about 75% of theirendogenous calcium with 42Ca.

Since ion microscopy is a high-vacuum technique,cryogenic procedures that preserve the native chemicaland morphological integrity of the cells must be used forthe sample preparation. The sandwich freeze-fracturesample preparation procedure used in this study(Chandra et al. 1986) has been checked for thesecriteria (Ausserer et al. 1988; Chandra and Morrison,1985; Chandra and Morrison, 1992). In addition, thismethod is quite rapid and takes about 30 s per samplefor sample preparation. However, the samples must befreeze-dried prior to ion microscopic analysis.

Our observations indicate that calcium to calciumexchange can be directly imaged at a subcellularresolution by using a stable isotope in the nutrientmedium and the technique of ion microscopy. On thebasis of the kinetics of this exchange, we were able toestimate the magnitude of rapidly exchangeable, mod-erately rapidly exchangeable and slowly exchangeablepools of intracellular calcium in the nucleus, the Golgiregion and the remaining cytoplasm of LLC-PKi cells.

The ability of the cell nucleus to exchange asignificant portion of its calcium rapidly (180 [iM in 2min and 309 pM in 20 min; Table 2) suggests that aloosely bound pool of calcium is stored in the nucleus. Itshould be noted that ion microscopic isotopic imagingdata are based on a slice through the nucleus and theother spatially resolved compartments and, therefore,individual compartments are free from any homogeniz-ation effects such as the contribution of the overlyingcytoplasm to the nucleus signals. It is difficult tospeculate on the nature of the calcium-storing sites inthe nucleus, since there are no known intranuclearmembrane structures that can be assigned for seques-tering calcium. The majority of the calcium stored in thenucleus has to be in the bound form, otherwiseintranuclear ionized Ca2+ would be very high. It isplausible that the calcium-binding proteins (Bachs andCarafoli, 1988; Harper et al. 1980; Schibeci andMartonosi, 1980) and the other nuclear material mayplay a role in storing calcium, a portion of which isstored in a loosely bound form. In isolated liver nuclei,

an inositol 1,4,5-trisphosphate-sensitive (Ins(l,4,5)P3)Ca2+ pool has been demonstrated (Pierluigi et al.1990). In order to clearly assign an Ins(l,4,5)P3-sensitive Ca2+ pool in the cell nucleus, direct obser-vations are needed on the nuclei of intact cells.

The Golgi apparatus exchanged about 376 juMcalcium in 2 min and about 734 /iM calcium in 20 min(Table 2). Because of the small volume of the Golgi, thecontribution of the Golgi apparatus to the overallcalcium exchange on a total cell basis is small.Observations on isolated Golgi vesicles have indicatedthat the Golgi apparatus may play a role in intracellularCa2+ homeostasis in human neutrophils (Krause andLew, 1987). The Golgi apparatus may provide apathway for removal of Ca2+ from cells by exocytosis ofsecretory vesicles originating in the Golgi, and if thisprocess is efficient it would affect the calcium exchangerate of the cell cytoplasm. Further observations areneeded to understand these processes and to evaluatethe role of the Golgi apparatus in cellular Ca2+

homeostasis.The cell cytoplasm (excluding Golgi) exchanged

about 410 fiM calcium (30.6% of the total cell calcium)in 2 min and about 678 [iM calcium (50.6% of the totalcell calcium) in 20 min (Table 2). On the basis of thefraction of the total cell calcium, the cytoplasm containsthe largest pool of exchangeable calcium. Smallerorganelles that sequester calcium, such as the endoplas-mic reticulum, mitochondria, recently discovered cyto-plasmic structures called "calciosomes" that contain thecalcium-binding protein calsequestrin (Volpe et al.1988), and a variety of other calcium-binding proteinsincluding calmodulin (Heizmann and Hunziker, 1991),may play a role together in this exchangeable pool ofcytoplasmic calcium. To separate different cytoplasmiccompartments for their calcium storage and exchange-ability in individual cells is a challenging task that isbeyond the capability of current microanalytical tech-niques. Although electron probe X-ray microanalysisoffers a better spatial resolution than ion microscopy, itcannot distinguish between the isotopes and, therefore,is unsuitable for the present study. At present, ionmicroscopy is the only technique for isotopic imagingstudies of single cells at a subcellular resolution.

This study is a direct demonstration of calciumexchange at a subcellular level. The unique ability ofion microscopy to image intracellular calcium separ-ately from the influx of calcium from the extracellularpool, and at the same time providing measurements forevaluating the health status of the cell, has enormouspotential for studies of calcium transport in spatiallyresolved single cells under physiological, pathologicaland toxicological conditions. The implications of thisapproach are not limited to resting cells alone andshould be realized in elucidating agonist-induced influxand efflux of calcium at a subcellular resolution in bothexcitable and non-excitable cells.

The National Institutes of Health and the Department of

424 S. Chandra and others

Energy are thanked for funding our biological ion microscopyprogram at Cornell. The support for the CAMECA IMS-3fion microscope was provided by the National ScienceFoundation. E.P.W. Kable and K. Hodges are thanked fortheir assistance in culturing cells. E.P.W. Kable is alsothanked for making measurements with laser scanningconfocal microscope. C.M.S. Fewtrell, C.S. Fullmer, D.A.Holowka, P.J. Millard and D.R. Sandison are thanked forconstructive criticisms of the manuscript. The CornellNIH/NSF Developmental Resource for Biophysical Imagingand Optoelectronics was used in culturing the cells.

References

Ausserer, W. A., Chandra, S. and Morrison, G. H. (1988).Morphological and elemental integrity of freeze-fractured, freeze-dried cultured cells during ion microscopic analysis. J. Microsc.(Oxford) 154, 39-57.

Ausserer, W. A., Ling, Y.-C, Chandra, S. and Morrison, G. H.(1989). Quantitative imaging of boron, calcium, magnesium,potassium and sodium distributions in cultured cells with ionmicroscopy. Analyt. Chem. 61, 2690-2695.

Bachs, O. and Carafoli, E. (1987). Calmodulin and calmodulin-binding proteins in liver cell nuclei. J. Biol. Chem. 262, 10786-10790.

Bonventre, J. V. and Cheung, J. Y. (1986). Cytosolic free calciumconcentration in cultured renal epithelial cells. Amer. J. Physiol.250, F329-F338.

Borle, A. B. (1972). Kinetic analysis of calcium movements in cellcultures V. intracellular calcium distribution in kidney cells. J.Membr. Biol. 10, 45-66.

Brundage, R. A., Fogarty, K. E., Tuft, R. A. and Fay, F. S. (1991).Calcium gradients underlying polarization and chemotaxis ofeosinophils. Science 254, 703-706.

Burns, M. S. (1982). Application of secondary ion mass spectrometry(SIMS) in biological research: a review. J. Microsc. (Oxford) 127,237-258.

Burns, M. S. (1984). Selection of calcium isotopes for secondary ionmass spectrometric analysis of biological material. J. Microsc.(Oxford) 135, 209-212.

Burns, M. S. (1992). Stable isotope analysis of ion movementin retina. Proc. Eighth Int. Conf. Secondary Ion MassSpectrometry (SIMS VIII), Amsterdam, September 15-20, 1991.(in press).

Castaing, R. and Slodzian, G. (1962). Microanalyse par emissionionique secondaire. J. Microscopic 1, 395-410.

Chandra, S., Ausserer, W. A. and Morrison, G. H. (1987).Evaluation of matrix effects in ion microscopic analysis of freeze-fractured, freeze-dried cultured cells. J. Microsc. (Oxford) 148,223-239.

Chandra, S., Fullmer, C. S., Smith, C. A., Wasserman, R. H. andMorrison, G. H. (1990). Ion microscopic imaging of calciumtransport in the intestinal tissue of vitamin D-deficient and vitaminD-replete chickens: a ^Ca stable isotope study. Proc. Nat. Acad.Sci. USA 87, 5715-5719.

Chandra, S., Kable, E. P. W., Morrison, G. H. and Webb, W. W.(1991). Calcium sequestration in the Golgi apparatus of culturedmammalian cells revealed by laser scanning confocal microscopyand ion microscopy. J. Cell Sci. 100, 747-752.

Chandra, S. and Morrison, G. H. (1985). Imaging elementaldistribution and ion transport in cultured cells with ion microscopy.Science 228, 1543-1544.

Chandra, S. and Morrison, G. H. (1988). Ion microscopy in biologyand medicine. Methods Enzymol. 158, 157-179.

Chandra, S. and Morrison, G. H. (1992). Sample preparation ofanimal tissues and cell cultures for secondary ion massspectrometry (SIMS) microscopy. Biol. Cell. 74, 31-42.

Chandra, S., Morrison, G. H. and Wolcott, C. C. (1986). Imagingintracellular elemental distribution and ion fluxes in cultured cells

using ion microscopy: a freeze-fracture methodology. J. Microsc.(Oxford) 144, 15-37.

Fassel, V. A. and Kniseley, R. N. (1974). Inductively coupled plasmas.Analyt. Chem. 46, 1155A-1164A.

Finch, E. A., Turner, T. J. and Goldin, S. M. (1991). Calcium as acoagonist of inositol, 1, 4, 5-trisphosphate-induced calcium release.Science 252, 443-446.

Galle, P. (1986). Tissue microlocalization of isotopes by ionmicroscopy and by microautoradiography. Proc. Fifth Int. Conf.Secondary Ion Mass Spectrometry (SIMS V) (ed. Benninghoven,A., Colton, R. J., Simon, D. S. and Werner, H. W.), pp. 495-497.Springer-Verlag, Berlin.

Harper, J. F., Cheung, W. Y., Wallace, R. W., Huang, H.-L., Levine,S. N. and Steiner, A. L. (1980). Localization of calmodulin in rattissues. Proc. Nat. Acad. Sci. USA 77, 366-370.

Heizmann, C. W. and Hunziker, W. (1991). Intracellular calcium-binding proteins: more sites than insights. Trends Biochem. Sci. 16,98-103.

Hernandez-Cruz, A., Sala, F. and Adams, P. R. (1990). Subcellularcalcium transients visualized by confocal microscopy in a voltage-clamped vertebrate neuron. Science 247, 858-862.

Krause, K.-H. and Lew, P. D. (1987). Subcellular distribution of Ca2+

pumping sites in human neutrophils. J. Gin. Invest. 80, 107-116.Kuwata, J. H. and Langer, G. A. (1989). Rapid, non-perfusion-

limited calcium exchange in cultured neonatal myocardial cells. J.Mol. Cell. Cardiol. 21, 1195-1208.

Langer, G. A., Frank, J. S., Nudd, L. M. and Seraydarian, K. (1976).Sialic acid: effect of removal on calcium exchangeability of culturedheart cells. Science 193, 1013-1015.

Levi-Setti, R., Crow, G. and Wang, Y. L. (1985). High-resolutiontopographic and isotopic imaging with a 40 KeV Ga + scanning ionmicroprobe. Proc. 20th Annu. Conf. of the Microbeam AnalysisSociety (Armstrong, J. T., ed.), pp. 209-218. San Francisco Press,Inc., Microbeam Analysis.

Ling, Y-C, Bernius, M. T. and Morrison, G. H. (1987). SIMIPS:secondary ion mass image processing system. J. Chem. Inf.Comput. Sci. 27, 86-94.

Linton, R. W. (1986). Biological ion microanalysis using SIMS: areview. Proc. Fifth Int. Conf. Secondary Ion Mass Spectrometry(SIMS V) (ed. Benninghoven, A., Colton, R. J., Simons, D. S. andWerner, H. W.), pp. 420-425. Springer-Verlag, Berlin.

Lipsky, N. G. and Pagano, R. E. (1985). A vital stain for the Golgiapparatus. Science 228, 745-747.

Mantus, D. S. and Morrison, G. H. (1990). Ion image detection with amicrochannel plate evaluated by using a charge coupled devicecamera. Analyt. Chem. 62, 1148-1155.

McGuffee, L. J., Hurwitz, L., Little, S. A. and Skipper, B. E. (1981).A 45Ca autoradiographic and stereological study of freeze-driedsmooth muscle of the Guinea Pig vas deferens. J. Cell Biol. 90, 201-210.

Morrison, G. H. and Slodzian, G. (1975). Ion microscopy. Analyt.Chem. 47, 932A-943A.

Pagano, R. E. and Sleight, R. G. (1985). Denning lipid transportpathways in animal cells. Science 229, 1051-1057.

Pierluigi, N., Orrenius, S., Nilsson, T. and Berggren, P.-O. (1990).An inositol 1, 4, 5-trisphosphate-sensitive Ca2+ pool in the livernuclei. Proc. Nat. Acad. Sci. USA 87, 6858-6862.

Sampson, H. W., Matthews, J. L., Martin, J. H. and Kunin, A. S.(1970). An electron microscopic localization of calcium in the smallintestine of normal, rachitic, and vitamin-D-treatcd rats. Calcif.Tiss. Res. 5, 305-316.

Schibeci, A. and Martonosi, A. (1980). Ca2+-binding proteins in thenuclei. Eur. J. Biochem. 113, 5-14.

Snedecor, G. W. and Cochran, W. G. (1989). Statistical Methods, 8thedn, pp. 255-264. Iowa State University Press, Ames, Iowa.

Thevenod, F., Dehlinger-Kremer, M., Kemmer, T. P., Christian, A.-L., Potter, B. V. L. and Schultz, I. (1989). Characterization ofinositol 1, 4, 5-trisphosphate sensitive (IsCaP) and -Insensitive(IisCaP) nonmitochondrial Ca2+ pools in rat pancreatic acinarcells. J. Membr. Biol. 109, 173-186.

Tsien, R. Y. and Poenie, M. (1986). Fluorescence ratio imaging: a new

Subcellular imaging of calcium-calcium exchange 425

window into intracellular ionic signaling. Trends Biochem. Sci. 11, Wells, K. S., Sandison, D. R., Strickler, J. and Webb, W. W. (1990).450-455. In Quantitative Fluorescence Imaging with Laser Scanning Confocal

Volpe, P., Krause, K-H., Hashimoto, S., Zorzato, F., Pozzan, T., Microscopy, Handbook of Biological Microscopy. Plenum, NewMeldolesi, J. and Lew, D. P. (1988). "Calciosome," a cytoplasmic York,organelle: the inositol 1, 4, 5-trisphosphate-sensitive Ca store ofnonmuscle cells? Proc. Nat. Acad. Sci. USA 85, 1091-1095. (Received 18 February 1992 - Accepted 6 April 1992)