MICROSCOPY RESEARCH AND TECHNIQUE 46:370?379 (1999) Correlated Measurements of Free and Total Intracellular Calcium Concentration in Central Nervous System Neurons LUCAS D. POZZO-MILLER,1,2* NATALIA B. PIVOVAROVA,2,3 JOHN A. CONNOR,4 THOMAS S. REESE,2,3 AND S. BRIAN ANDREWS2,3 1Department of Neurobiology, University of Alabama at Birmingham, Birmingham, Alabama 35294 Biological Laboratory, Woods Hole, Massachusetts 02543 3Laboratory of Neurobiology, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20892 4Department of Neuroscience, University of New Mexico, Albuquerque, New Mexico 87131 2Marine KEY WORDS endoplasmic reticulum; mitochondria; Ca2? imaging; electron probe X-ray microanalysis ABSTRACT Transient changes in the intracellular concentration of free calcium ([Ca2?]i) act as a trigger or modulator for a large number of important neuronal processes. Such transients can originate from voltage- or ligand-gated fluxes of Ca2? into the cytoplasm from the extracellular space, or by ligand- or Ca2?-gated release from intracellular stores. Characterizing the sources and spatio-temporal patterns of [Ca2?]i transients is critical for understanding the role of different neuronal compartments in dendritic integration and synaptic plasticity. Optical imaging of fluorescent indicators sensitive to free Ca2? is especially suited to studying such phenomena because this approach offers simultaneous monitoring of large regions of the dendritic tree in individual living central nervous system neurons. In contrast, energy-dispersive X-ray (EDX) microanalysis provides quantitative information on the amount and location of intracellular total, i.e., free plus bound, calcium (Ca) within specific subcellular dendritic compartments as a function of the activity state of the neuron. When optical measurements of [Ca2?]i transients and parallel EDX measurements of Ca content are used in tandem, and correlated simultaneously with electrophysiological measurements of neuronal activity, the combined information provides a relatively general picture of spatio-temporal neuronal total Ca fluctuations. To illustrate the kinds of information available with this approach, we review here results from our ongoing work aimed at evaluating the role of various Ca uptake, release, sequestration, and extrusion mechanisms in the generation and termination of [Ca2?]i transients in dendrites of pyramidal neurons in hippocampal slices during and after synaptic activity. Our observations support the long-standing speculation that the dendritic endoplasmic reticulum acts not only as an intracellular Ca2? source that can be mobilized by a signal cascade originating at activated synapses, but also as a major intracellular Ca sink involved in active clearance mechanisms after voltage- and ligand-gated Ca2? influx. Microsc. Res. Tech. 46:370?379, 1999. r 1999 Wiley-Liss, Inc. INTRODUCTION In neurons of the mammalian central nervous system (CNS), transient changes in the intracellular concentration of free calcium ([Ca2?]i) following neuronal activity regulate important functions such as neuronal excitability, dendritic integration, and synaptic plasticity (reviewed by Berridge, 1998). For free Ca2? to act as an intracellular signal transducer, these transients must be efficiently buffered to resting levels rather quickly, but the processes that accomplish this are not well understood. Depending on the magnitude of the [Ca2?]i transients, neurons may employ any of several buffering, sequestering, and extrusion mechanisms, including binding to cytoplasmic proteins, ATP-mediated uptake into smooth endoplasmic reticulum (ER), electrochemically driven uptake into mitochondria, and extrusion across the plasma membrane by a Ca2? exchanger and a Ca2?-ATPase (reviewed by Carafoli, 1987; Miller, 1991). r 1999 WILEY-LISS, INC. Digital optical imaging of Ca2?-sensitive fluorescent indicators has become the method of choice for measuring spatio-temporal changes of [Ca2?]i in individual CNS neurons (reviewed by Connor, 1988; Connor et al., 1994). Cultures of dissociated embryonic cells (Connor, 1986; Murphy et al., 1987), acutely isolated neurons from young animals (Connor et al., 1988), and acute brain slices (Llano et al., 1991; Regehr et al., 1989; Tank et al., 1988), are the most commonly used in vitro preparations for optical imaging of [Ca2?]i in the mammalian CNS. In acutely isolated neurons and dissociated cell cultures, individual neurons are easily identiContract grant sponsor: Grass Foundation; Contract grant sponsor: Lakian Foundation; Contract grant sponsor: NIH Intramural Research Program; Contract grant sponsor: NIH; Contract grant number: NS35644. *Correspondence to: Lucas D. Pozzo-Miller, Ph.D., Department of Neurobiology, University of Alabama at Birmingham, Civitan International Research Center 429-A2, 1719 6th Avenue South, Birmingham, AL 35294. E-mail: [email protected] Received 28 January 1999; accepted in revised form 7 May 1999 FREE AND TOTAL CALCIUM IN CENTRAL NEURONS fied and the optical properties of the preparation are optimal. However, in the case of dissociated cell cultures from embryonic tissue, the separation and subsequent random reaggregation of the cells, leads to a neuronal network that lacks the characteristic features of the intact brain region from which they were derived. The acute slice preparation retains much of the original cellular and synaptic organization of the brain region in situ, but the high degree of absorbance and scattering produced by damaged tissue at the slice surface presents limitations for the application of high-resolution optical imaging of fluorescent signals. Organotypic slice cultures (reviewed by Ga?hwiler et al., 1997) are suitable preparations for combined electrophysiological and optical imaging in individual CNS neurons. Hippocampal slice cultures maintained in vitro over a period of several days undergo a progressive thinning, thereby reducing scattering and absorbance from extraneous tissue while retaining much of the cellular and synaptic connectivity of the hippocampal cortex in situ (Ga?hwiler 1984; Pozzo-Miller et al., 1993; Stoppini et al., 1991). Pyramidal neurons in such slice cultures mature and differentiate, acquiring morphological (Buchs et al., 1993; Pozzo-Miller and Landis, 1993; Pozzo-Miller et al., 1997; Zimmer and Ga?hwiler, 1984) and physiological (Charpak and Ga?hwiler, 1991; Ga?hwiler, 1984; Llano et al., 1988; Malouf et al., 1990; Muller et al., 1993; Petrozzino et al., 1995; Pozzo-Miller et al., 1993, 1995, 1996, 1999) characteristics similar to those in acute slices from adult animals. Thus, slice cultures offer a favorable preparation for optical imaging of Ca2?-sensitive fluorescent dyes in an organotypically preserved neuronal network (Connor et al., 1994; Petrozzino et al., 1995; Pozzo-Miller et al., 1993, 1996, 1999). It is well known that the preponderance of cellular calcium (Ca) is in the bound form, with typical bound/ free ratios on the order of 103 (McBurney and Neering, 1987; Neher, 1995). Because cytosolic and intraorganelle pools of bound Ca are so large, they exert a profound influence over the magnitude, shape, and time course of [Ca2?]i transients. While free Ca2? transients are the direct effectors of many important neuronal functions, the bound Ca pool nevertheless comes into play by, for example, siphoning large fractions of Ca2? currents away from the signaling pool or by sequestering significant amounts of Ca within noncytosolic compartments. The magnitude and importance of such effects can vary widely, in accordance with the concentrations, distributions, affinities, and kinetics of Ca2?-binding proteins, pumps, and leaks. Thus, it is uniquely informative to measure how concentrations of total (bound plus free) calcium ([Ca]) in specific subcellular locations change in parallel with [Ca2?]i. We illustrate this point here by using defined changes in [Ca] to identify intradendritic Ca sequestration sites. Energy-dispersive X-ray (EDX) microanalysis, carried out on cryosections of rapidly frozen tissues, was the first and is still the most reliable method for quantitatively measuring the concentrations of Ca, as well as of other physiologically important elements such as Na and K, within identified subcellular compartments. The physical principles and instrumentation for rapid freezing (Heuser et al., 1976; Van Harreveld and Crowell, 1964) and EDX analysis (Hall and Gupta, 371 1983) have been known and available for decades; results obtained from the combination of these methodologies, as pioneered by the Somlyos and colleagues (Shuman et al., 1976; Somlyo et al., 1977), have contributed much to our current understanding of the ionic changes that underlie cell physiology. Nonetheless, EDX microanalysis requires highly specialized instrumentation and faces several technical problems that are peculiar to biological specimens. Thus, EDX results have had to date only limited impact in the field of neuronal Ca regulation. Over the last decade, however, several major problems have been largely ameliorated by significant advances in instrumentation and techniques. Among the most important of these are: (1) the ability to prepare high-quality truly ultrathin cryosections of unfixed, rapidly frozen tissues that have structural detail comparable to conventional preparations (Buchanan et al., 1993; Michel et al., 1992), while still retaining native distributions of diffusible elements; and (2) improved detectors and analytical microscopes (e.g., Leapman et al., 1993) that can achieve, in a clean environment, the high sensitivity and resolution necessary to characterize biologically relevant changes in cellular [Ca]. In the present communication, we review our studies aimed at examining the role and relative importance of the diverse Ca2? buffering and extrusion mechanisms in neurons of mammalian brain slices. To begin that investigation, we first characterized the conditions necessary to elicit widespread dendritic Ca2? entry by using simultaneous digital imaging of Ca2?-sensitive dyes and electrical recording from pyramidal neurons in hippocampal slice cultures during high-frequency afferent fiber stimulation. We then replicated those conditions in parallel experiments, rapidly froze the slices at defined times after the high-frequency stimulation of afferent axons, and carried out EDX microanalysis in ultrathin cryosections to determine the amount and distribution of total Ca in pyramidal dendrites. The results illustrate well the advantages of combining microfluorometric digital optical imaging of intracellular free Ca2? concentrations using fluorescent dyes with high-resolution EDX microanalysis of total Ca concentrations at the electron microscopic level. FREE [Ca2?]i MEASUREMENTS USING FLUORESCENT DYES AND DIGITAL OPTICAL IMAGING WITH SIMULTANEOUS ELECTROPHYSIOLOGICAL RECORDING Pyramidal neurons in slice cultures had resting membrane potentials (-57?2 mV, mean ? SEM), input resistances (50 ? 8 M?; Fig. 1A), and resting levels of [Ca2?]i (49 ? 5 nM) similar to those from acute hippocampal slices (Brown and Johnston, 1983; Regehr et al., 1989; Schwartzkroin, 1975). In response to a depolarizing current pulse applied through the recording electrode, they exhibited the characteristic train of action potentials, spike adaptation, and slow afterhyperpolarization (Fig. 1B). Figure 1C shows the [Ca2?]i transient in a neuron dialysed with the Ca2? indicator bis-fura-2 via the recording pipette in the patch-clamp whole-cell configuration, evoked by a train of action potentials in response to a depolarizing current pulse. Figure 1C (left) shows a fluorescence image (top) and a pseudocolor map of the [Ca2?]i (bottom) at 250 ms from the Fig. 1. Fig. 2. FREE AND TOTAL CALCIUM IN CENTRAL NEURONS beginning of the depolarizing pulse, computed using the pixel-by-pixel ratio of bis-fura-2 fluorescence. Figure 1C (right) illustrates the time course of the [Ca2?]i transient in two dendritic and one somatic location (colored boxes in the fluorescence image) temporally aligned with the train of action potentials, and the depolarizing current injection (bar). The faster rise time and larger amplitude of the dendritic [Ca2?]i transients compared to those at the somatic location provides additional evidence that the high (L-type) voltage-dependent Ca2? channels activated by the membrane depolarization are indeed preferentially clustered in proximal dendritic processes (Westenbroek et al., 1990). Fig. 1. Membrane electrical properties and voltage-dependent Ca2? transients in a pyramidal neuron in hippocampal slice culture. A: Upper trace shows the membrane voltage response to a hyperpolarizing current pulse (lower trace) under current-clamp in the whole-cell patch-clamp configuration, used to estimate the input resistance of the neuron. B: Upper traces show the membrane voltage responses to depolarizing current pulses (lower traces). The characteristic action potential frequency adaptation is observed on the left trace. The expanded view on the right shows the typical afterhyperpolarization that follows action potential trains. C: Membrane voltage response and associated Ca2? transients evoked by a depolarizing current pulse. Left: Fluorescence image at 357 nm excitation (top), and a pseudocolor map calculated from the pixel-by-pixel ratio of background-subtracted fluorescence images obtained at 357 nm (isosbestic point, Ca2?insensitive) and 380 nm (Ca2?-sensitive) excitation (bottom). Intracellular Ca2? concentration traces shown on the right (at 50 ms frame interval) were calculated from the regions of interests marked with boxes in the fluorescence image after background correction from a region over the slice but outside the dye-filled cell. Membrane voltage trace shows the train of action potentials evoked by the depolarizing current pulse (1 second, 0.1 nA; black bar). The membrane voltage response was digitized at 10 kHz, acquired synchronously with Ca2? concentration traces (at 50-ms interval), and is shown with the same time base. The spatial location of the color traces of Ca2? concentration correspond to the colored boxes in the fluorescence image (left, top). All experiments were performed in normal extracellular buffer, and using patch pipettes containing K?-based intracellular solution and 500 然 of the Ca2? indicator bis-fura-2 (Molecular Probes); the pipette can be seen attached to the soma in the fluorescence image (top). All simultaneous electrophysiological recording and microfluorometric digital imaging of Ca2?-sensitive dyes were performed in pyramidal neurons maintained in cultured hippocampal slices using sharp microelectrodes or whole-cell patch pipettes in inverted or upright microscopes (Axiovert, or Axioskop FS, Zeiss), using dry (20X, 0.5NA) or water-immersion (63X, 0.9NA) objectives, and equipped with 12-bit, digital, cooled charged-coupled device (CCD) camera systems (Series 200, or PXL-37; Photometrics; Petrozzino et al., 1995; Pozzo-Miller et al., 1993, 1996, 1999). Alternating dye excitation light was provided by filter wheels (Lambda-10, Sutter Instruments), or by a galvanometric monochromator (Polychrome I, TILL Photonics). Electrophysiological recordings were performed using current-clamp amplifiers (Axoclamp2A, Axon Instruments). Synchronized electrical and optical recording, afferent stimulus trigger, and control of the cooled CCD camera and galvanometric monochromator was performed using a single Power Macintosh computer (9500/120 MHz), running custom-developed software (TI Workbench). Fig. 2. Synaptically activated free Ca2? transient in a pyramidal cell in hippocampal slice culture. A: Membrane voltage response to one 500 ms tetanus at 60 Hz given to presynaptic afferent fibers. Sharp microelectrode recording under current-clamp at resting membrane potential. B: Same evoked response as in A, but on a longer time scale to show the prolonged afterhyperpolarization. C: Fluorescence ratios of the fura-2 signals, before (a) and after (b? k) the tetanus, which evoked the response shown in B. The tip of the sharp recording microelectrode contained 10 mM of the Ca2? indicator fura-2 in 200 mM KCl, and was backfilled with 3 M KCl; it appears as blue wedge on the lower left. The electrical and optical recording was performed as described in Figure1. Modified with permission from Pozzo-Miller et al. (1993). 373 Endogenous excitatory neurotransmitter released from presynaptic afferent fibers within the slice can elicit suprathreshold membrane depolarizations that trigger regenerative Na?-dependent action potential discharge in postsynaptic pyramidal neurons. Highfrequency stimulation of these afferent axons activates several mechanisms of dendritic Ca2? entry in postsynaptic pyramidal neurons, including influx through locally-activated, low voltage-gated Ca2? channels (Magee et al., 1995) and through the N-methyl-D-aspartate (NMDA) subtype of glutamate receptors (Malinow et al., 1994; Perkel et al., 1993; Petrozzino et al., 1995; Pozzo-Miller et al., 1996, 1999), as well as more widespread voltage-gated influx triggered by back-propagating, Na?-dependent dendritic action potentials (Jaffe et al., 1992; Miyakawa et al., 1992; Spruston et al., 1995). Figure 2A shows the electrical response evoked in a postsynaptic pyramidal neuron filled with the Ca2? indicator fura-2 via the recording microelectrode, by a single high-frequency stimulation to the afferent presynaptic fibers indicated by the arrow; Figure 2B shows the same response on a slower time scale. A series of action potentials followed by a slow (?10 second) membrane depolarization and a prolonged (?1 min) afterhyperpolarization, were evoked by the high-frequency afferent stimulation. This robust, long-lasting action potential firing was accompanied by a rapid change of [Ca2?]i in the primary dendrites (to 290 ? 60 nM), and a delayed [Ca2?]i transient in the somatic regions (to 220 ? 60 nM). Proximal dendrites were the first regions where [Ca2?]i transients occurred (Fig. 2C). Panel b in Figure 2C shows an image acquired between the trigger of the stimulus and 600 ms after the synaptic stimulation, displaying the rapid dendritic [Ca2?]i response to synaptic stimulation. A somatic [Ca2?]i response was observed (panel c), including within the cell nucleus, only several seconds after the initiation of the stimulation (panel d). The [Ca2?]i response decayed in all dendritic and somatic regions with a similar time course (panels e to g), returning to resting levels 15 seconds after triggering of the synaptic stimulus (panel h). This decay corresponded to the time course of the slow depolarization evidenced in the membrane potential recordings (Fig. 2A and B). No changes in [Ca2?]i were observed during the prolonged hyperpolarization, suggesting that a persistent intracellular free Ca2? signal is not required to drive this process. The long-lasting synaptic responses to tetanic stimulation in the slice cultures might arise as a consequence of the particular circuitry that has developed in the isolated slices after weeks in vitro. Activation of the B subtype of ?-aminobutyric acid receptors (GABAB) has been implicated in the agonist- (Ga?hwiler and Brown, 1985) and synapticallyinduced (Malouf et al., 1990) slow hyperpolarization observed in pyramidal neurons in slice cultures. Interestingly, the time course of the long-lasting depolarizations induced by synaptic stimulation in our slice cultures resemble those observed in response to bath application of glutamate and selective metabotropic glutamate receptor (mGluR) agonists (Charpak and Ga?hwiler, 1991), indicating the involvement of mGluRs in the prolonged depolarizations and [Ca2?]i transients induced by afferent fiber stimulation in our cultures (Pozzo-Miller et al., 1995, 1996). 374 L.D. POZZO-MILLER ET AL. EXPERIMENTAL CONSIDERATIONS REGARDING INTRACELLULAR FLUORESCENT INDICATORS Fluorescent indicator molecules for Ca2? ion concentration have revolutionized the study of intracellular Ca2? transients in small neuronal processes. So far, two methods have been used to introduce the indicator molecules into living cells: bath-loading of membranepermeable ester derivatives (AM forms), and passive diffusion of the membrane-impermeant forms (K? salts) from a patch pipette (or iontophoresis from a sharp microelectrode). The potential drawbacks of both techniques of indicator loading have been reviewed (Connor, 1988), and experimentally addressed (Connor, 1994). They include: (1) excessive Ca2? buffering, which may affect normal neuronal processing; (2) association with intracellular constituents, which may affect indicator binding and spectral properties; (3) alteration of absorption and fluorescence spectra by physical factors such as cytoplasmic viscosity; and (4) transport of indicator molecules out of the cell or into intracellular compartments. The self-loading of membrane-permeant forms adds another set of potential artifacts, such as: (1) incomplete de-esterification, leading to a mix of responsive and unresponsive indicator; (2) deesterification and trapping within organelles, resulting in artifactual report of Ca2? concentrations; and (3) photobleaching (more severe in bath-loading conditions, since after loading and de-esterification there is a limited amount of indicator within the cell, as compared to the unlimited supply of indicator when diffusing from a patch pipette). If caution is exercised to control some of these experimental drawbacks, i.e., by careful calibration of the experimental setup, rational use of indicator concentrations, and limited UV illumination, optical imaging of Ca2? concentration using ratiometric fluorescent indicators will provide meaningful quantitative measurements of Ca2? dynamics in MEASUREMENT OF TOTAL CALCIUM CONCENTRATIONS WITHIN SUBCELLULAR DENDRITIC ORGANELLES BY EDX MICROANALYSIS IN THE ELECTRON MICROSCOPE EDX microanalysis ? performed as depicted schematically in Figure 3, and described in detail elsewhere (Andrews et al., 1994; Buchanan et al., 1993; Leapman and Andrews, 1991; Pozzo-Miller et al., 1997) ? was used to obtain direct measurements of persistent changes in the distribution of total Ca within dendritic compartments of pyramidal neurons after synaptic stimulation. We focused on the three subcellular compartments that comprise essentially the entirety of these dendrites, namely, endoplasmic reticulum (ER), mitochondria, and cytoplasm. These dendritic compartments can be readily identified in unstained, ultrathin cryosections (e.g., Fig. 3E). EDX analysis of cryosections from unstimulated slices, rapidly frozen either in control buffer or in the presence of the Na? channel blocker tetrodotoxin (TTX, 5 然, to block action potential-dependent spontaneous synaptic transmission) revealed that the concentrations of Ca in all three dendritic compartments (Fig. 4) were comparable to basal concentrations in a variety of other cells (Andrews et al., 1988; Somlyo et al., 1985). The mean concentration of Ca within dendritic ER ([Ca]ER) was 5.1 ? 1.1 mmol/kg dry weight. The units mmol/kg dry weight are used throughout for total Ca concentrations because these units follow naturally from EDX data processing procedures. This concentration unit is approximately related to a more familiar unit ? mmol/l of hydrated cell volume ? according to the dry mass fraction of a given cell compartment. For ER, the relationship [Ca]ER (mmol/l hydrated cell volume) ? 0.25 x [Ca]ER (mmol/kg Fig. 3. Schematic diagram illustrating the main steps in the EDX microanalysis. Sequence starts at upper left and proceeds clockwise. A: Following defined stimulation in the recording chamber, the filter membrane supporting the slice culture on a Millicell-CM (Millipore) culture plate insert is cut out, and using the membrane as a handle so as to avoid touching the tissue, it is placed on an agar/gelatin pad already mounted on a specially designed freezing disc that fits both a modified LifeCell CF-100 cold-metal-block freezing machine, and the arms of Leica cryoultramicrotomes. A few seconds before freezing, extraneous is wicked from the tissue, again being careful not to touch the slice. The assembly is magnetically mounted onto the freezing machine and frozen at controlled times. Special culture plate inserts with pre-attached nichrome stimulating wires, as well as custom freezing stages to accommodate these inserts without the necessity of cutting the membrane, have been designed for electrical stimulation of the slices directly on the stage of the freezing machine. This permits times as short as 1 second between delivery of synaptic stimulation and the instant of impact on the freezing block. B: Frozen cultures are transferred under liquid nitrogen to a Leica Ultracut S/FCS or Ultracut 4E/FC4E cryoultramicrotome, mounted in en face orientation, trimmed to a 250 ? 250 痠 block face (maximum). Ribbons of ultrathin cryosections (nominal thickness typically ?80nm) were prepared from the superficial 5?20 痠 of the hippocampal slice cultures. C: Ribbons were manipulated with an eyelash probe and transferred onto carbon- and formvar-coated, 200-mesh, thin-bar, copper EM grids. The grids were cryotransferred into an analytical electron microscope, either a VG HB501 dedicated STEM or a LEO 912 Omega TEM equipped with Proscan 1024 ? 1024 slow-scan CCD camera. Cryosections were freeze-dried in the microscope at a specimen stage temperature of ? ?100蚓 over the course of 1?2 hours, then recooled to below ?170蚓 for imaging and analysis. D: Images of freeze-dried cryosections are recorded under relatively low-dose conditions (?103 e-/nm2), using digital dark-field detection in the VG STEM or the zero-loss signal and CCD detection in the LEO 912 (diagram illustrates the VG STEM). E: Morphological detail can be quite good in micrographs of satisfactory cryosections, as illustrated by the lowdose, digital dark-field STEM image of the proximal dendritic shaft of a pyramidal neuron. Cytoplasm rich in microtubules (MT), elongated mitochondria (M), and cisterns of ER in both long- and cross-section (arrows) are easily identified. Micrographs are evaluated for tissue integrity and freezing quality, and optimal areas of dendrites selected for EDX analysis. The usual probe size for EDX microanalysis is indicated by the dot within the box. Bar ? 1痠. F: Lastly, electron beam intensity is increased to ?5 nA and focused to a small probe on an area of interest, so that characteristic and continuum x-rays are generated by inelastic collisions. Probe size is typically 100 nm or less, depending on the size of the target organelle; the usual probe size for analysis of ER content is indicated by the dot within the box on E. Emitted X-rays are captured for ?100 sec (total dose ?108 e-/nm2) by an ultrathin-window energy-dispersive detector (Tracor Micro-ZHV or Oxford Pentafet), and sorted by a dedicated multichannel analyzer with associated electronics to produce an EDX spectrum that, like the one illustrated here, is essentially a plot of X-ray counts vs. energy. All biologically interesting elements between C (Z ? 6) and Ca (Z ? 20) produce X-rays at characteristic energies within the energy range 0?10 keV (see labels on F), which can be detected, identified, and quantified. Subsequent processing of spectra, as described (Shuman et al., 1976; Buchanan et al., 1993; Pozzo-Miller et al., 1997), provides compartmental concentrations of elements in units of mmol/kg dry weight. living neurons and their finest processes, such as dendrites and spines. FREE AND TOTAL CALCIUM IN CENTRAL NEURONS dry weight), so that 5.1 ? 1.1 mmol/kg dry weight for control [Ca]ER becomes ?1.3 mmol/l. These relationships are discussed in more detail in Buchanan et al. (1993). Following the stimulation procedures outlined above for measurements of free [Ca2?]i, afferent fibers were given either a single high-frequency train of stimuli, or four such tetani at 30-second intervals. In slices rapidly frozen 3 minutes after a single tetanus, i.e., at a time well after the decay of free Ca2? transients (see Fig. 2 375 and Petrozzino et al., 1995; Pozzo-Miller et al., 1993), [Ca]ER had increased to ?15 mmol/kg dry weight in 40% of dendritic ER cisterns; this population of cisterns had a mean of 41 ? 7 mmol/kg (Fig. 4). The Ca content of the remaining 60% was unchanged. We have proposed that the 40% of cisterns showing large Ca increases must represent a buffering component of the ER. It is noteworthy that responsive and non-responsive ER cisterns can exist quite close to each other within the same dendritic segment (Pozzo-Miller et al., 1997). Given the intercon- Fig. 3. 376 L.D. POZZO-MILLER ET AL. Fig. 4. Histograms comparing Ca concentrations within hippocampal dendritic organelles 3 minutes after different stimulation protocols. Only Ca-accumulating ER takes up large amounts of Ca following synaptic stimulation by 1 or 4 action potential-inducing trains (50 Hz, 1 second) to afferent axons (1 Train and 4 Trains). Ca2? uptake was graded with increased stimulation (1 Train vs. 4 Trains), and was inhibited by blocking synaptic transmission with tetrodotoxin (4 Trains ? TTX) or by blocking SERCA pumps with thapsigargin (4 Trains ? TG). In the presence of thapsigargin Ca2? uptake was diverted to dendritic mitochondria, within which Ca was localized as small ??hot spots,?? termed inclusions, contained within an otherwise low-Ca matrix. Total Ca content in ER cisterns returned to resting levels after a 15?30 minute recovery period (4 Trains ? Rec). There were no large changes in the Ca content of any other organelle at this time point. Asterisks indicate values that were significantly different (P ? 0.05) from appropriate controls. Modified from Pozzo-Miller et al. (1997). nected nature of the dendritic ER network, it is likely that, at least in some instances, these functionally distinct cisterns are actually in physical continuity. This possibility provides further support for the emerging realization (Golovina and Blaustein, 1997; Meldolesi and Pozzan, 1998) that functional regio-specialization of the ER, as exemplified by the terminal cisternae of skeletal muscle sarcoplasmic reticulum, may be more common than is implied by the continuity and structural homogeneity of the ER network. In contrast to the large increase in Ca within some ER cisterns, the concentration of cytoplasmic total Ca ([Ca]cyto), increased only slightly (not significant). At this time point, there was no significant change in the Ca concentration of dendritic mitochondria ([Ca]mito), nor in any other analyzed element in any cellular compartment (Pozzo-Miller et al., 1997). Three minutes after 4 tetani at 30-second intervals, [Ca]ER increased to a mean of 72 ? 8 mmol/kg, in the 45% of ER cisterns that exceeded 15 mmol/kg (Fig. 4). The Ca concentrations achieved in the most avidly accumulating cisterns occasionally exceeded 100 mmol/ kg, which is an extraordinarily high level. As with the single afferent train, there was a residual pool of non-responsive ER cisterns, which was reduced only slightly by the recruitment of cisterns into the responsive pool; this suggests that inactive cisterns do not have the capacity to accumulate Ca. There was still no increase in [Ca]mito, and only a modest rise in [Ca]cyto at this time point. Following four tetani at 30-second intervals in the presence of 5 然 TTX, no Ca uptake occurred in any cellular compartment (Fig. 4), demonstrating that Ca sequestration in ER was a strict consequence of synaptic activity. In addition, [Ca]ER completely returned to basal levels after a recovery period of ?15 minutes from the last high-frequency stimulus (Fig. 4). These results identify a specific component of ER as the high-capacity Ca buffer in hippocampal dendrites. This finding is largely consistent with previous work demonstrating that the ER is a major physiological sink for Ca (Henkart et al., 1978; Horikawa et al., 1998; Somlyo et al., 1985; reviewed in Carafoli et al., 1995; Meldolesi and Pozzan, 1998; Pozzan et al., 1994). However, three new findings emerge. First, the levels of [Ca]ER attained by some cisterns in hippocampal neurons is extraordinary, reaching levels otherwise found only in the terminal cisternae of skeletal muscle sarcoplasmic reticulum (Somlyo et al., 1977). Even so, this large accumulation is apparently within the range of normal physiological activity for these dendrites, since the associated electrical responses and [Ca2?]i transients (?500 nM) are well within physiological levels (Fig. 2), the parallel elevation of cytoplasmic total Ca is in comparison quite small (Fig. 4), and ER Ca sequestration was reversible (Fig. 4). Second, there is a clear functional distinction between Ca-sequestering and non-sequestering elements of ER, even though cisterns expressing this difference are structurally and spatially indistinguishable. The invariant Ca content of the non-responsive ER pool under conditions that induce large changes in both free and total intracellular calcium concentration suggests that this pool is not capable of Ca accumulation. These findings support the concept that regional specialization of the ER is an important and common feature of many cell types, including neurons. Finally, the recovery phase of [Ca]ER has a surprisingly long lifetime, on the order of 15?30 minutes. Since ER stores will presumably be loaded and primed for Ca2?- or ligandgated release during this phase (see Pozzo-Miller et al., 1996), it is possible that ER cisterns carry a memory trace of previous synaptic activity over the time frame of minutes. Our results detected negligible mitochondrial accumulation at the described time (3 minutes after synaptic stimulation) and conditions of synaptically-mediated Ca2? entry, which is interesting and perhaps surprising in view of increasing evidence that mitochondria can play a major role in neuronal Ca2? regulation (Babcock et al., 1997; Friel and Tsien, 1994; Herrington et al., FREE AND TOTAL CALCIUM IN CENTRAL NEURONS 377 Fig. 5. Estimated time course curves illustrating hypothetical temporal relationships between intracellular free Ca2? and total Ca concentrations in various dendritic compartments of pyramidal neurons following synaptic activity-evoked Ca2? influx. The diagram emphasizes the following points: (1) both mitochondria and a subset of ER can respond with a large increase in their Ca content; (2) in both organelles, the increased Ca load evoked by Ca2? entry persists long after the elevation and decay of the [Ca2?]i transient, with the ER Ca load being particularly longlived; and (3) ER and mitochondria appear to accumulate and release Ca in distinct, and perhaps co-operative, time domains. Traces for total Ca concentrations were generated by extrapolating smooth curves between real data points at 0, 0.5, 1, 3, and 15 minutes. The trace for free Ca2? concentration was drawn after representative responses from real bisfura-2 imaging experiments using an acquisition rate of 20 frames-per-second. Note different scales and units on Y-axes. 1996). Mitochondrial Ca2? transport can in fact be brought into play, as revealed the action of thapsigargin, an inhibitor of ER Ca2? (SERCA) ATPase pumps (Fig. 4). Ca sequestration activity of the ER was abolished by incubation with 10 然 thapsigargin, and under these conditions Ca sequestration is diverted to mitochondria, many of which exhibited large focal increases in Ca (Fig. 4). Furthermore, when hippocampal slices were rapidly frozen 30 seconds after a single high frequency stimulus, i.e., just after termination of the free dendritic Ca2? transient (see Fig. 2 and Petrozzino et al., 1995; Pozzo-Miller et al., 1993), both the ER and mitochondria show significant increases in Ca content as compared to unstimulated controls, with [Ca]mito ?5 mmol/kg dry weight in approximately 50% of mitochondria, even while [Ca]ER continues to rise (Pivovarova et al., 1997). It appears that the ER and mitochondria of these pyramidal neuron dendrites may engage in a coordinated interaction to ensure control over [Ca2?]i. The emerging view of neuronal calcium homeostasis, illustrated schematically in Figure 5, is similar to previous models (Lawrie et al., 1996; Rizzuto et al., 1993, 1994) and supports the idea that the ER and mitochondria participate in a highly choreographed interplay, each dominating neuronal Ca2? regulation over specific [Ca2?]i and time regimes, as determined by the characteristics of their respective pumps, channels, and transporters. Ca2? buffering is another aspect of neuronal Ca2? regulation that has received recent attention (Neher, 1995). While a detailed discussion of this topic is beyond the scope of this article, we would like to point out that EDX measurements of total Ca concentrations, specifically [Ca]cyto, are potentially useful in this context. Because, as mentioned, EDX directly detects the ?99% of Ca, which is not in the free Ca2? pool, tandem measurements of [Ca]cyto and [Ca2?]i can be used to infer Ca2? binding ratios, as well as the dependence of this parameter on cellular Ca loads. As an example, it is straightforward to calculate the bound/free Ca ratio of dendritic cytosol at rest as follows: an average [Ca]cyto of 0.7 mmol/kg dry weight (Fig. 4) is equivalent to 120 然, calculated as described (Buchanan et al., 1993; and assuming a cytoplasmic water fraction of 85%); given that [Ca2?]i ??0.1 然, the bound/free Ca ratio is ?1200. During activity-dependent Ca2? influx, the most dramatic changes occur in the free Ca2? pool. For example, stimulations that increase [Ca2?]i by about one order of magnitude above resting levels (from ?0.1 to ?1然) will increase [Ca]cyto only a factor of about 4 above the normal range (from ?1 to ?4 mmol/kg dry weight). The reason(s) for this decrease in bound/free Ca ratio (from ?1,000 to ?500) with increased Ca2? influx is an open question, although we may speculate that it is due, at least in part, to a transition into a non-linear buffering regime and/or to a significant reservoir of tightly bound Ca. Future experiments will hopefully clarify the cellular basis for the apparently dynamic relationship between free and total Ca. TECHNICAL ASPECTS OF CALCIUM ANALYSIS IN THE ELECTRON MICROSCOPE The studies reviewed here represent the first direct correlation of the total Ca concentration in Ca2?regulating subcellular compartments with the physiological activity of central vertebrate neurons. Although EDX microanalysis is a well-established technique (Somlyo et al., 1985) for quantitatively measuring elemental concentrations with subcellular resolution, the refined level of analysis demanded in studies such as this one have been greatly facilitated by recent technical advances in biological EDX (Buchanan et al., 1993; Leapman and Andrews, 1991). These advances have largely eliminated the problems that plagued many early attempts at biological microanalysis using immature techniques and instruments. Modern EDX analysis now offers accurate and reliable performance, with a demonstrated sensitivity of fewer than 50 atoms of Ca in very small (?104 nm2) analytical volumes (Shi et al., 1996). For measuring total Ca in biological specimens, there are only a limited number of alternative approaches. Secondary ion mass spectrometry (SIMS; Chandra et al., 1989) can detect low elemental concentrations but 378 L.D. POZZO-MILLER ET AL. has spatial resolution that limits subcellular analysis. Another emerging technique is electron energy loss spectroscopy (EELS; reviewed by Leapman et al., 1994), which has been experimentally demonstrated (Leapman et al., 1993; Shuman and Somlyo, 1987) to be 3?4 times more sensitive than EDX for Ca analysis, with better spatial resolution. The improved sensitivity of EELS translates to a time savings of ?15X for equivalent statistical precision, making it feasible to quantitatively map, rather than just point probe, Ca distributions at subcellular resolution. Mapping provides an intuitively useful structural context for analytical data, and eliminates the need to subjectively choose a few locations for analysis in the hope that these are representative of the cell as a whole. There are two complementary approaches to element mapping in the electron microscope, STEM/EELS spectrum imaging (Hunt and Williams, 1991; Jeanguillaume and Colliex, 1989), and energy-filtered transmission EM (EFTEM; Grohovaz et al., 1996; Ottensmeyer, 1986). While both have been used with some preliminary success, STEM spectrum-imaging would appear to have an advantage for applications involving physiologically relevant concentrations of Ca, since Ca produces complex signals with low peak/background ratios that require the full range of spectral information to quantify properly (Leapman et al., 1993). CONCLUSIONS The correlated application of optical digital imaging of intracellular free Ca2? concentrations using fluorescent dyes combined with high-resolution EDX measurements of the subcellular distribution of total Ca has directly demonstrated that in dendrites of hippocampal pyramidal neurons, a subset of specific ER cisterns is the major subcellular compartment responsible for Ca sequestration in the minutes following neuronal activity. They also demonstrate that the Ca load associated with dendritic Ca2? transients persists for many minutes following the decay of these transients, and that within this time frame, the ER is the dominant sequestration compartment. The results have further characterized the following properties of the ER network in relation to Ca2? homeostasis: (1) ER Ca2? buffering is dependent on and graded with neuronal activity; (2) it is extremely powerful yet still reversible, even in extreme instances of repeated high-frequency afferent activity; and (3) it is dependent on a SERCA-type Ca2? pump. The last observation suggests that ER Ca sequestration and inositol-1,4,5-inositol trisphosphate- and/or calcium-induced Ca2? release may be alternative functions of the same organelles. Taken together, these observations support the long-standing speculation (Henkart et al., 1978) that the dendritic ER not only acts as an intracellular source of Ca2? that can be mobilized by a signal cascade originated at activated synapses, but also as the major intracellular sink involved in active clearance mechanisms following voltage- and ligand-gated Ca2? influx. ACKNOWLEDGMENTS We thank M.F. O?Connell for expert cryosectioning, J. Chludzinski, R. Sun, and L. Verselis for technical assistance, Dr. T. Inoue for image acquisition and analysis software, and Dr. R.D. Leapman for invaluable advice and discussions. This work was supported by Fellowships at the Marine Biological Laboratory from the Grass Foundation and the Lakian Foundation awarded to L.D.P.-M., the NIH Intramural Research Program, and NIH Research Grant NS35644 awarded to J.A.C. REFERENCES Andrews SB, Leapman RD, Landis DMD, Reese TS. 1988. Activitydependent accumulation of calcium in Purkinje cell dendritic spines. Proc Natl Acad Sci USA 85:1682?1685. Andrews SB, Buchanan RA, Leapman RD. 1994. Quantitative darkfield mass analysis of ultrathin cryosections in the field-emission scanning transmission electron microscope. Scan Microsc 8:13?24. Babcock DF, Herrington J, Goodwin PC, Park YB, Hille B. 1997. Mitochondrial participation in the intracellular Ca2? network. J Cell Biol 136:833?844. Berridge MJ. 1998. Neuronal calcium signaling. Neuron 21:13?26. Brown TH, Johnston D. 1983. Voltage-clamp analysis of mossy fiber synaptic input to hippocampal neurons. J Neurophysiol 50:487?507. Buchanan RA, Leapman RD, O?Connell MF, Reese TS, Andrews SB. 1993. Quantitative scanning transmission electron microscopy of ultrathin cryosections: subcellular organelles in rapidly frozen liver and cerebellar cortex. J Struct Biol 110:244?255. Buchs P-A, Stoppini L, Muller D. 1993. Structural modifications associated with synaptic development in area CA1 of rat hippocampal organotypic cultures. Dev Brain Res 71:81?91. Carafoli E. 1987. Intracellular calcium homeostasis. Ann Rev Biochem 56:395?433. Carafoli E Chiesi M, Gazzotti P. 1995. Membrane carriers related to intracellular calcium regulation. In: Laragh JH, Brenner BM, editors. Hypertension: pathophysiology, diagnosis and management. New York: Raven Press, p 1245?1259. Chandra S, Gross D, Ling YC, Morrison GH. 1989. Quantitative imaging of free and total intracellular calcium in cultured cells. Proc Natl Acad Sci USA 86:1870?1874. Charpak S, Ga?hwiler BH. 1991. Glutamate mediates a slow synaptic response in hippocampal slice cultures. Proc R Soc Lond B 243:221? 226. Connor JA. 1986. Digital imaging of free calcium changes and of spatial gradients in growing processes in single, mammalian central nervous system cells. Proc Natl Acad Sci USA 83:6179?6183. Connor JA. 1988. Fluorescence imaging applied to the measurement of Ca2? in mammalian neurons. In: Grinnell AD, Armstrong D, Jackson MB, editors. Calcium and ion channel modulation. New York: Plenum Press, p 395?406. Connor JA. 1994. Dissecting signals from Fura?2 in cytosol, nucleus, and other organelles in living cells. Neuroprotocols 5:25?34. Connor JA, Wadman WJ, Hockberger PE, Wong RKS. 1988. Sustained dendritic gradients of Ca2? induced by excitatory amino acids in CA1 hippocampal neurons. Science 240:649?653. Connor JA, Pozzo-Miller LD, Petrozzino JJ, Mu?ller W. 1994. Calcium signaling in dendritic spines of hippocampal neurons. J Neurobiol 25:234?242. Friel DD, Tsien RW. 1994. An FCCP-sensitive Ca2? store in bullfrog sympathetic neurons and its participation in stimulus-evoked changes in [Ca2?] i . J Neurosci 14:4007?4024. Ga?hwiler BH. 1984. Development of the hippocampus in vitro: cell types, synapses and receptors. Neuroscience 11:751?760. Ga?hwiler BH, Brown DA. 1985. GABAB-receptor-activated K? current in voltage-clamped CA3 pyramidal cells in hippocampal cultures. Proc Natl Acad Sci USA 82:1558?1562. Ga?hwiler BH, Capogna M, Debanne D, McKinney RA, Thompson SM. 1997. Organotypic slice cultures: a technique has come of age. Trends Neurosci 20:471?477. Golovina VA, Blaustein MP. 1997. Spatially and functionally distinct Ca2? stores in sarcoplasmic and endoplasmic reticulum. Science 275:1643?1648. Grohovaz F, Bossi M, Pezzati R, Meldolesi J, Tarelli FT. 1996. High resolution ultrastructural mapping of total calcium: electron spectroscopic imaging/electron energy loss spectroscopy analysis of a physically/chemically processed nerve-muscle preparation. Proc Natl Acad Sci USA 93:4799?4803. Hall TA, Gupta BL. 1983. The localization and assay of chemical elements by microprobe methods. Q Rev Biophys 16:279?339. Henkart MP, Reese TS, Brinley FJ Jr. 1978. Endoplasmic reticulum sequesters calcium in the squid giant axon. Science 202:1300?1303. FREE AND TOTAL CALCIUM IN CENTRAL NEURONS Herrington J, Park YB, Babcock DF, Hille B. 1996. Dominant role of mitochondria in clearance of large Ca2? loads from rat adrenal chromaffin cells. Neuron 16:219?228. Heuser JE, Reese TS, Landis DMD. 1976. Preservation of synaptic structure by rapid freezing. Cold Spring Harbor Symp Quant Biol 40:17?24. Horikawa Y, Goel A, Somlyo AP, Somlyo AV. 1998. Mitochondrial calcium in relaxed and tetanized myocardium. Biophys J 74:1579? 1590. Hunt J, Williams DB. 1991. Electron energy-loss spectrum imaging. Ultramicroscopy 38:47?73. Jaffe DB, Johnston D, Lasser-Ross N, Lisman JE, Miyakawa H, Ross WN. 1992. The spread of Na? spikes determines the pattern of dendritic Ca2? entry into hippocampal neurons. Nature 357:244? 246. Jeanguillaume C, Colliex C. 1989. Spectrum-image: the next step in EELS digital acquisition and processing. Ultramicroscopy 28:252? 257. Lawrie AM, Rizzuto R, Pozzan T, Simpson AW. 1996. A role for calcium influx in the regulation of mitochondrial calcium in endothelial cells. J Biol Chem 271:10753?10759. Leapman RD, Andrews SB. 1991. Analysis of directly frozen macromolecules and tissues in the field emission STEM. J Microsc 161:3?19. Leapman RD, Hunt JA, Buchanan RA, Andrews SB. 1993. Measurement of low calcium concentrations in cryosectioned cells by parallelEELS mapping. Ultramicroscopy 49:225?234. Leapman RD, Sun SQ, Hunt JA, Andrews SB. 1994. Biological electron energy loss spectroscopy in the field emission STEM. Scan Microsc 8:245?259. Llano I, Marty A, Johnson JW, Ascher P, Ga?hwiler BH. 1988. Patchclamp recording of amino acid-activated responses in ??organotypic?? slice cultures. Proc Natl Acad Sci USA 85:3221?3225. Llano I, Dreessen J, Kano M, Konnerth A. 1991. Intradendritic release of calcium induced by glutamate in cerebellar Purkinje cells. Neuron 7:577?583. Magee JC, Christofi G, Miyakawa H, Christie B, Lasser-Ross N, Johnston D. 1995. Subthreshold synaptic activation of voltage-gated Ca2? channels mediates a localized Ca2? influx into the dendrites of hippocampal pyramidal neurons. J Neurophysiol 74:1335?1342. Malinow R, Otmakhov N, Blum KI, Lisman J. 1994. Visualizing hippocampal synaptic function by optical detection of Ca2? entry through the N-methyl-D-aspartate channel. Proc Natl Acad Sci USA 91:8170?8174. Malouf AT, Robbins CA, Schwartzkroin PA. 1990. Phaclofen inhibition of the slow inhibitory postsynaptic potential in hippocampal slice cultures: a possible role for the GABAB-mediated inhibitory postsynaptic potential. Neuroscience 35:53?61. McBurney RN, Neering IR. 1987. Neuronal calcium homeostasis. Trends Neurosci 10:164?169. Meldolesi J, Pozzan T. 1998. The endoplasmic reticulum Ca2? store: a view from the lumen. Trends Biochem Sci 23:10?14. Michel M, Gna?gi H, Mu?ller M. 1992. Diamonds are a cryosectioner?s best friend. J Microsc 166:43?56. Miller RJ. 1991. The control of neuronal Ca2? homeostasis. Prog Neurobiol 37:255?285. Miyakawa H, Ross WN, Jaffe D, Callaway JC, Lasser-Ross N, Lisman JE, Johnston D. 1992. Synaptically activated increases in Ca2? concentration in hippocampal CA1 pyramidal cells are primarily due to voltage-gated Ca2? channels. Neuron 9:1163?1173. Muller D, Buchs P-A, Stoppini L. 1993. Time course of synaptic development in hippocampal organotypic cultures. Dev Brain Res 71:93?100. Murphy SN, Thayer SA, Miller RJ. 1987. The effects of excitatory amino acids on intracellular calcium in single mouse striatal neurons in vitro. J Neurosci 7:4145?4158. Neher E. 1995. The use of fura?2 for estimating Ca buffers and Ca fluxes. Neuropharmacology 34:1423?42. Ottensmeyer FP. 1986. Elemental mapping by energy filtration: advantages, limitations, and compromises or Electron spectroscopic imaging: parallel energy filtering and microanalysis in the fixedbeam electron microscope. Ann NY Acad Sci 483:339?53. Perkel DJ, Petrozzino JJ, Nicoll RA, Connor JA. 1993. The role of Ca2? entry via synaptically activated NMDA receptors in the induction of long-term potentiation. Neuron 11:817?823. Petrozzino JJ, Pozzo-Miller LD, Connor JA. 1995. Micromolar Ca2? 379 transients in dendritic spines of hippocampal pyramidal neurons in brain slice. Neuron 14:1223?1231. Pivovarova NB, Pozzo-Miller LD, Leapman RD, O?Connell M, Reese TS, Andrews SB. 1997. Rapid calcium accumulation by endoplasmic reticulum and mitochondria following synaptic activation of CA3 pyramidal dendrites in hippocampal slices. Soc Neurosci Abstr 23:1127. Pozzan T, Rizzuto R, Volpe P, Meldolesi J. 1994. Molecular and cellular physiology of intracellular calcium stores. Physiol Rev 74:595?636. Pozzo-Miller LD, Landis DMD. 1993. Cytoplasmic structure in organotypic cultures of rat hippocampus prepared by rapid freezing and freeze substitution fixation. Synapse 13:195?205. Pozzo-Miller LD, Petrozzino JJ, Mahanty NK, Connor JA. 1993. Optical imaging of cytosolic calcium, electrophysiology, and ultrastructure in pyramidal neurons of organotypic slice cultures from rat hippocampus. NeuroImage 1:109?120. Pozzo-Miller LD, Petrozzino JJ, Connor JA. 1995. G protein-coupled receptors mediate a fast excitatory postsynaptic current in CA3 pyramidal neurons in hippocampal slices. J Neurosci 15:8320?8330. Pozzo-Miller LD, Petrozzino JJ, Golarai G, Connor JA. 1996. Ca2? release from intracellular stores induced by afferent stimulation of pyramidal neurons in hippocampal slice. J Neurophysiol 76:554? 562. Pozzo-Miller LD, Pivovarova N, Leapman RD, Buchanan RA, Reese TS, Andrews SB. 1997. Activity-dependent calcium sequestration in dendrites of hippocampal neurons in brain slices. J Neurosci 17:8729?8738. Pozzo-Miller LD, Inoue T, DiEuliis Murphy D. 1999. Estradiol increases spine density and NMDA-dependent Ca2? transients in spines of CA1 pyramidal neurons from hippocampal slices. J Neurophysiol 81:1404?1411. Regehr WG, Connor JA, Tank DW. 1989. Optical imaging of calcium accumulation in hippocampal pyramidal cells during synaptic activation. Nature 341:533?536. Rizzuto R, Brini M, Murgia M, Pozzan T. 1993. Microdomains with high Ca2? close to IP3-sensitive channels that are sensed by neighboring mitochondria. Science 262:744?777. Rizzuto R, Bestianutto C, Brini M, Murgia M, Pozzan T. 1994. Mitochondrial Ca2? homeostasis in intact cells. J Cell Biol 126:1183? 1194. Schwartzkroin PA. 1975. Characteristics of CA1 neurons recorded intracellularly in the hippocampal in vitro slice preparation. Brain Res 85:423?436. Shi S, Sun S, Rews SB, Leapman RD. 1996. Thickness measurement of hydrated and dehydrated cryosections by EELS. Microsc Res Tech 33:241?250. Shuman H, Somlyo AP. 1987. Electron energy loss analysis of neartrace-element concentrations of calcium. Ultramicroscopy 21:23?32. Shuman H, Somlyo AV, Somlyo AP. 1976. Quantitative electron probe microanalysis of biological thin sections: methods and validity. Ultramicroscopy 1:317?339. Somlyo AP, Bond M, Somlyo AV. 1985. Calcium content of mitochondria and endoplasmic reticulum in liver frozen rapidly in vivo. Nature 314:622?625. Somlyo AV, Shuman H, Somlyo AP. 1977. Elemental distribution in striated muscle and the effects of hypertonicity. Electron probe analysis of cryosections. J Cell Biol 74:828?857. Spruston N, Schiller Y, Stuart G, Sakmann B. 1995. Activitydependent action potential invasion and calcium influx into hippocampal CA1 dendrites. Science 268:297?300. Stoppini L, Buchs L-A, Muller D. 1991. A simple method for organotypic cultures of nervous tissue. J Neurosci Methods 37:173?182. Tank DW, Sugimori M, Connor JA, Llina?s RR. 1988. Spatially resolved calcium dynamics of mammalian Purkinje cells in cerebellar slice. Science 242:773?777. Van Harreveld A, Crowell J. 1964. Electron microscopy after rapid freezing on a metal surface and substitution fixation. Anat Rec 149:381?386. Westenbroek DB, Ahlijanian MK, Catterall WA. 1990. Clustering of L-type Ca2? channels at the base of major dendrites in hippocampal pyramidal neurons. Nature 347:281?284. Zimmer J, Ga?hwiler BH. 1984. Cellular and connective organization of slice cultures of the rat hippocampus and fascia dentata. J Comp Neurol 228:432?446.