Journal of Cell Science 108, 1895-1909 (1995) 1895
Printed in Great Britain © The Company of Biologists Limited 1995
Direct visualization of a vast cortical calcium compartment in Paramecium by
secondary ion mass spectrometry (SIMS) microscopy: possible involvement
Nicole Stelly1,*, Sylvain Halpern2, Gisèle Nicolas3, Philippe Fragu2 and André Adoutte1
1Laboratoire de Biologie Cellulaire 4 (CNRS, URA 1134), Bâtiment 444, Université Paris-Sud, 91405 Orsay Cedex, France
2Equipe de Microscopie Ionique (INSERM U66), Institut Gustave Roussy, 94800 Villejuif, France
3Centre Interuniversitaire de Microscopie Electronique (CNRS, URA 1488) et Laboratoire de Cytologie, Université Pierre et Marie
Curie, 7 quai St Bernard, Bâtiment A, 75252 Paris Cedex 05, France
*Author for correspondence
The plasma membrane of ciliates is underlaid by a vast particular, the emitting zone was still seen in mutants
continuous array of membrane vesicles known as cortical totally lacking trichocysts, the large exocytotic organelles
alveoli. Previous work had shown that a puriﬁed fraction docked at the cell surface, indicating that they make no
of these vesicles actively pumps calcium, suggesting that major direct contribution to the emission. Calcium con-
alveoli may constitute a calcium-storage compartment. centration within alveoli was quantiﬁed for the ﬁrst time in
Here we provide direct conﬁrmation of this hypothesis SIMS microscopy using an external reference and was
using in situ visualization of total cell calcium on sections found to be in the range of 3 to 5 mM, a value similar to
of cryoﬁxed and cryosubstituted cells analyzed by SIMS that for sarcoplasmic reticulum. After massive induction of
(secondary ion mass spectrometry) microscopy a method trichocyst discharge, this concentration was found to
never previously applied to protists. A narrow, continuous, decrease by about 50%, suggesting that the alveoli are the
Ca-emitting zone located all along the cell periphery was main source of the calcium involved in exocytosis.
observed on sections including the cortex. In contrast, Na
and K were evenly distributed throughout the cell. Various
controls conﬁrmed that emission was from the alveoli, in Key words: calcium, SIMS, Paramecium
INTRODUCTION electron probe microanalysis of tissue sections (Somlyo, 1985;
Andrews et al., 1987), and an indirect one, seeking to identify
The question of the subcellular location of calcium (Ca) stores the organelles by virtue of the presence of a set of ‘marker’
in eukaryotic cells has attracted considerable attention in recent proteins involved in Ca2+ homeostasis: Ca2+-ATPases, Ca2+
years (see Koch, 1990; Meldolesi et al., 1990; Tsien and Tsien, channels (ryanodine receptor, inositol-triphosphate (InsP3)
1990; Lytton and Nigam, 1992; Meldolesi and Villa, 1993, for receptor) and Ca-binding proteins (calsequestrin, calreticulin,
review). The measurement of free, cytosolic calcium (Ca2+) etc.). These proteins were detected either, by physiological and
concentration and its oscillation in single cells has become biochemical methods (such as by measurement of Ca2+ uptake
possible with the advent of ﬂuorescent probes (Grynkiewicz et and release in subcellular fractions; see Pietrobon et al., 1990
al., 1985; see Williams and Fay, 1990, for review). This con- for review) or by immunocytological approaches (subcellular
centration, however, is several orders of magnitude lower than localization, at the EM level, of compartments reacting with
that of total cellular calcium; part of this large amount of antibodies directed against the marker proteins; e.g. see Volpe
calcium is bound to cytosolic proteins but the majority, by far, et al., 1988).
appears to be segregated inside membrane-bounded intracellu- The results of these approaches are summarized in recent
lar organelles where it is complexed with low-affinity, high- reviews (cited above). Except for the striated muscle cell, an
capacity proteins (Carafoli, 1987; Koch, 1990). Release of extensively studied model system in which the situation is con-
Ca2+ from these organelles plays a key role in several siderably clariﬁed because of the ampliﬁcation of the Ca
processes, especially in the response to various extracellular storage compartment achieved in the form of the sarcoplasmic
stimuli. reticulum, the number, location and functional role of Ca
Identiﬁcation of these Ca-sequestering organelles has storage compartments in eukaryotic cells has until recently
proven difficult, however. Two major approaches have been been unclear. Except for mitochondria, which do not appear to
used, a direct one, seeking to visualize the element itself by be involved in Ca2+ storage, the major intracellular Ca store
1896 N. Stelly and others
has, for many years, been assumed to be the endoplasmic methods used in biological microanalysis, see Linton and
reticulum. Currently, at least two and more probably three or Goldsmith, 1992).
even four types of compartments are implicated (Burgoyne and Paramecia have already been studied by microanalysis at the
Cheek, 1991; Lytton and Nigam, 1992; Sitia and Meldolesi, EM level both in the EPMA (Schmitz et al., 1985; Zierold et
1992; Meldolesi and Villa, 1993), depending on the cell type al., 1989) and the EELS mode (Knoll et al., 1993), using
analyzed, on the basis of the receptors, channels, pumps and excellent techniques of cryoﬁxation followed by either freeze-
calcium-binding proteins that they contain. All of these com- drying or cryosubstitution. The major result of these studies
partments appear to belong to the general intracellular protein was the identiﬁcation of a peripheral Ca-containing zone in the
sorting compartments (RER, Golgi, endosomes, etc.) although cortex of Paramecium (not including the exocytotic organelles
they are more or less distantly connected to it. known as trichocysts) and preliminary evidence for redistrib-
In this paper, we present an approach complementary to ution of this calcium after induction of exocytosis in the EELS
those just cited, aimed at directly visualizing total cellular Ca. high-resolution study (Knoll et al., 1993) but not in the EPMA
It involves the use of both a new cell type and a different one (Zierold et al., 1989). Here, we extend these studies using
method. The cell type is the ciliated protozoan Paramecium, the different SIMS approach, which we adapted for fast-
in which we have recently shown that a vast vesicular network swimming single cells. We conﬁrm the occurrence of an
lying just beneath the plasma membrane actively pumps Ca 2+ intensely emitting peripheral rim of calcium and, through the
(Stelly et al., 1991). This ‘primitive’ organism therefore use of mutants devoid of trichocysts, demonstrate that the exo-
appeared to offer a naturally ampliﬁed Ca storage compart- cytotic organelle is not the location of the ion. By a variety of
ment (akin to the sarcoplasmic reticulum), facilitating direct controls, we show that this rim cannot be due solely to arte-
visualization. The method is secondary ion mass spectrome- factual displacement or external adhesion of calcium. This
try (SIMS; Castaing and Slodzian, 1962) microscopy, which provides deﬁnitive conﬁrmation of the existence of a vast sub-
has been extensively used in solid state physics to character- membranal calcium compartment in this cell, in which most of
ize surface composition of samples, but much less in biolog- the cell calcium is stored (at least one order of magnitude more
ical applications (see reviews by Chandra and Morrison, than in the rest of the cytoplasm). We also provide the ﬁrst
1988; Fragu et al., 1992). A primary ion beam is focused onto quantiﬁcation of the amount of calcium by SIMS microscopy
the surface of a tissue section, leading to the sputtering of the using an internal standard; this concentration is in the mil-
most superﬁcial atoms, themselves partly in the form of ions. limolar range, equivalent to that found in the sarcoplasmic
These secondary ions are then collected, analyzed with a mass reticulum. Finally, we provide preliminary evidence that this
spectrometer and the corresponding image is reconstructed. calcium is involved in the exocytotic process, by observing a
One therefore obtains an image of the distribution of a speciﬁc 50% reduction of its amount in cells ﬁxed 15 to 30 seconds
atom at the surface of the specimen analyzed. This method after massive induction of exocytosis.
offers three main advantages. First, it allows visualization of
all the elements of Mendeleiev’s table as well as discrimina-
tion between many of their stable and radioactive isotopes. MATERIALS AND METHODS
By successively eroding the same section one can map several
different ions (for example, Ca2+, Na+, K+, etc.) in the same Biological material and sample preparation
tissue and cells. It should be stressed that ions are highly Paramecium
prone to extraction from cytological preparations during Strains and culture conditions
specimen preparation (see Mentré and Escaig, 1988). A pre- The wild-type (WT) cells used in these experiments were from stock
requisite for all the approaches just outlined is to take suitable d4-2 of Paramecium tetraurelia. Cells were grown at 27°C in
precautions to avoid loss and/or redistribution of diffusible phosphate-buffered wheat grass powder infusion, bacterized the day
compounds. This can be achieved by rapid freezing of the before use with Klebsiella pneumoniae and supplemented with 0.5
cells, avoiding the use of any ﬁxative, then either cryosubsti- µg/ml β-sitosterol.
tuting the samples (as done in the present work) or using Two mutants were used in this study: mutant tam8, whose tri-
freeze-fracturing followed by freeze-drying (Chandra and chocysts are never attached to the cell surface (Beisson and Rossignol,
1975; Lefort-Tran et al., 1981); and a thermosensitive mutant, nd9,
Morrison, 1992). Second, its sensitivity is at least as good as
whose trichocysts are attached at the cell surface but cannot be dis-
that of X-ray microanalysis (EPMA) and probably slightly charged at 27°C (Beisson et al., 1976).
better (see below). This level of sensitivity is comparable or
slightly inferior to that of electron energy loss spectrometry Microscopy
methods (EELS); Third, images of ion distribution can be Cells were harvested from early stationary phase cultures and the
obtained over large areas of cells in a short time, making the pellet was ﬁxed either: (1) in 0.5% glutaraldehyde plus 2%
method especially valuable when there is a need to observe paraformaldehyde, 50 mM sodium cacodylate buffer, pH 7.4, for 20
whole tissues or extended portions of large cells. The major minutes at 4°C, or in the same ﬁxative followed by postﬁxation in 2%
drawback of SIMS microscopy is its limited lateral resolution, OsO4 in the same buffer. Cells were then pre-embedded in ﬁbrinogen
especially when compared with EELS, being of the order of pellets, dehydrated and embedded in LR White or Epon-Araldite; or
(2) by fast-freeze ﬁxation by slamming the specimen against a cold
0.5 µm for the present instruments, yielding images equiva-
copper block cooled by liquid helium (Escaig, 1983) followed by
lent to those from a light microscope. This drawback is partly freeze substitution at −86°C for 72 hours in acetone in the presence
compensated by the sensitivity of SIMS, the very low level of 20 mM oxalic acid, then warmed to −30°C, maintained for two
of background noise and the relative ease with which data hours at −30°C, and ﬁnally warmed to room temperature and
from large areas can be collected (for a detailed comparison embedded in Epon-Araldite.
of the merits and limitations of the various microprobe Wild-type cells were also cryoﬁxed and cryodehydrated without
Calcium stores visualized by SIMS in Paramecium 1897
oxalic acid as controls and embedded in Epon-Araldite or cryoem- In our IMS 3F, the secondary ion beam intensity is measured
bedded in Lowicryl K4M. directly with the electron multiplier; the measurements are performed
on selected areas, which are limited by adapted apertures. In the
Exocytosis present work, the measured areas were always of 8 µm diameter. In
Synchronous exocytosis can be achieved with AED (amino ethyl order to obtain statistically signiﬁcant results, sets of at least 10 mea-
dextran), which causes instantaneous release of most of the trichocysts surements were carried out on each holder for each of the domains
(Plattner et al., 1984, 1985; Kerboeuf and Cohen, 1990). AED, kindly under analysis (Ca rim, cytoplasm, etc.). The beam was centered on
provided by J. Cohen and D. Kerboeuf, was used at 6 µM on a pellet the area to be measured and the location of the area was recorded over
of cells. About 30 seconds after stimulation, the cells were slammed on the image. Concerning the Ca rim located at the cell periphery, which
the cryobloc and cryodehydrated in the same way as the untreated cells. is the main subject of this paper, the diameter of the measured circle
Aliquots of cells were further treated with picric acid and observed with is larger than that of the rim. The rim was therefore positioned in the
a dark-ﬁeld microscope to check the extend of exocytosis. center of the measured ﬁeld. We checked that a slight move of the
rim toward the borders of the ﬁeld did not signiﬁcantly modify the
Muscle recorded values.
The cutaneous muscle of frog was taken up in Ringer’s solution. Very Usually, several ions were recorded from the same section (most
small pieces were cryoﬁxed and cryodehydrated exactly the same way frequently Ca2+, Na+, K+ and Mg2+) and, in most cases, a histologi-
as for Paramecium. cal section immediately following or preceding those analyzed by
To achieve a good preservation of cells (Paramecium and muscle), SIMS was stained with Toluidine Blue on a glass slide, to be observed
ultrathin sections were obtained and stained with uranyl acetate at the light microscope to provide a reference pattern.
followed by lead citrate. They were observed in Siemens Elmiskop
102 electron microscope.
SIMS microscopy study RESULTS
Serial semi-thin sections were deposited on glass slides for optical
examination and on ultrapure gold holders for ion analysis. Sections (1 Fast-freezing of wild-type Paramecium allows good
µm) on glass slides were stained with Toluidine Blue and sections (3 ultrastructural preservation
µm) for ion analysis were deposited over a microdrop of water on the In order to prevent ﬁxation-induced ion loss and redistribution,
gold holder, and heated to 60°C. In some control experiments sections
were deposited over the gold holder without any contact with water.
we adapted the cryoblock method of Escaig (1983) to use with
The instrument used in this study was an IMS 3F (CAMECA, a continuously fast-swimming, fragile cell such as Para-
Courbevoie, France), ﬁtted with two primary ion sources and mecium. We used small pieces of ﬁlter paper to trap the cells
connected to an image processing system. This system (Olivo et al., in a thin layer while keeping them alive and healthy. Under the
1989) allows digitalization of images (512×512 pixels), high-speed conditions used, only those cells that happen to be in the most
signal integration of the ionic images (to improve signal/noise ratio), superﬁcial portion of the ﬁlter facing the copper bloc were
histogram equalization (to increase the grey-scale contrast and to cooled rapidly, a condition essential for preventing formation
decrease background noise) and, ﬁnally, image superposition. of ice crystals. After rapid freezing, there are two main alter-
The ion microscope was operated with the O2+ primary source with natives for preserving the intracellular distribution of ions,
a 15 keV primary beam current of 200 nA. The image ﬁeld diameter freeze-substitution or even better, freeze-drying. The problem
was 150 µm. A mass resolution (M/∆M) of 2000 was used in order
to eliminate interferences between cluster ions and the speciﬁc
of freeze-drying and freeze-sectioning in SIMS is that the
elements under study. Under these conditions, elemental mapping of sections obtained in that way adhere very poorly to the gold
Ca, K, Na and Mg is easily achieved. holders used in the following step. Sod et al. (1990) have
For quantiﬁcation and calibration, internal reference elements were developed indium holders to overcome this difficulty for
used. In SIMS, the secondary ion beam current intensity (Ia) is a animal cell cultures, but this approach is difficult to apply to
function of the concentration of the analyzed element (Ca), the area single cells. In addition, in order to keep the paramecia in a
to be analyzed (S), the useful ion yield (Ya) and the primary ion beam state as close as possible to the physiological one, we avoided
current intensity (Ip). Thus, Ia = Ca.S.Ya.Ip. However, when dealing using high concentrations of centrifuged cells. Under the con-
with insulating specimens such as embedded biological samples, part ditions used, the density of cells found in the ﬁlter paper is low.
of the positive primary ion beam is repelled by charge effects when
This condition in addition to the previous one precluded the
negative secondary ions are extracted. The real intensity of the
primary ion beam current (Ip) is therefore variable, and the relation use of the freeze-fracture and freeze-drying technique
between Ia and Ca is not directly applicable. developed by Chandra et al. (1986). Therefore we resorted to
Nevertheless, a calibration is possible by measuring the intensity freeze-substitution and inclusion in resin, realizing that some
of the secondary ion beam using an internal reference element (Ir), redistribution of ions may occur at the ﬁnal step of inclusion
which is present at a large homogeneous and constant concentration in the resin when cells are brought back to room temperature.
in the specimen. Then Ia/Ir = K.Ca, where K is a proportionality Controls embedded in Lowicryl at low temperature were thus
constant which can be determined using a standard with increasing included to check for major redistribution (see below). The
concentration of the tested element to generate a calibration curve. As frozen pieces of ﬁlter paper were then taken through the
the carbon content of biological specimens and embedding resins are various steps of cryosubstitution in acetone, in the presence (or
virtually similar, this element can be used as an internal reference. In
absence) of oxalic acid, inclusion in epon resin (or in Lowicryl)
order to quantify Ca, calcium octoate (Calcium-Norol by
SICCANOR, 59282 Douchy-les-Mines, France) was used as a and sectioning (see Materials and Methods). Sections were ﬁrst
reference. Since calcium octoate is soluble in Epon-Araldite, samples observed by both light and conventional transmission electron
containing varying Ca concentrations, from 0.05 mM to 5 mM, were microscopy, to ascertain the quality of structural preservation.
prepared. Sections of 3 µm thickness were obtained and deposited By conventional light microscopy (after Toluidine Blue
without any water onto the gold holders. Emission was measured over staining) the cell contours and many cellular organelles were
ﬁelds of 150 µm in diameter. readily recognized. In Fig. 1, for example, we can recognize,
1898 N. Stelly and others
vesicles and even those of mitochondria were difficult to
recognize. Thus, mitochondria were revealed more by their
overall shape, distribution, compactness and ghosts of cristae
than through the conventional architecture of outer and inner
membranes. Similarly, the membrane which normally
surrounds the trichocysts was not visible. Treatment of the
sections with osmic acid did not modify these patterns. In
summary, cytoplasmic membranes appeared to be extracted.
A converse result was repeatedely observed for the alveolar
lumen: while in chemically ﬁxed cells the lumen of the alveoli
is usually swollen and essentially electron transparent
(‘empty’), as in Fig. 2a, in all types of cryoﬁxed cells the lumen
is more ﬂattened and appears to be ﬁlled with a meshwork of
electron-dense material distributed evenly throughout all of the
luminal volume (Fig. 2b,c). Note (Fig. 2c) that this ﬂuffy
material appears to be more abundant along the inner alveolar
membrane, i.e. on the surface facing the epiplasm. A similar
appearance can be seen in pictures published by Glas-Albrecht
et al. (1991) of Paramecium cells which were rapidly ﬁxed and
freeze-substituted by a somewhat different method. The
presence of the meshwork does not depend on the presence of
oxalic acid in the cryosubstitution method, is not affected by
Fig. 1. Rapidly frozen WT Paramecium: semi-thin section stained the type of embedding resin used and is seen even on unstained
with Toluidine Blue and observed with an optical microscope. N, sections, thus indicating that the meshwork is not an artefact
macronucleus; V, food vacuole; T, trichocysts attached to the cortex
C. Bar, 10 µm.
due to oxalate precipitation or staining. Cryoﬁxation therefore
reveals the presence of a genuine meshwork that had appar-
ently collapsed or was extracted in all previous studies using
from the periphery to the inside of the cell: cilia in the form of The overall ultrastructural conservation at ﬁrst sight
patches in some portions of the section, the cortex consisting appeared to be slightly less good in Lowicryl than in Epon-
of adjacent typical cup-shaped cortical units, the carrot-shaped Araldite. However, the cytoplasm and, especially, the
dark trichocysts perpendicular to the surface when cut longi- reticulum membranes appeared to be less extracted, and more
tudinally, the oral depression or oral apparatus, and ﬁnally the material seemed to be preserved in various complex organelles
dense cytoplasm with food vacuoles containing bacteria in such as the axonemes. All further studies presented in this
various stages of digestion and, occasionally, a portion of the paper, except for the Lowicryl control, were carried out with
macronucleus. The only abnormality observed is that the cells Epon-Araldite-embedded material because: the ultrastructural
located at the front are slightly deformed by the compression preservation appeared to be quite acceptable, these sections
generated during slamming on the copper block. adhered more easily to the gold substratum required for SIMS
Fig. 2 illustrates the ultrastructural characteristics of stained microscopy and the loss of ions was more limited during the
sections from cryoﬁxed wild-type cells (b,c) as compared to late stages of processing of Epon sections than Lowicryl ones
chemically ﬁxed ones (a). Two other variables were also (especially during recovery of sections on water).
analyzed (not shown): presence or absence of oxalic acid in the
cryosubstitution medium and inclusion in Lowicryl instead of SIMS reveals a Ca compartment at the periphery of
Epon. Oxalic acid was added in order to promote in situ pre- cryoﬁxed Paramecium cells
cipitation of calcium (see Nicaise et al., 1989); Lowicryl was Cell sections of both chemically ﬁxed and cryoﬁxed cells were
used in order to check the quality of ultrastructural preserva- analyzed by SIMS microscopy as described in Materials and
tion it provided compared with traditional resins, since its use Methods. The distribution of a large number of ions was
could prove useful both for ion distribution studies (by examined; Fig. 3 shows an example of the results observed on
allowing polymerization at low temperature after cryosubsti- cryoﬁxed cells with two physiologically important ones, Na+
tution, thus further preventing ion diffusion) and for immuno- and Ca2+, as compared to the light microscopic appearance of
cytochemical studies. In cells located close to the frozen front, an adjacent section. Fig. 3a corresponds to the Toluidine Blue-
good ultrastructural preservation was observed: the cortex with stained section and shows portions of seven cells, three of
its typical organelles was easily recognized, and even some of which, in the upper part of the picture, lie along the edge of
the cytoskeletal networks which underly the cortex such as the the frozen front. The various organelles pointed out in Fig. 1
epiplasm and the ﬁlamentous infraciliary lattice were seen can be recognized.
(Allen, 1971) (Fig. 2b,c). The major differences noted with The most striking differences in spatial distribution of ions
respect to conventional chemical ﬁxation are: ﬁrst, in the depend on the ﬁxation method concerned calcium: in chemi-
appearance of membranes within the cytoplasm and; second, cally ﬁxed cells, a weak Ca signal was uniformly distributed
in that of the alveolar lumen. Although the cytoplasm throughout the cells (not shown; see Fragu et al., 1992), while
displayed its classical ribosome- and glycogen-studded appear- in cryoﬁxed ones (Fig. 3b), the signal was restricted to a rela-
ance, membranes of the rough endoplasmic reticulum, of small tively narrow, bright band located around the cell periphery.
Calcium stores visualized by SIMS in Paramecium 1899
Fig. 2. Electron microscopy
of ultrathin sections of WT
Paramecium. alv, cortical
alveola; ci, cilium; T,
mitochondrion; pm, plasma
membrane; oam, outer
alveolar membrane; iam,
inner alveolar membrane;
ep, epiplasm; kf,
kinetodesmal ﬁbers. All
specimens were embedded
in Epon-Araldite. The
sections were stained with
uranyl acetate followed by
lead citrate. (a) Chemical
ﬁxation with 0.5%
glutaraldehyde in 50 mM
cacodylate buffer, followed
by 2% OsO4 in the same
buffer: cortical alveoli are
swollen and empty. Bar, 0.5
µm. (b) Cryoﬁxation and
acetone in the presence of
20 mM oxalic acid: cortical
alveoli are more ﬂattened
and are ﬁlled with a
meshwork of electron-dense
material. Trichocysts have
barely decondensed. Bar,
0.5 µm. (c) Detail of a
cortical alveola of a
cryoﬁxed cell: note the good
preservation of plasma and
cortical membranes and the
presence of dense material
throughout the alveolar
lumen and especially facing
the inner alveolar
membrane. Bar, 0.1 µm.
Thus, as suspected, chemical ﬁxation induced a drastic redis- yielded a homogeneously and intensely emitting cytoplasm in
tribution of ions, most clearly seen with Ca2+. Redistribution which only some large circular areas, most probably corre-
was less striking with Na+, K+ and Mg2+ because these sponding to food vacuoles, were not labelled (Fig. 3c). The
elements were uniformly distributed in cryoﬁxed cells: Na+ same was true for Mg2+; K+ yielded a slightly more granular
1900 N. Stelly and others
Fig. 4. SIMS image of Ca in cryoﬁxed WT Paramecium: the semi-
thin section was placed over the gold holder without water. Calcium
is located around the cell periphery as in Fig. 3f. Image ﬁeld is 150
immediately heated, we prepared some specimens that had no
contact with water at any stage. These are difficult to prepare
because of the tendency of sections to roll over themselves on
the gold holder in the total absence of water. SIMS of such a
‘dry’ specimen is shown in Fig. 4; Ca emission, identical to
that seen on sections deposited on water, is evident. As an addi-
tional control, cryoﬁxed cells were included in Lowicryl after
cryosubstitution in order to carry out the whole procedure at
low temperature. Although the appearance of the cells in SIMS
microscopy was slightly more hazy than in the case of Epon
embedding, the Ca rim was clearly visible, indicating that the
rim is not due to ion redistribution occurring during inclusion
In fact, when using low levels of integration, a punctate
peripheral Ca pattern with a periodicity identical to that of the
Fig. 3. SIMS and light microscopy of adjacent cell sections. The adjacent cortical alveoli was often observed (see for example,
sections traverse a large number of cells located at the edge of the the three cell sections to the right of Fig. 3b). Three points
sample (i.e. facing the frozen copper block). (a) Toluidine Blue- required evaluation, however, before a deﬁnitive acceptance of
stained section; (b) and (c) Ca and Na SIMS observed section
adjacent to that in (a). Note the regularity of the Ca peripheral signal
this hypothesis. First, although the difference observed
in (b), its independence from the presence of trichocysts and its between cryo- and chemically ﬁxed cells was encouraging and
absence over the macronucleus. Bar, 30 µm. suggested a speciﬁc Ca localization in cryoﬁxed cells, could
we be totally conﬁdent of our method? One way of answering
this question would be to show that our methods reveal the
or patchy labelling of the cytoplasm, again excluding vacuoles well-known speciﬁc Ca localizations in striated muscle.
(not shown). In chemically ﬁxed cells, Na+ and K+ were less Second, since the large exocytotic vesicles known as tri-
uniformly distributed, with K+ concentrated into large precip- chocysts are docked beneath the cell surface in Paramecium in
itates both inside and outside the cells. The general intensity an highly regular arrangement, could they be contributing to
of emission also appeared to be reduced (not shown; see Fragu the Ca signal? Indeed, many types of exocytotic vesicles
et al., 1992). In the remainder of this paper we will therefore contain large amounts of Ca (reviewed by Nicaise et al., 1992).
only be describing cryoﬁxed cells. The presence or absence of This second question could be explored by using cells devoid
oxalic acid in the cryosubstitution medium did not modify the of trichocysts. Third, did careful observation of a large number
results. Because some loss of elements can occur during the of sections indeed conﬁrm that both the size of the Ca-emitting
ﬁnal stages of specimen preparation, when the sections are zone and its precise shape and distribution agree with what is
brieﬂy deposited over a water drop on the gold holder and known for cortical alveoli at the EM level? In particular, does
Calcium stores visualized by SIMS in Paramecium 1901
Fig. 5. Frog cutaneous
muscle. (a) A semi-thin
section of a portion of muscle
ﬁber cryoﬁxed and observed
for Ca by SIMS. The bright
bands correspond to the
cisternae within the I band.
Bar, 25 µm. (b) Toluidine
Blue-stained sections of the
same ﬁber. Bar, 25 µm.
(c) Ultrastructure of the same
cryoﬁxed ﬁber where the
terminal cisternae (TC) are
aligned along the Z disk
within the I band. M,
mitochondrion. Bar, 1 µm.
the width of the compartment agree with what is known of the Trichocyst-deprived cells still display the peripheral
size of alveoli in electron microscopy? This last question could Ca compartment
be analyzed by carefully comparing successive sections of Two approaches were used to obtain cells devoid of tri-
material, some being observed by conventional light chocysts at the cortex; massive induction of trichocysts
microscopy to locate the major organelles and consecutive discharge by AED from wild-type cells (Plattner et al., 1984),
ones by SIMS microscopy, and also by comparing the distrib- or use of a mutant (tam8) lacking attached trichocysts at the
ution of Ca2+ with that of another ion such as Na+ to check cortex (Beisson and Rossignol, 1975; Lefort-Tran et al.,
whether extracellular adhesion of Ca occurred.These three 1981). The wild-type cells were frozen within 30 seconds after
points are addressed below. discharge; the extent of discharge was monitored by
observing, in a dark-ﬁeld microscope, aliquots of the AED-
SIMS identiﬁes the expected Ca compartment in treated cell suspension to which picric acid had been added.
muscle cells Although sometimes irregular, AED-induced discharge was
Samples of frog cutaneous muscle were prepared using exactly usually quite effective as shown by light and EM microscopy
the same methods as for Paramecium, i.e. by rapid freezing, controls. In these cells rapidly frozen after discharge, some
cryosubstitution in the presence of oxalic acid and embedding modiﬁcations in the ultrastructural aspect of the alveoli are
in Epon-Araldite. Conventional light microscopy shows the apparent: they appear to be somewhat collapsed, with the
typical striated appearance of sarcomeres (Fig. 5b) and electron outer membrane disjointed from the plasma membrane, and to
microscopy indicates a reasonably good ultrastructural preser- contain less ﬂuffy material than the controls. The mutant cells
vation (Fig. 5c). I and A bands are readily identiﬁed and the were processed as wild-type cells. tam8, clearly, lacked
terminal cisternae are clearly seen in the I bands. In this tissue, attached trichocysts at the cortex; its trichocysts lay randomly
EM electron-probe analysis of ultrathin cryosections has within the cytoplasm. One additional mutant strain was used,
shown that 60 to 70% of total ﬁber Ca is localized in the nd9. This is a conditional mutant, defective in exocytosis at
terminal cisternae, within the I bands (Somlyo et al., 1981). 27°C, but displaying a normal complement of trichocysts
SIMS microscopy (Fig. 5a) shows a regular alternation of attached at the cortex. It is therefore defective in only the very
emitting and non-emitting bands for Ca, corresponding, ﬁnal steps of exocytosis (Beisson et al., 1980). This strain
respectively, to the light and dark striations seen in light therefore allows discrimination between effects due to the
microscopy, and therefore to I and A bands, respectively. As absence of trichocysts at the surface from effects only due to
expected, Ca is thus restricted to the I bands in which the lack of exocytosis potential.
terminal cisternae of the sarcoplasmic reticulum are located, In all cases (AED-treated WT, tam8, nd9), the peripheral Ca
and no major loss or redistribution of calcium seem to have compartment was still observed by SIMS microscopy (Fig.
occurred during sample preparation. 6a,b,c). The Ca image of nd9 cells appeared to be identical to
The periphery of the ﬁber also strongly emitted Ca, that of WT ones. While in AED-treated WT cells and in tam8
probably because of the presence of a dense array of vesicles ones, the width and intensity of the Ca zone appeared to be
lying immediately beneath the sarcolemna, referred to as slightly reduced. This led us to a more quantitative study as
caveolae by Franzini-Armstrong (1970) and which possibly described below. In any case, it became clear that trichocysts
contain high amounts of calcium originating from the exta- cannot be the major contributors to the peripheral Ca signal,
cellular medium. since this signal was always present in tam8 cells, where con-
1902 N. Stelly and others
Fig. 6. Ca images of different types of
paramecia showing conservation of the
peripheral calcium signal. (a) WT
Paramecium after stimulation by AED; (b)
mutant nd9; (c) mutant tam8; (d) Toluidine
Blue-stained optical view of a section from
the same tam8 cell as that in c. V, vacuoles;
T, trichocysts; G, gullet. Bars, 20 µm.
ventional light microscopy of adjacent cells revealed the adjacent sections observed by light microscopy (Figs 3 and 6)
absence of trichocysts at the cortex and their scattered presence yields the following observations.
in the cytoplasm (see, for example, Fig. 6d). (1) The Ca-emitting zone exactly follows the periphery of
A reverse argument can also be made: in many instances, the cells, with all its deformations and, in particular, clearly
when a dense array of longitudinally sectioned trichocysts was outlines the shape of the oral depression (which is also
observed in control sections, the corresponding SIMS image bordered by alveoli for the most part; see Allen, 1974).
of an adjacent section often showed a Ca-emitting zone which (2) The inner cytoplasmic portion of the sections gives a
was narrower than the zone occupied by the trichocyst bodies much lower Ca signal except, occasionally, for food vacuoles.
(see Fig. 3a, for example). There may be some correlation between the ‘age’ of vacuoles
and the intensity of their signal, young vacuoles showing
Ca emission at the cell periphery is strictly higher signal than old ones (see Fig. 6c,d). No Ca emission was
correlated with the presence and integrity of the observed from these massive DNA-containing macronuclei.
cortex (3) When a portion of a cell was torn during preparation, and
Side by side comparison of SIMS images and corresponding lost its cortex, no Ca signal was observed in this area.
Calcium stores visualized by SIMS in Paramecium 1903
Fig. 7. Ca and Na saturation
analysis. The same section was
submitted to recordings of
increasing duration for Ca
(a,b,c,d) and Na (e,f,g,h),
corresponding, respectively, to
250, 500, 1,000 and 2,000
integrations. Note the increasing
intensity and thickness of the Ca-
emitting zone from (a) to (d) and
the absence of spillover of the Na
signal out of the cells in (h).
Image ﬁeld is 150 µm diameter.
(4) When a section was tangential to the surface and SIMS microscopy, seems to be broader than would be expected
contained many cortical units such as the lower part of the cell if only the lumen of the alveoli were emitting. We examined
in Fig. 1, the Ca-emitting zone was much enlarged. this question, with two types approaches, image saturation
(5) Cell sections located far from the freezing front tend to studies and morphometric studies involving superposition of
display a much less regular Ca rim, the farther ones sometimes the Ca and Na images.
being devoid of it. For saturation studies, an increasing number of images
Thus, the Ca-emitting zone appears to be tightly correlated were summed, starting from low levels of integration
with that known to contain alveoli and to depend on good (detecting high Ca concentration) to very high ones (detecting
freeze-ﬁxation. However, the width of this zone, as seen by low Ca concentration) (see Materials and Methods). At low
1904 N. Stelly and others
Fig. 8. Na and Ca image superposition. The same section was submitted to Ca analysis (a) followed by Na analysis (b) and the two resulting
images were superposed in the computer (c). Note that a peripheral yellow to orange rim is obtained in c, reﬂecting the good superposition of
the Ca rim within the limits of the Na border. Also note that the most rapidly frozen front is to the right of the image, as can be seen by the
slight deformation of the cells due to their slamming over the copper block and by the presence of extracellular calcium-rich deposists. In
contrast, note that cells deeper in the sample (to the left of the image), although they are detected in the Na image (b), lack the red Ca rim (a),
probably because of poor ﬁxation. Bar, 10 µm.
levels, a ﬁne Ca line ﬁrst appears all along the cell periphery, the quasi-uniform distribution of the much smaller amount of
only very slightly spilling over (as based on the Na and K total cytoplasmic Ca than that located in the alveoli.
images; see below), but it thickens with increasing integra- Taken together, these saturation studies indicate the
tions, reaching a width of approximately 5 µm (Fig. 7a to d). presence of a thin intracellular peripheral compartment with a
This is broader than the largest width of alveoli as measured very high Ca concentration, surrounded, on the outside, by a
on EM images (approx. 3 µm). Thus, the width of the Ca- narrow Ca-containing zone and, on the inside, by a broader Ca-
emitting zone is broader than what strictly corresponds to the containing domain where Ca concentration decreases in a
alveoli. The same phenomenon was observed using ‘dry’ graded manner towards the cytoplasm of the cell.
sections; it is therefore not due to diffusion induced by the
water drop used on the gold holder. Under the same satura- Quantiﬁcation of alveolar Ca reveals a two-fold
tion conditions, the Na, K and Mg emissions did not spill over decrease after exocytosis
across the cell boundary but remained conﬁned within their As brieﬂy indicated above, SIMS analysis revealed, in both
initial area of distribution (Fig. 7g and h), again indicating tam8 cells and the wild type cells in which a massive release
that the situation observed for Ca reﬂects a real in situ dis- of trichocysts was induced, a Ca emission lower than that of
tribution and not nonspeciﬁc diffusion. It should be stressed, control cells. This had to be quantiﬁed rigorously because
however, that the level of integration required to see the section to section comparison in SIMS microscopy may not be
widening of the calcium rim is at least an order of magnitude reliable. A proportionality coefficient (Ca/C) was therefore
higher than that needed to see the initial peripheral signal. established for each of our measurements by measuring simul-
Thus, the amount of the excess calcium is much lower than taneously the Ca and C beam intensities within the same
that seen at low integration levels. volume of sample. Each measure provided corresponds to the
Widening of the Ca-emitting zone occurred mainly inward. mean of 10 measurements carried out over a surface of 8 µm
This was established by measurements on the micrographs, diameter. This allows a sample to sample comparison (see
superposition of hand-drawn tracings and, most directly, by Materials and Methods). In addition, absolute quantiﬁcation of
computer superposition of the calcium and sodium images Ca was achieved using an internal standard consisting of
(Fig. 8), using the programs of Olivo et al. (1989). On such calcium octoate included in Epon-Araldite (see Materials and
images, it can be seen that at low integration levels, the red rim Methods).
(Ca) superposes well over the green background (Na) giving a Table 1 summarizes the quantitative observations in
yellow-orange border, with no indication of a red external different types of cell sections. As expected, Ca concentration
margin; this remains the case even at higher levels of integra- is signiﬁcantly higher (about 7-fold) at the cell periphery than
tion. This shows: ﬁrst, that the calcium rim detected at low within the cytoplasm in all types of cells and conditions (2 to
levels of integration is located inside the cell; and, second, that 3 mM vs 0.3 to 0.4 mM). It is interesting to compare the Ca
the outer boundary of calcium distribution does not extend concentration within the cortex of Paramecium with that in
much beyond that of Na. muscle, using SIMS. We found the muscle values to oscillate
At extremely high integration levels (10 to 20-fold higher around 10 mM. This is quite comparable to the value found by
than necessary to see the rim), a weak Ca signal was eventu- Somlyo et al. (1981) using EM microanalysis: their values
ally seen throughout the cytoplasm, most probably reﬂecting range from 10 to 117 mmol/kg dry weight, the highest value
Calcium stores visualized by SIMS in Paramecium 1905
Table 1. Quantitative evaluation of Ca amounts The idea that the cortical alveoli, a network of large, inter-
Cortex n1† n2 Cytoplasm n1 n2 connected membrane vesicles directly underlying the plasma
Wild type 17.9±8.4.10−1* 35 22 3.0±1.2.10−1 34 30
membrane in Paramecium and other ciliates might correspond
(3 exp.) 3.4±1.8 mM 0.4±0.2 mM to a Ca-sequestering compartment akin to the sarcoplasmic
Wild type + 10.5±5.10−1 36 20 3.7±1.5.10−1 15 11
reticulum of muscle cells is old (Allen and Eckert, 1969; Satir
AED 1.8±1.2 mM 0.6±0.4 mM and Wissig, 1982). It received strong support when we
(2 exp.) succeeded in purifying these vesicles and showed that they
tam8 12.4±6.2.10−1 41 14 4.4±3.10−1 11 11 actively pump Ca2+ in an ATP- and Mg2+-dependent process
(1 exp.) 2.3±1.2 mM 0.7±0.6 mM (Stelly et al., 1991). Additional evidence was provided by the
tam8 + AED 10.6±4.5.10−1 19 9 4.8±1.7.10−1 7 6 work of Schmitz et al. (1985), Zierold (1991) and Knoll et al.
(1 exp.) 1.8±1.0 mM 0.8±0.4 mM (1993) using EM microanalysis methods. Here, we have sought
nd9 17.9±5.1.10−1 25 12 3.3±1.3.10−1 6 6 to provide direct visualization of this compartment over large
(1 exp.) 3.4±1.2 mM 0.5±0.2 mM areas of many cells using a new approach, that of SIMS
microscopy. The main advantage of this method lies in the fact
*The upper line represents the mean of Ca count/C count ± s.d. The second that it provides images of total Ca distribution over individual
line is the Ca concentration in mM, calculated using Calcium reference.
†The numbers in the n1 columns refer to the number of surface spots over cell sections and is thus especially suited for identifying Ca
which measurements were carried out; the numbers in the n2 columns refer to storage sites (as compared to methods detecting only free
the number of different cell sections examined. Ca2+). Since the lateral resolution of the method is limited,
however, the compartment must be of sufficient size to be iden-
tiﬁable. Because cortical alveoli are typically about 0.2 to 2 µm
corresponding to a position of their narrow (20 µm) probe × 1 to 3 µm in size, we hoped that, if they indeed contained
within the terminal cisternae. When converted to mM, the units large amounts of Ca, SIMS microscopy would allow their visu-
used in the present work, by taking into account the water alization. In addition, there was a clear prediction as to the
amount, their values are 3 to 30 mM, i.e. with an average very expected intracellular location of the signal, namely through-
close to our 10 mM value obtained by SIMS with a much larger out the cell periphery.
probe diameter. These predictions were fulﬁlled remarkably: a continuous,
Thus, the total Ca concentration in the cortex is in the same peripheral, Ca-emitting zone was immediately seen, provided
range as that of the sarcoplasmic reticulum. that cells were ﬁxed by rapid freezing.
After induction of massive exocytosis by AED, a marked The good ultrastructural conservation of cells after cryoﬁx-
decrease in Ca concentration in the cortex was repeatedly ation, the fact that the peripheral location of Ca strictly
observed (from 3.4±1.8 mM to 1.8±1.2 mM). This decrease is depended on avoiding chemical ﬁxation, and the excellent cor-
statistically highly signiﬁcant using the t-test (P<0.0001). relation observed between the occurrence of the Ca signal and
Interestingly, this decrease in the cortex appears to be corre- the presence of a strip of well frozen cortex in the corre-
lated with an increase in Ca concentration within the sponding section, all argue against artifacts. In addition, the
cytoplasm, but this is barely signiﬁcant statistically. The fact that it is found also in cortices devoid of trichocysts
kinetics of these changes was not studied. All the data demonstrates that the Ca does not emanate predominantly from
presented are from experiments with cells that were cryoﬁxed these exocytotic organelles. The most likely Ca-containing
about 30 seconds after exocytosis. compartment therefore remaining the alveolar one. Additional
Concerning trichocyst mutant strains, nd9 displayed a evidence on these two points was recently provided by electron
cortical Ca concentration identical to that of wild type, indi- probe microscopy. Using conventional X-ray microanalysis,
cating that, when trichocysts are attached at the cortex, inca- we found amounts of Ca in a number of alveoli ranging from
pacity to carry out exocytosis does not lead per se to a modi- 5 to 10 mM (Stelly, Halpern and Nicaise, unpublished). No Ca
ﬁcation of Ca concentration. signal was observed on trichocysts and this is all the more sig-
In contrast, tam8 displayed a signiﬁcantly lower amount, niﬁcant, since these organelles are easily identiﬁed in the
placing it at an intermediate level between normal wild type unstained sections used for microanalysis. Conﬁrmation of
and wild type after exocytosis. Treatment of tam8 cells with these two points can be found in a recent study of Knoll et al.
AED further decreased the Ca concentration in the cortex but (1993) in which one electron energy loss spectrum image of
this was only marginally signiﬁcant statistically (P<0.02). ultrarapidly ﬁxed Paramecium is provided, showing the
Trichocyst exocytosis or lack of attached trichocysts at the presence of Ca in alveoli (not in trichocysts) and its redistrib-
cortex therefore lead to a signiﬁcant decrease in cortical Ca ution after exocytosis. It should be pointed out that previous
concentration. X-ray microanalysis studies had already clearly indicated the
presence of Ca below the cell surface but not in the trichocysts
(Schmitz et al., 1985). In summary, all the available evidence
DISCUSSION converges to indicate that trichocysts make little or no contri-
bution to the peripheral calcium signal and that the signal
The major result presented in this paper is the visualization of emanates predominantly from alveoli.
a subcortical Ca compartment in Paramecium using a new The width of the Paramecium band, as seen in SIMS
method, SIMS microscopy, applied to sections of ultrarapidly microscopy, can be estimated as 3 to 5 µm, while the EM
frozen cells. In addition, the amount of Ca in this compartment observations indicate that the alveoli do not exceed 2 µm. In
was found to decrease substantially after massive induction of addition, saturation studies show that this width can reach 10
exocytosis. to 20 µm with a narrow Ca zone observed on the outside of the
1906 N. Stelly and others
cell, a phenomenon not seen with Na or K and Mg. Several newly discovered material in Paramecium alveoli. We
explanations can account for this observation. These fall into searched for a homologue of calsequestrin in the cortical
two broad categories: artefactual diffusion of Ca from the fraction through immunoblotting, Stains-all decoration and
alveoli during sample preparation, or normal presence of Ca in radioactive Ca-overlay of gel blots but did not observed a
the proximity of the alveolar compartment. The ﬁrst hypothe- signal in the calsequestrin Mr zone.
sis cannot be totally excluded and, in fact, some diffusion The second corollary of the high intra-alveolar Ca concen-
would not be surprising in view of all the steps involved in tration is the likelihood of a tight control over its release. Pre-
sample preparation. We have excluded the possibility, viously, we have focused on the analysis of Ca2+ uptake into
however, that rapid ﬂotation of sections has a major effect, alveoli in vitro (Stelly et al., 1991). We are currently charac-
since sections obtained by completely avoiding any contact terizing the release system kinetically and pharmacologically.
with water showed an identical distribution of all the major There are indirect indications for the operation of an InsP3
ions analyzed. In addition, inclusion in Lowicryl at low tem- system in Paramecium (Beisson and Ruiz, 1992) and various
perature yielded exactly the same Ca rim as that in Epon- components of the inositide cascade are present (Freund et al.,
Araldite, indicating that no major diffusion occurs at the time 1992) although InsP3 itself has proven elusive. Similarly, the
when the cryosubstituted samples are warmed up. Finally, this occurrence of a Ca2+-induced/Ca2+-release cascade has been
extracellular emitting zone is Ca-speciﬁc and was not observed suggested to underly morphogenetic waves during cell division
for several other ions, indicating that it does not reﬂect a gen- (Le Guyader and Hyver, 1991), but remains to be demon-
eralized diffusion from the inside of the cell. The second strated.
hypothesis is therefore much more likely and at least two As for the function of alveoli in Ca2+ regulation, at least
possible explanations for a genuine in situ broader Ca zone can three roles might be considered (Stelly et al., 1991), especially
be suggested. First, the cell is covered by cilia and both the when it is noted that the alveoli are in close proximity to three
plasma and ciliary membranes may be expected to bind a sub- Ca2+-controlled organelles, the cilia, the trichocysts and
stantial amount of Ca by means of the negatively charged phos- cytoskeletal networks: (1) general homeostasis of intracellular
pholipids and the glycosylated cell coat, thus spreading the Ca2+ by active sequestration above a given cytosolic level; this,
signal towards the outside of the cell. Second, concerning the for example, might be the case when a surge of Ca2+ occurs
spread of the signal towards the cell’s interior, it should be through depolarization of the ciliary membrane. The alveoli
recalled that a vast ﬁlamentous network, the infraciliary being immediately adjacent to basal bodies might pump the
network, recently shown to be made up of Ca-binding proteins Ca2+ that has entered the ciliary lumen; (2) release of at least
(Garreau de Loubresse et al., 1991), makes up the deepest of part of the Ca2+ required for trichocyst exocytosis. In this
the cortical cytoskeletal layers several micrometers below the respect, it must be stressed that alveolar membranes are tightly
alveoli. This network may well yield a signal beneath the apposed to the membranes of the tip of trichocysts; (3) release
alveoli on the sections. In addition in Paramecium, especially of the Ca2+ required for the disassembly of the several
in stationary phase cells, mitochondria tend to be concentrated cytoskeletal networks adjacent to the cortex when they undergo
just below the cortex, providing a further possible weaker reorganization during division (Iftode et al., 1989). Of these,
calcium-emitting zone (see Girard et al., 1991, and Rizzuto et one of the thickest is the infraciliary lattice, which is made up
al., 1993, for recent evidence of calcium pumping of ER- of Ca-binding contractile proteins (Garreau de Loubresse et al.,
released calcium by mitochondria). It appears, therefore, that 1991). In the case of this network, the alveoli may also provide
three Ca domains can be identiﬁed at the cell periphery in Ca2+ to regulate contractility.
Paramecium: (i) a narrow zone of very high Ca concentration The present results clearly support the idea that the alveoli
corresponding to alveoli, surrounded by: (ii) a small extracel- provide at least some of the Ca2+ required during exocytosis,
lular zone probably corresponding to membrane- and cilium- since after massive exocytosis the amount of Ca2+ inside the
bound Ca; and (iii) a larger intracellular zone displaying a Ca alveoli dropped by 50%. It should be stressed that such a
concentration decreasing towards the inside of the cell. massive drop in Ca2+ is quite similar to what occurs in the sar-
Through calibration of the absolute amount of Ca using a coplasmic reticulum during muscle contraction (Somlyo et al.,
calcium octoate derivative embedded and analyzed in the same 1981). In fact, we may not have captured the point of lowest
conditions, the absolute amount of Ca contained in various Ca2+ concentration, since the cryoﬁxation device used imposes
areas of the sections was approximated. Within the peripheral a delay of about 30 seconds between the application of the sec-
band, the mean value was 3.4±1.8 mM. This concentration is retagogue (AED) and the ﬁxation of cells. Using faster ﬁxation
higher by several orders of magnitude than that of free methods, Knoll et al. (1993) recently observed a profound
cytosolic Ca2+ (10−4 mM; Eckert, 1972). There are several redistribution of alveolar Ca2+ within 80 milliseconds after
implications of this considerable difference. First, it is most AED-induced exocytosis in Paramecium. This observation
likely that Ca is associated with a Ca-sequestering protein agrees very well with our results.
within the lumen of the alveoli. In fact, we wonder whether the The question of the origin of the Ca2+ required for exocyto-
ﬂuffy material identiﬁed in the alveoli only after ultra-rapid sis has been extensively discussed both for Paramecium
freezing corresponds to such hypothetical sequestering (Plattner et al., 1991; Knoll et al., 1992, 1993; Cohen and
proteins. A tempting cytological analogy with calsequestrin Kerboeuf, 1993) and other systems. Our data indicate that a
can indeed be made: only when rapid ﬁxation was used was a contribution of alveoli to the process is very likely. Interest-
diffuse, ﬂuffy material clearly seen in terminal cisternae, which ingly, Cohen and Kerboeuf (1993), on the basis of a completely
was later identiﬁed as calsequestrin (Jorgensen and Campbell, different approach, also concluded that at least part of the Ca2+
1984; Jorgensen et al., 1985). Previously, conventional involved in trichocyst exocytosis originates from intracellular
chemical ﬁxation had failed to reveal it, as is the case with the stores, which, they suggest, might correspond to the alveoli.
Calcium stores visualized by SIMS in Paramecium 1907
The exact path followed by alveolar Ca2+ during exocytosis is contain a ryanodine receptor (McPherson et al., 1992, Sardet
still unknown but the tight apposition of alveolar and trichocyst et al., 1992). The ciliate Ca stores may therefore be evolution-
membranes, at the tip of trichocysts, suggests a direct ﬂow, ary homologues of this ER-derived network and, indeed, of
through speciﬁc transmembrane proteins. other Ca storage compartments of ‘higher’ eukaryotes, but this
Assuming that trichocysts themselves contain very little is an unproven generalization. The search for proteins homol-
calcium, as we concluded above, the fact that exocytotic ogous to those of the compartments of higher eukaryotes in
mutants lacking attached trichocysts show a much decreased alveolar sacs has been hampered by the great evolutionary
amount of Ca2+ in alveoli, while those having attached tri- distance separating ciliates from metazoa, which usually leads
chocysts have normal amounts, provides a further hint as to the to failure of cross-reaction when most antibodies to metazoan
regulation of alveolar amounts. It suggests that trichocyst proteins are tested in Paramecium. In fact, the exact relation-
attachment per se, independently of exocytotic capability, ship of alveoli to the endoplasmic reticulum in ciliates and their
provides a regulatory loop inducing normal Ca2+ pumping and mode of biogenesis are still unclear. It is being studied at
sequestration in alveoli. It should be recalled in this context present using EM immunocytochemistry by our group, with
that mutant tam8, which lacks cortex-attached trichocysts, was plasma membrane markers (Charret et al., unpublished;
shown in our previous work to pump Ca2+ in vitro as efficiently Capdeville et al., 1993).
as the wild type (Stelly et al., 1991). Thus, the in vivo decrease
in alveolar Ca described in the present work did not result from This work was supported by a grant from the Université Paris-Sud
an intrinsic defect of the Ca2+-pumping machinery but was due (Action Interdisciplinaire 8922). We thank Dr J. Cohen and D.
rather to an indirect, presumably regulatory, process. Kerboeuf for mutant strains and gift of AED, Drs J. P. Mauger and J.
How wide is the occurrence of such vesicular Ca compart- Cohen for critical reading of the manuscript, Dr M. Müller for his help
in improving the manuscript and Prof. G. Nicaise for sharing the X-
ments closely apposed to the plasma membrane? The obser- ray results and for comments on the manuscript. We are grateful to
vation made in Paramecium appears to be valid for other Mrs N. Narradon for expert photographic assistance and to Mrs C.
ciliates: preliminary SIMS observations carried out in Tetrahy- Couanon for her careful preparation of the manuscript.
mena, a genus relatively close to Paramecium in molecular
evolutionary terms, but also in Euplotes, a very distant hypotri- REFERENCES
chous ciliate (Baroin-Tourancheau et al., 1992), show the
Allen, R. D. and Eckert, R. (1969). A morphological system in ciliates
presence of the peripheral Ca band (Stelly, unpublished). comparable to the sarcoplasmic reticulum-transverse tubular system in
Cortical alveoli are in fact a shared ultrastructural character of striated muscle. J. Cell. Biol. 43 (2, Pt. 2), 4a (Abstr.)
three protist Phyla, ciliates, dinoﬂagellates and apicomplexa Allen, R. D. (1971). Fine structure of membranous and microﬁbrillar systems
(the Phylum comprising Plasmodium, Toxoplasma, gregarines in the cortex of Paramecium caudatum. J. Cell Biol. 49, 1-20.
and other parasitic protists); these three Phyla have been shown Allen, R. D. (1974). Food vacuole membrane growth with microtubule-
associated membrane transport in Paramecium. J. Cell Biol. 63, 904-922.
by molecular phylogenetic analysis to form a monophyletic Andrews, S. B., Leapman, R. D., Landis, D. M. D. and Reese, T. S. (1987).
group (the ‘alveolata’) (Gajadhar et al., 1991). It is tempting Distribution of calcium and potassium in presynaptic nerve terminals from
to suggest that in these three groups the alveoli play the role cerebellar cortex. Proc. Nat. Acad. Sci. USA 84, 1713-1717.
of Ca-sequestering organelles. In fact, in these three types of Baroin-Tourancheau, A., Delgado, P., Perasso, R. and Adoutte, A. (1992).
A broad molecular phylogeny of ciliates: Identiﬁcation of major evolutionary
organisms there are exocytotic functions that may correlate trends and radiations within the phylum. Proc. Nat. Acad. Sci. USA 89, 9764-
with the presence of alveoli. 9768.
Going to much more distant biological groups, the occur- Beisson, J. and Rossignol, M. (1975). Movements and positioning of
rence of plasma membrane-linked Ca compartments in various organelles in Paramecium aurelia. In Molecular Biology of
types of mammalian cells has been proposed (Putney, 1986). Nucleocytoplasmic Relationships (ed. S. Puiseux-Dao), pp. 291-294.
Elsevier Scientiﬁc Publishing Company, Amsterdam, The Netherlands.
In some highly differentiated metazoan cells, the relation of the Beisson, J., Lefort-Tran, M., Pouphile, M., Rossignol, M. and Satir, B.
endoplasmic reticulum to the plasma membrane becomes quite (1976). Genetic analysis of membrane differentiation in Paramecium.
specialized, the paradigm being skeletal muscle cells with the Freeze-fracture study of the trichocyst cycle in wild-type and mutant strains.
tight apposition of a specialized domain of the sarcoplasmic J. Cell Biol. 69, 126-143.
reticulum to the T-tubule membrane, at the level of which Beisson, J., Cohen, J., Lefort-Tran, M., Pouphile, M. and Rossignol, M.
(1980). Control of membrane fusion in exocytosis. Physiological studies on a
signal transduction operates (Caswell and Brandt, 1989). Paramecium mutant blocked in the ﬁnal step of the trichocyst extrusion
However, the closest analogy at the moment is with eggs of process. J. Cell Biol. 85, 213-227.
sea urchins (Luttmer and Longo, 1985) and of ascidians Beisson, J. and Ruiz, F. (1992). Lithium-induced respeciﬁcation of pattern in
(Gualtieri and Sardet, 1989). There are indeed a number of Paramecium. Dev. Genet. 13, 194-202.
Burgoyne, R. D. and Cheek, T. R. (1991). Locating intracellular calcium
striking similarities between the cortical alveoli of Parame- stores. Trends Biochem. Sci. 16, 319-320.
cium and a vesicular ER network in the cortex of the ova of Capdeville, Y., Charret, R., Antony, C., Delorme, J., Nahon, P. and
various sea urchin species (Gardiner and Grey, 1983; Sardet, Adoutte, A. (1993). Ciliary and plasma membrane proteins in Paramecium:
1984; Terasaki et al., 1991); both actively pump Ca2+ in vitro, Description, localization and intracellular transit. In Advances in Cell and
both appear to release at least some of the Ca2+ required for Molecular Biology of Membranes, vol. 2A: Membrane Traffic in Protozoa
(ed. H. Plattner), pp. 181-226. Jai Press Inc., Greenwich, USA.
exocytosis (trichocysts in Paramecium, cortical granules in sea Carafoli, E. (1987). Intracellular calcium homeostasis. Annu. Rev. Biochem.
urchin; Gillot et al., 1991) and both constitute the major 56, 395-433.
calcium store of the cell. Castaing, R. and Slodzian, G. (1962). Microanalyse par émission ionique
The sea urchin egg network has several additional proper- secondaire. J. Microsc. 1, 31-38.
ties which have not been detected in Paramecium: it contains Caswell, A. H. and Brandt, N. R. (1989). Open question, Does muscle
activation occur by direct mechanical coupling of transverse tubules to
calsequestrin (Henson et al., 1989), releases Ca2+ in response sarcoplasmic reticulum? Trends Biochem. Sci. 14, 161-165.
to InsP3 (Terasaki and Sardet, 1991) and also appears to Chandra, S., Bernius, M. T. and Morrison, G.H. (1986). Intracellular
1908 N. Stelly and others
localization of diffusible elements in frozen-hydrated biological specimens Knoll, G., Grässle, A., Braun, C., Probst, W., Höhne-Zell, B. and Plattner,
with ion microscopy. Anal. Chem. 58, 493-496. H. (1993). A calcium inﬂux in neither strictly associated with nor necessary
Chandra, S. and Morrison, G. H. (1988). Ion microscopy in biology and for exocytotic membrane fusion in Paramecium cells. Cell Calcium 14, 173-
medicine. Meth. Enzymol. 158, 157-179. 183.
Chandra, S. and Morrison, G. H. (1992). Sample preparation of animal Koch, G. L. E. (1990). The endoplasmic reticulum and calcium storage.
tissues and cell cultures for secondary ion mass spectrometry (SIMS) BioEssays 12, 527-531.
microscopy. Biol. Cell 74, 31-42. Lefort-Tran, M, Aufderheide, K., Pouphile, M., Rossignol, M. and Beisson,
Cohen, J. and Kerboeuf, D. (1993). Calcium and trichocyst exocytosis in J. (1981). Control of exocytotic processes: cytological and physiological
Paramecium: Genetics and physiological studies. In Advances in Cell and studies of trichocyst mutants in Paramecium tetraurelia. J. Cell Biol. 88,
Molecular Biology of Membranes, vol. 2A: Membrane Traffic in Protozoa 301-311
(ed. H. Plattner), pp. 61-81. Jai Press Inc., Greenwich, USA. Le Guyader, H. and Hyver, C. (1991). Duplication of cortical units on the
Eckert, R. (1972). Bioelectric control of ciliary activity. Locomotion in the cortex of Paramecium: A model involving a Ca2+ wave. J. Theor. Biol. 150,
ciliated protozoa is regulated by membrane-limited calcium ﬂuxes. Science 261-276.
176, 473-481. Linton, R. W. and Goldsmith, J. G. (1992). The role of secondary ion mass
Escaig, J. (1983). Techniques de congélation ultra-rapide de specimens spectrometry (SIMS) in biological microanalysis: technique comparisons
biologiques sans cryo-protecteur. In Microanalyse en Biologie (ed. C. and prospects. Biol. Cell 74, 147-154.
Quintana, and S. Halpern), pp. 105-120. Société Française de Microscopie Luttmer, S. and Longo, F. J. (1985). Ultrastructural and morphometric
Electronique, Paris. observations of cortical endoplasmic reticulum in Arbacia, Spisula and
Fragu, P., Briançon, C., Fourré, C., Clerc, J., Casiraghi, O., Jeusset, J., Mouse Eggs. Dev. Growth Differ. 27, 349-359.
Omri, F. and Halpern, S. (1992). SIMS microscopy in the biomedical ﬁeld. Lytton, J. and Nigam, S. K. (1992). Intracellular calcium: molecules and
Biol. Cell 74, 5-18. pools. Curr. Opin. Cell Biol. 4, 220-226.
Franzini-Armstrong, C. (1970). Studies of the triad. I. Structure of the McPherson, S. M., McPherson, P. S., Mathews, L., Campbell, K. P. and
junction in frog twitch ﬁbers. J. Cell Biol. 47, 488-499. Longo, F. J. (1992). Cortical localization of a calcium release channel in sea
Freund, W. D., Mayr, G. W., Tietz, C. and Schultz, J. E. (1992). Metabolism urchin eggs. J. Cell Biol. 116, 1111-1121.
of inositol phosphates in the protozoan Paramecium. Characterization of a Meldolesi, J., Madeddu, L. and Pozzan, T. (1990). Intracellular Ca2+ storage
novel inositol-hexakisphosphate-dephosphorylating enzyme. Eur. J. organelles in non-muscle cells: heterogeneity and functional assignment.
Biochem. 207, 359-367. Biochim. Biophys. Acta 1055, 130-140.
Gajadhar, A. A., Marquardt, W. C., Hall, R., Gunderson, J., Ariztia- Meldolesi, J. and Villa, A. (1993). Endoplasmic reticulum and the control of
Carmona, E.V and Sogin, M. L. (1991). Ribosomal RNA sequences of Ca2+ homeostasis. In Subcellular Biochemistry, vol. 21: Endoplasmic
Sarcocystis muris, Theileria annulata and Crypthecodinium cohnii reveal reticulum (ed. N. Borgese and J. R. Harris), pp. 189-207. Plenum Press, New
evolutionary relationships among apicomplexans, dinoﬂagellates, and York.
ciliates. Mol. Biochem. Parasitol. 45, 147-154. Mentré, P. and Escaig, F. (1988). Localization of cations by pyroantimonate.
Gardiner, D. M. and Grey, R. D. (1983). Membrane junctions in Xenopus I. Inﬂuence of ﬁxation on distribution of calcium and sodium. An approach
eggs: Their distribution suggests a role in calcium regulation. J. Cell Biol. 96, by analytical ion microscopy. J. Histochem. Cytochem. 36, 49-54.
1159-1163. Nicaise, G., Gillot, I., Julliard, A. K., Keicher, E., Blaineau, S., Amsellem,
Garreau de Loubresse, N., Klotz, C., Vigues, B., Rutin, J. and Beisson, J. J., Meyran, J. C., Hernandez-Nicaise, M. L., Ciapa, B. and Gleyzal, C.
(1991). Ca2+-binding proteins and contractility of the infraciliary lattice in (1989). X-ray microanalysis of calcium containing organelles in resin
Paramecium. Biol. Cell 71, 217-225. embedded tissue. Scanning Microsc. 3, 199-220.
Gillot, I., Ciapa, B., Payan, P. and Sardet, C. (1991). The calcium content of Nicaise, G., Maggio, K., Thirion, S., Horoyan, M. and Keicher, E. (1992).
cortical granules and the loss of calcium from sea urchin eggs at fertilization. The calcium loading of secretory granules. A possible key event in stimulus-
Dev. Biol. 146, 396-405. secretion coupling. Biol. Cell 75, 89-99.
Girard, J. P., Gillot, I., de Renzis, G. and Payan, P. (1991). Calcium pools in Olivo, J. C., Kahn, E., Halpern, S., Briançon, C., Fragu, P. and Di Paola, R.
sea urchin eggs: Roles of endoplasmic reticulum and mitochondria in relation (1989). Microcomputer system for ion microscopy digital imaging and
to fertilization. Cell Calcium 12, 289-299. processing. J. Microsc. 56, 105-114.
Glas-Albrecht, R., Kaesberg, B., Knoll, G., Allmann, K., Pape, R. and Pietrobon, D., Di Virgilio, F. and Pozzan, T. (1990). Structural and functional
Plattner, H. (1991). Synchronised secretory organelle docking in aspects of calcium homeostasis in eukaryotic cells. Eur. J. Biochem. 193,
Paramecium. Saltatory movement along microtubules transiently formed 599-622.
from ciliary basal bodies and selective exclusion of microinjected Plattner, H., Matt, H., Kersken, H., Haacke, B. and Stürzl, R. (1984).
heterologous organelles. J. Cell Sci. 100, 45-54. Synchronous exocytosis in Paramecium cells. I. A novel approach. Exp. Cell
Grynkiewicz, G., Poenie, M. and Tsien., R. Y. (1985). A new generation of Res. 151, 6-13.
Ca2+ indicators with greatly improved ﬂuorescence properties. J. Biol. Chem. Plattner, H., Stürzl, R. and Matt, H. (1985). Synchronous exocytosis in
360, 3440-3450. Paramecium cells. IV. Polyamino compounds as potent trigger agents for
Gualtieri, R. and Sardet, C. (1989). The endoplasmic reticulum network in repeatable trigger-redocking cycle. Eur. J. Cell Biol. 36, 32-37.
the Ascidian egg: Localization and calcium content. Biol. Cell 65, 301-304. Plattner, H., Lumpert, C. J., Knoll, G., Kissmehl, R., Höhne, B., Momayezi,
Henson, J. H., Begg, D. A., Beaulieu, S. M., Fishkind, D. J., Bonder, E. M., M. and Glas-Albrecht, R. (1991). Stimulus-secretion coupling in
Terasaki, M., Lebeche, D. and Kaminer, B. (1989). A calsequestrin-like Paramecium cells. Eur. J. Cell Biol. 55, 3-16.
protein in the endoplasmic reticulum of the sea urchin: Localization and Putney, J. W. Jr (1986). A model for receptor-regulated calcium entry. Cell
dynamics in the egg and ﬁrst cell cycle embryo. J. Cell Biol. 109, 149-161. Calcium 7, 1-12.
Iftode, F., Cohen, J., Ruiz, F., Torres Rueda, A., Chen-Shan, L., Adoutte, Rizzuto, R., Brini, M., Murgia, M. and Pozzan, T. (1993). Microdomains
A. and Beisson, J. (1989). Development of surface pattern during division in with High Ca2+ close to IP3-sensitive channels that are sensed by neighboring
Paramecium. I. Mapping of duplication and reorganization of cortical mitochondria. Science 262, 744-747.
cytoskeletal structures in the wild type. Development 105, 191-211. Sardet, C. (1984). The ultrastructure of the sea urchin cortex isolated before
Jorgensen, A. O. and Campbell, K. P. (1984). Evidence for the presence of and after fertilization. Dev. Biol. 105, 196-210.
calsequestrin in two structurally different regions of myocardial Sardet, C., Gillot, I., Ruscher, A., Payan, P., Girard, J. P. and de Renzis, G.
sarcoplasmic reticulum. J. Cell Biol. 98, 1597-1602. (1992). Ryanodine activates sea urchin eggs. Dev. Growth Differ. 34, 37-42
Jorgensen, A. O., Shen, A. C. Y. and Campbell, K. P. (1985). Ultrastructural Satir, B. H. and Wissig, S. L. (1982). Alveolar sacs of Tetrahymena:
localization of calsequestrin in adult rat atrial and ventricular muscle cells. J. Ultrastructural characteristics and similaritites to subsurface cisterns of
Cell Biol. 101, 257-268. muscle and nerve. J. Cell Sci. 55, 13-33.
Kerboeuf, D. and Cohen, J. (1990). A Ca2+ inﬂux associated with exocytosis Schmitz, M., Meyer, R. and Zierold, K. (1985). X-ray microanalysis in
is speciﬁcally abolished in a Paramecium exocytotic mutant. J. Cell Biol. cryosections of natively frozen Paramecium caudatum with regard to ion
111, 2527-2535. distribution in ciliates. Scanning Electron Microsc. 1, 433-445.
Knoll, G., Kerboeuf, D. and Plattner, H. (1992). A rapid calcium inﬂux Sitia, R. and Meldolesi, J. (1992). Endoplasmic reticulum: A dynamic
during exocytosis in Paramecium cells is followed by a rise in cyclic GMP patchwork of specialized subregions. Mol. Biol. Cell 3, 1067-1072.
within 1 s. FEBS Lett. 304, 265-268. Sod, E. W., Crooker, A. R. and Morrison G. H. (1990). Biological
Calcium stores visualized by SIMS in Paramecium 1909
cryosection preparation and practical ion yield evaluation for ion Tsien, R. W. and Tsien, R. Y. (1990). Calcium channels, stores, and
microscopic analysis. J. Microsc.160, 55-65. oscillations. Annu. Rev. Cell Biol. 6, 715-760.
Somlyo, A. V., Gonzalez-Serratos, H., Shuman, H., McClellan, G. and Volpe, P., Krause, K. H., Hashimoto, S., Zorzato, F., Pozzan T., Meldolesi,
Somlyo, A. P. (1981). Calcium release and ionic changes in the sarcoplasmic J. and Lew, D. P. (1988). ‘Calciosome,’ a cytoplasmic organelle: The
reticulum of tetanized muscle: an electron-probe study. J. Cell Biol. 90, 577- inositol 1,4,5-trisphosphate-sensitive Ca2+ store of nonmuscle cells? Proc.
594. Nat. Acad. Sci. USA 85, 1091-1095.
Somlyo, A. P. (1985). Cell calcium measurement with electron probe and Williams, D. A. and Fay, F. S. (1990). Intracellular calibration of the
electron energy loss analysis. Cell Calcium 6, 197-212. ﬂuorescent calcium indicator Fura-2. Cell Calcium 11, 75-83.
Stelly, N., Mauger, J. P., Claret, M. and Adoutte, A. (1991). Cortical alveoli Zierold, K., Gerke, I. and Schmitz, M. (1989). X-ray microanalysis of fast
of Paramecium: a vast submembranous calcium storage compartment. J. exocytotic processes. In Electron Probe Microanalysis. Applications in
Cell Biol. 113, 103-112. Biology and Medicine. Springer Series in Biophysics, vol. 4, pp. 281-292,
Terasaki, M. and Sardet, C. (1991). Demonstration of calcium uptake and Springer, Berlin.
release by sea urchin egg cortical endoplasmic reticulum. J. Cell Biol. 115, Zierold, K. (1991). Cryoﬁxation methods for ion localization in cells by
1031-1037. electron probe microanalysis: a review. J. Microsc. 161, 357-366.
Terasaki, M., Henson, J., Begg, D., Kaminer, B. and Sardet, C. (1991).
Characterization of sea urchin endoplasmic reticulum in cortical
preparations. Dev. Biol. 148, 398-401. (Received 12 August 1994 - Accepted 20 February 1995)