HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Lee, S.-H. Articles by Neher, E. Search for Related Content PubMed PubMed Citation Articles by Lee, S.-H. Articles by Neher, E.

Differences in Ca²⁺ buffering properties between excitatory and inhibitory hippocampal neurons from the rat

Suk-Ho Lee, Christian Rosenmund, Beat Schwaller * and Erwin Neher

MS 9975 Received 11 August 1999; accepted after revision 6 March 2000.

	ABSTRACT

Top Abstract Introduction Methods Results Discussion References

1. Endogenous calcium binding ratios (_S) in dendrites of cultured hippocampal neurons were estimated according to the single compartment model for transients in intracellular Ca²⁺ concentration ([Ca²⁺]). In addition, the electrophysiological characteristics of neurons were classified by their autaptic currents and intrinsic firing patterns. These data were analysed in order to determine whether a correlation between Ca²⁺ buffers and electrophysiological type exists.

2. Ca²⁺ binding ratios of endogenous buffers were estimated by eliciting [Ca²⁺] transients with short depolarizations, while cells were loaded with fura-2. Two types of estimates could be obtained: one termed _S(), based on analysing time constants () of [Ca²⁺] transients, and another termed _S(dCa), derived from an analysis of initial amplitudes of [Ca²⁺] transients.

3. Values for _S() and _S(dCa) were estimated as 57 ± 10 (mean ± s.d., n = 10) and 60 ± 14 (n = 10), respectively, in excitatory neurons, and 130 ± 50 (n = 11) and 150 ± 70 (n = 11), respectively, in inhibitory neurons. The _S values of excitatory and inhibitory cells were significantly different from each other, regardless of the measurement method (Student's t test, P < 0·01). However, there was no significant difference in _S between the groups classified according to firing patterns.

4. Although _S() values were well matched to those of _S(dCa) in most excitatory cells, the two values did not agree in three out of the fourteen inhibitory cells investigated. In these cells, the first few [Ca²⁺] transients after obtaining the whole cell configuration displayed a double exponential decay, suggesting that buffers with slow binding kinetics, such as parvalbumin, are involved. This hypothesis is further explored in an accompanying paper.

	INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

Spatial and temporal fine tuning of the intracellular Ca²⁺ level ([Ca²⁺]_i) allows Ca²⁺ to serve a multitude of cellular functions as a second messenger. The spatio-temporal extent of the [Ca²⁺]_i increase upon perturbation from its resting levels is largely determined by calcium influx channels, extrusion mechanisms, and calcium buffers. Within a given calcium flux system, calcium buffers have profound effects on the amplitude of [Ca²⁺] transients, the diffusional range, and the relaxation time constant of these signals. Lumped Ca²⁺ intracellular buffer capacity can be quantified by measuring the differential increment of Ca²⁺ bound buffer divided by the free calcium increase (the calcium binding ratio, ) using the single compartment model (Neher & Augustine, 1992; Helmchen et al. 1996). Such measurements have been described for chromaffin cells (Neher & Augustine, 1992), somata of cerebellar Purkinje neurons (Fierro & Llano, 1996), dendrites of pyramidal neurons in cerebral cortex (Helmchen et al. 1996), and motoneurons (Lips & Keller, 1998). Considerable differences in endogenous calcium binding ratios (_S), as estimated in these different types of neurons, prompted us to survey _S in hippocampal neurons and to attempt a correlation between this quantity and their electrophysiological characteristics. Hippocampal neurons cultured in low density offer an excellent model system, in which inhibitory and excitatory cells can be readily identified based on the type of autaptic current (Bekkers & Stevens, 1991), thereby allowing unambiguous comparison of Ca²⁺ dynamics between inhibitory and excitatory cells.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Hippocampal neuron culture

Astrocyte feeder wells were made 4-5 days before plating neurons by plating astrocytes on 18 mm coverslips coated with a 1:4 mixture of rat tail collagen and poly-D-lysine. Newborn Sprague-Dawley rats were decapitated according to the rules of the state animal welfare committee. The brains were cleaned of meninges and vascular tissue, and hippocampi were dissected in cold physiological salt solution. The tissue was enzymatically dissociated with 2 units ml^-1 papain (Worthington, Lakewood, NJ, USA) in Dulbecco's minimal essential medium (DMEM, Life Technologies, Karlsruhe, Germany) for 60 min at 37°C, transferred to DMEM containing 5 % fetal calf serum (FCS, Biochrom, Berlin, Germany), 50 U ml^-1 penicillin, and 50 µg ml^-1 streptomycin, and gently triturated by passage through a Pasteur pipette. The suspension was plated into astrocyte feeder wells to a final cell density of 400-800 cells cm^-2. Within 2 days in culture, cells were treated with 8 µM 5-fluoro-2'-deoxyuridine and 20 µM uridine in order to halt glial proliferation.

Solutions

The standard internal dialysis solution for hippocampal cultured neurons contained (mM): 170 potassium gluconate, 10 Hepes, 5 Mg-ATP, 5 Na₂-creatinin-phosphate, 5 KCl, with 100-200 µM K₅-fura-2. The standard external solution for bathing neurons during experiments contained (mM): 168 NaCl, 2·8 KCl, 10 Na-Hepes, 2 MgCl₂, 4 CaCl₂, 11 glucose. To block synaptic currents, 10 µM 6-nitro-7-sulphamoylbenzo[f]quinoxaline-2,3-dione (NBQX) and 20 µM bicuculline were added to the external solution. The pH of internal and external solutions was adjusted to 7·2 and 7·3, respectively, with the bases of the main cation in the given solution. The final osmolarity of external solutions was adjusted to that of the medium of the given culture and was typically within 360 ± 5 mosmol l^-1 on day 14. All experiments were performed at room temperature (22-25°C). All chemicals were obtained from Sigma, except NBQX and bicuculline (Tocris Cookson, Bristol, UK), fura-2 (Molecular Probes, Eugene, OR, USA) and chemicals for cell culture (Life Technologies).

Imaging of [Ca²⁺] transients in neuronal dendrites

For fura-2 fluorescence (F) excitation, a polychromatic light source (xenon-lamp based, Polychrome-II, TILL Photonics, Martinsried, Germany), providing a band (±10 nm) of monochromatic light, was coupled to the epi-illumination port of an inverted microscope (Axiovert 135 TV, Zeiss, Jena, Germany) via a quartz light-guide and a UV condenser. Imaging was performed with a × 40 water immersion objective lens (NA = 1·2, C-Apochromat, Zeiss) and a water-cooled slow-scan CCD camera (TILL Photonics Imago CCD camera). The monochromator and CCD were controlled by a PC running VisION Software (TILL Photonics). A dichroic mirror (DC400LP, Omega Optical, Brattleboro, VT, USA) was used for reflecting fura-2 excitation light, and a long pass emission filter (Q515LP, Omega Optical) was used to separate the emission light and residual scattered excitation light.

Calibration parameters were determined using an in vivo calibration (Neher, 1989). Briefly, the fura-2 fluorescence ratios at minimal and maximal Ca²⁺ concentrations, R_min and R_max (see below), were determined by loading neurons with the standard internal solutions plus 10 mM K₅-BAPTA and 10 mM CaCl₂, respectively. The effective dissociation constant (K_eff) was determined by loading cells with an intracellular solution containing 3·33 mM K₅-BAPTA and 6·66 mM Ca-BAPTA. K_eff was calculated from the equation:

Keff = [Ca2+](Rmax - Rint)/(Rint - Rmin),

(1)

where [Ca²⁺] was entered as 444 nM (assuming a dissociation constant (K_d) for BAPTA of 222 nM at pH = 7·2), and R_int is the fluorescence ratio measured under the same condition. The K_d of fura-2 was calculated from:

Kd = Keff ( + Rmin)/( + Rmax),

(2)

where is the isocoefficient (Zhou & Neher, 1993). The estimated values for R_min, R_max and K_eff (in µm) were typically 0·5, 5·23 and 1·70, respectively.

For high time resolution and minimization of the photobleaching effect, images taken with single wavelength excitation at 380 nm (F₃₈₀), at a frequency of 10-20 Hz, were preceded and followed by images taken with dual excitation at wavelengths of 350 and 380 nm. During off-line analysis, a region of interest (ROI) was set on the proximal dendrite for determining fura-2 fluorescence (F). Adjacent to this ROI, a second ROI was chosen in an area with no neuronal structures in order to determine background fluorescence (F_b). The fluorescence intensities from the two ROIs were then averaged to get F and F_b. The value of F - F_b was regarded as the relevant fura-2 fluorescence of the ROI. Subsequently, isosbestic fluorescence (F_iso) was calculated from images of the double wavelength excitation period using the equation:

Fiso = F350 + F380.

(3)

The values thus obtained were linearly interpolated between points just before and just after the period of single wavelength excitation. The ratios R = F₃₅₀/F₃₈₀ and R' = F_iso/F₃₈₀ were then converted to [Ca²⁺] using the equations:

[Ca2+] = Keff(R - Rmin)/(Rmax - R)

(4)

and

[Ca2+] = Keff(R' - (Rmin + ))/((Rmax + ) - R'),

(5)

respectively.

Estimation of dendritic calcium binding ratios in single neurons

Within 20 s of establishing the whole cell configuration we started to apply short depolarizing pulses (to +10 mV for 3 ms) every 20-60 s in order to evoke [Ca²⁺] transients in the dendritic tree. F₃₈₀ images at 10-20 Hz were taken. Furthermore, these images were preceded and followed by double wavelength image pairs, so that the Ca²⁺ concentration could be calculated as described in the previous section. Because calcium transients were largely abolished by the addition of 1 µM TTX to the bath solution (data not shown), these dendritic transients are most probably evoked by back-propagating action potentials, which may be generated by the mechanism of voltage escape.

During off-line analysis, F_iso was calculated, and the time course of F_iso was regarded as the loading curve of fura-2. The concentration of fura-2 in dendrites was estimated assuming that the F_iso at the plateau of the loading curve represents full loading of fura-2. In cases where the F_iso measured in a ROI set at a large distance from the soma did not plateau, an exponential fit of the loading curve was extrapolated to infinity and regarded as the value of F_iso in the fully loaded state.

Electrophysiological characterization of the cells

At the end of each experiment, the electrophysiological characteristics of the cell under study were determined according to autaptic current and firing pattern in response to long (1 s) depolarizing current injections. To yield a high proportion of cells that had autaptic currents without restricted dendritic growth, cells were cultured at low density (400-800 cells cm^-2 at the time of seeding) on top of an astrocyte feeder layer. Furthermore, relatively isolated cells were selected for the investigation. Within these constraints, roughly half of the cells investigated displayed autaptic currents when examined with 3 ms depolarizing pulses (from -70 to 0 mV) while perfused with bath solutions lacking blockers (Fig. 1). The type of autaptic current could be differentiated based on its relaxation time course, direction, and sensitivity to specific blockers (NBQX or bicuculline). Consistent with a previous report (Bekkers & Stevens, 1991), excitatory autaptic currents (probably glutamatergic) relaxed with shorter time courses than inhibitory ones (probably GABAergic). Time constants of the latter were longer than 30 ms. Since low concentrations of chloride in the internal dialysis solutions (calculated chloride equilibrium potential E_Cl -92 mV) were used, the two kinds of current could be distinguished by their polarity. Under these experimental conditions, excitatory autaptic currents were inwardly directed and sometimes accompanied by Na⁺ spikes, while inhibitory autaptic currents were outwardly directed at -70 mV (Fig. 1B).

		View larger version [in this window] [in a new window]
Figure 1. Representative example of excitatory (A) and inhibitory (B) autaptic currents Glutamatergic and GABAergic currents were blocked by superfusion of the cells with 10 µM NBQX and 20 µM bicuculline, respectively (dotted traces, which are closer to the x-axis in both cases).

In the presence of NBQX and bicuculline, the firing pattern in response to 1 s of depolarizing current injections was characterized, since close correlation between fast-spiking neurons and GABAergic non-pyramidal neurons containing parvalbumin in hippocampus has been reported previously (Kawaguchi et al. 1987). The intrinsic firing pattern was classified as fast spiking (FS), regular spiking (RS), and irregular or burst spiking (IS), according to their frequency accommodation property and maximal firing frequency. Because FS cells could not be found, i.e. neurons with no frequency accommodation (Connors & Gutnick, 1990), the firing patterns were instead classified based on the criteria summarized in Table 1.

Table 1. Types and criteria for classifying firing pattern

Intrinsic firing pattern	Maximal frequency	Accommodation factor (a)
Fast spiking	> 35 Hz	< 30 %
Regular spiking	< 30 Hz	30 %
Irregular spiking	n.d.	n.d.

The maximal frequency (f_max) and accommodation factor (a) of a given cell were estimated from a plot of instantaneous discharge frequency (calculated as the inverse of the interval between a given spike and the next one) against the time when the first spike occurred. The maximal frequency was determined as the average instantaneous frequency within 0·2-0·8 s after onset of stimulation, using 500-700 pA current injections. Recording at room temperature might be responsible for the lower maximal frequency compared to other reports (Kawaguchi et al. 1987; Connors & Gutnick, 1990), which were performed at higher temperatures. The accommodation factor was defined as:

a = (f(0·05 - 0·15) - f(0·9-1))/f(0·05 - 0·15),

(6)

where f_{(t0 - t1)} represents the mean instantaneous discharge frequency averaged within the interval (in seconds) between onset of stimulation (t₀) and the occurrence of the first spike (t₁). f_max and a could not be evaluated in IS cells, since these cells showed bursting or otherwise unpredictable firing patterns.

Theory of the single compartment and linear approximation model

Ca²⁺ binding ratios of endogenous buffers (_S) were estimated according to the single compartment and linear approximation model (Neher & Augustine, 1992; Helmchen et al. 1996; Neher, 1998). Within this framework, calcium transients (d[Ca²⁺](t)) following short pulses of Ca²⁺ influx can be described by the following equations:

d[Ca2+](t) = A exp(-t/),

(7)

A = d[Ca]T/(1 + S + B),

(8)

= (1 + S + B)/,

(9)

where A is the initial amplitude, d[Ca]_T is the total intracellular Ca²⁺ increase evoked by the influx, is the calcium extrusion rate, and _B is the Ca²⁺ binding ratio of the Ca²⁺ indicator dye (fura-2). _B is defined as:

B [BCa]/ [Ca2+]i = [B]TKd,B/([Ca2+]i + Kd,B)2,

(10)

where [BCa] is the concentration of the fura-2-Ca complex, [B]_T is the total concentration of fura-2 and K_d,B is its dissociation constant. In practice, however, _B was replaced by the incremental calcium binding ratio (_B'), since _B is not completely linear within the dynamic range of [Ca²⁺]_i in this study. _X', the incremental Ca²⁺ binding ratio of a buffer X, is defined as:

X' = [X]TKd,X/{\123}([Ca2+]i,rest + Kd,X)([Ca2+]i,peak + Kd,X)},

(11)

where [Ca²⁺]_i,rest and [Ca²⁺]_i,peak represent [Ca²⁺]_i values before and at the peak perturbation, [X]_T is the total concentration of Ca²⁺ buffer X, and K_d,X is the Ca²⁺ dissociation constant of X (Neher & Augustine, 1992).

According to eqns (8) and (9), plotting A and values measured at different levels of _B (during wash-in of fura-2) provides two estimations of _S: the first is obtained from the x-intercept of the straight line fit to the plot of versus _B, and the second from a fit of 1/A versus _B. These two estimates will be referred to as _S() and _S(dCa), respectively. Each [Ca²⁺] transient was fitted with an exponential curve, K₀ + K₁ exp(-t/), where K₀ was constrained such that the fit reaches the resting [Ca²⁺] level as determined from 5-10 points preceding the stimulus.

This model predicts that the product A should be independent of added buffer, according to the equation:

([Ca2+]i - [Ca2+]rest) dt = A = d[Ca]T/.

(12)

To quantify the validity of this model, the product A for every cell was evaluated. Specifically, all the A values measured from a ROI were pooled, and their standard deviation regarded as a measure of the uniformity of A.

Immunohistochemistry for parvalbumin (PV)

Hippocampal neurons grown on glass coverslips were fixed with 4 % paraformaldehyde in Tris-buffed saline (TBS, pH 7·3). After permeabilizing the fixed cells with TBS containing 0·1 % Triton X-100, endogenous peroxidase was blocked by incubation with methanol-H₂O₂ 50:1 (v/v) for 30 min. After washing with TBS, cells were incubated in a moist chamber for 3 days at 4°C with primary anti-PV antibody (1:2000, PV-28, SWant, Bellinzona, Switzerland) in TBS containing 10 % bovine serum in a moist chamber. The second antibody (1:200, biotinylated goat anti-rabbit IgG) was added at room temperature for 2 h. This was followed by incubation with the avidin-biotin complex (1:200, VECTOR Laboratories, Burlingame, CA, USA). The antibody complex was visualized by incubation with the substrate 3,3'-diaminobenzidine (DAB)-HCl-hydrogen peroxide.

Data analysis and numerical simulation

Data were analysed on a PC with IgorPro (version 3.12, WaveMetrics, Lake Oswego, OR, USA) and compared statistically using Student's t test. The statistical data are presented as means ± S.D., with n indicating the number of cells and P < 0·01 indicating a significant difference.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Spatial homogeneity of [Ca²⁺] along dendrites

The single compartment model requires that diffusional equilibration of [Ca²⁺] across the compartment under consideration is rapid. Cylindrical regions of dendrites, which are typically 3 µm in diameter and 10 µm long, were examined in order to verify the spatial homogeneity of [Ca²⁺] along this direction. Specifically, line profiles of F/F along the dendrites (x-axis) were calculated from F₃₈₀ images, sampled every 20 or 100 ms during [Ca²⁺] transients. From these line profiles, images were constructed (Fig. 2Ab and Bb) by assigning consecutive line profiles to consecutive lines in the image (similar to line-scan displays in laser scan microscopy).

		View larger version [in this window] [in a new window]
Figure 2. Composed line profile images for examining spatial homogeneity

Aa, hippocampal cultured neuron loaded with 150 µM fura-2 from a somatic patch pipette. Scale bar = 20 µm. Ab, F/F image reconstructed from a time series (y-axis) of line profiles along the dendrite (x-axis) as described below. The horizontal axis represents distance along the line profile, while the vertical axis shows time. For the sixth image, a 3 ms depolarization was applied through the pipette to the soma (marked with ), and images were sampled every 20 ms at 380 nm exitation. During off-line analysis, two lines of interest (LOI) were drawn: one was along the dendrite (marked with in A) for F_dend (fluorescence of dendrite), and the other was 25 pixels away from the cell margin for F_b (background fluorescence). The difference F_dend(x) - F_b(x), obtained by point-to-point subtraction, was regarded as real fluorescence originating from fura-2 (F_fura2). The line profile of fura-2 fluorescence at the resting state (F_rest(x)) was calculated by time-averaging the first 5 line profiles before depolarization. Finally, F/F as a percentage value was calculated according to 100 × [{\123}F_fura2(x) - F_rest(x)}/F_rest(x)], and the time series of F/F line profiles were combined to construct the image illustrated. The marks on the lower LOI in the F₃₈₀ image (Aa) correspond to the triangles () on the x-axis of Ab. Ac, line profile along a LOI on a dendrite just after the depolarization. The values at each pixel were obtained by averaging F/F values at corresponding pixels in 5 temporally consecutive line profiles. Ba, fluorescence image of a hippocampal cultured neuron, which has an excitatory autaptic current (1·0 nA, data not shown). Scale bar = 20 µm. Bb and c were obtained as described in Ab and c. Again, for the sixth image, a 3 ms depolarization was applied through the pipette to the soma (marked with ) and images were sampled every 100 ms at 380 nm excitation.

When synaptic blockers (NBQX and bicuculline) were not included in the bath solution, hot spots were observed in the dendrites of the excitatory cells (Fig. 2B; n = 2 out of 4). A hot spot was defined as a local area on a dendrite, for which the value F/F was higher by more than 2 S.D. than mean values in other dendritic regions. After including synaptic blockers in the bath solutions, no dendritic hot spots were observed in composite images such as those of Fig. 2Ab (n = 8). Because synaptic inputs might be responsible for such dendritic hot spots, subsequent experiments were performed in the presence of the synaptic blockers. Even when hot spots persisted, the _S values estimated in the ROIs which contained these hot spots were not significantly different from those of other ROIs, which did not contain hot spots. Therefore, such inhomogeneities in dendritic Ca²⁺ profiles do not significantly affect the estimates of _S (data not shown). Similarly, we found no significant dependence of on the distance of a ROI from the soma, when _S values were estimated from two or three ROIs in a single cell (Fig 3 and Fig 5C). In contrast, the _S values estimated at branching points were sometimes (although not always) higher than those in dendritic shafts (note asterisks in Fig. 5C). Therefore, ROIs were always placed at a specified distance from branching points for routine evaluation of _S values.

		View larger version [in this window] [in a new window]
Figure 3. Representative examples for measurements of S in an excitatory neuron Aa, fura-2 fluorescence image of a hippocampal cultured neuron. The image was taken 14 min after establishment of the whole cell configuration with excitation at 380 nm (20 ms exposure). This cell had been cultured for 14 days. Its electrophysiological properties were classified as those of a regularly spiking excitatory cell (fmax = 30 Hz, a = 46 %). Two ROIs (marked ROI1 and ROI2) were chosen for the measurement of dendritic fluorescence at 26 and 43 µm (ROI1 and ROI2, respectively) from the margin of the soma. Two adjacent ROIs were selected for background fluorescence (Fb). Scale bar = 20 µm. Ab, time course of isosbestic fluorescence (Fiso) in ROI1 () and ROI2 () during whole cell patch recording with a somatic pipette containing 150 µM fura-2. The abscissa represents the time elapsed after break-in (s). The left and right ordinates represent fluorescence from ROI1 and ROI2, respectively. Both ordinates have the same units (adu, which refers to analog-to-digital units of the CCD camera). The first point (at t = 0) was taken in the cell-attached configuration and was regarded as the autofluorescence from the dendrite in the ROI. Exponential curves (continuous lines) were fitted to Fiso of ROI1 and ROI2 with time constants of 149·5 and 188·7 s, respectively. Ac, plot of the decay time constants () of [Ca2+] transients (shown in B) against the incremental calcium binding ratio of fura-2 (B') evaluated at the same time point. The x-axis intercepts of the regression lines fitted to values of ROI1 (, continuous line) and ROI2 (, dotted line) were 51·8 and 56·5, respectively, and provide values designated as S() for ROI1 and ROI2, respectively. Ad, plot of the inverse of the [Ca2+] transient amplitudes as a function of B'. Based on eqn (8), S(dCa) values for ROI1 (, continuous regression line) and for ROI2 (, dotted line) were estimated to be 54·0 and 69·7, respectively. The 3 points with the highest B' from each ROI were excluded from the fit. Ae, plot of the product A versus B', showing the lack of dependence of A on B'. B, dendritic [Ca2+] transients measured in the two ROIs. The 8 exemplar [Ca2+] transients shown here were obtained by 3 ms depolarizations evoked during the period marked in the loading curve (

Ab, between 26 and 242 s after break-in). The numbers between the two columns denote the whole cell recording time in seconds (the time elapsed after the whole cell configuration was established).

Endogenous calcium binding ratios in dendrites of excitatory and inhibitory neurons

Estimates for _S() and _S(dCa) were obtained from serial images of [Ca²⁺] transients, which were elicited repetitively while cells were being loaded with fura-2. The procedure is described in the Methods section and exemplar analyses for excitatory and inhibitory cells are detailed in Fig 3 and Fig 4, respectively.

		View larger version [in this window] [in a new window]
Figure 4. Representative example for S estimation in an inhibitory neuron A, fluorescence image of an inhibitory hippocampal cultured neuron (culture age was 17 days). [Ca2+] transients were obtained from the average fluorescence of a ROI, set on the dendrite. Another ROI was selected for background fluorescence (Fb). Scale bar = 20 µm. B, time course of isosbestic fluorescence (Fiso) of the ROI during whole cell patch recording with a somatic pipette containing 150 µM fura-2. Both axes have the same meaning as those of the loading curves shown in Fig. 3. Fura-2 diffused into the ROI with a time constant of 137·5 s, as calculated from the loading curve. C, dendritic [Ca2+] transients measured in the ROI. The 9 exemplar [Ca2+] transients were evoked by 3 ms depolarizations between 24 and 249 s after break-in. The numbers at the right side of each [Ca2+] transient denote the whole cell recording time in seconds. D, estimation of S in the ROI. Da, plot of time constants () of the decay in [Ca2+] transients (shown in C) against the incremental Ca2+ binding ratio of fura-2 (B') evaluated at the same time point. The S(), measured as the x-axis intercept of the regression line fitted to values (continuous line) was 130. Db, plot of the inverse of [Ca2+] transient amplitudes as a function of B' based on eqn (8), where S(dCa) was estimated to be 138. The last point was excluded from the linear fits to and 1/dCa. Dc, plot of the product A versus B'.

For a comparison between excitatory and inhibitory cells, only cells whose _S values satisfied the following conditions were chosen for analysis: (1) ROIs could be set on the proximal dendritic shaft, located between 15 and 40 µm from the edge of the soma, at some distance from any visible branching point; (2) the difference between _S() and _S(dCa) divided by the average of the two values was less than 0·3; and (3) the standard deviations of the A values from the ROIs included for analysis were less than 5 × 10^-3 µm s. Cases satisfying all these requirements were pooled according to their electrophysiological type and then statistically compared (see Fig. 5A). The _S values from individual groups, classified according to their firing patterns, are listed in Table 2 and Table 3.

		View larger version [in this window] [in a new window]
Figure 5. Summary of endogenous Ca2+ binding ratios estimated in excitatory and inhibitory hippocampal cultured neurons A, graphic representation of mean S() and S(dCa) values measured in excitatory and inhibitory cells. Error bars represent S.D. Statistical comparison of values from excitatory and inhibitory neurons revealed statistically significant differences (P < 0·01) for both S() (*) and S(dCa) (**) values. B, plot of S versus culture age of neurons. No statistically significant dependence on age is resolved within the range of ages investigated. C, plot of S(dCa) versus the distance of the ROI from the margin of the soma. The S(dCa) values were measured in 27 different ROIs from 10 excitatory cells () and 24 different ROIs from 11 inhibitory cells (). All S values marked with were measured in the same cell. The ROI of the value at 30 µm was set on a branching point while the others were located on the dendritic shaft.

Table 2. Endogenous Ca²⁺ binding ratios (_S) and calcium extrusion rates () in cells classified according to synapse type

Type of autaptic current	n	S()	S(dCa)	(s-1)
Excitatory	10	57 ± 10	60 ± 14	340 ± 94
Inhibitory	11	130 ± 50	152 ± 68	305 ± 77

Spiking pattern	n	fmax (Hz)	a	Number of neurons			S(dCa)
				Exc.	Inh.	n.d.
Fast spiking	4	45 ± 14	0·17 ± 0·06	0	3	1	134 ± 45
Regular spiking	11	18·5 ± 15·3	0·57 ± 0·30	4	6	1	132 ± 83
Irregular spiking	4	n.d.	n.d.	3	1	0	64 ± 30

The _S values of excitatory and inhibitory cells were significantly different from each other, regardless of the type of measurement considered (Student's t test, P < 0·01). The _S values of the inhibitory cells were more dispersed than those of excitatory cells, consistent with previous observations that interneurons in the hippocampus are so diverse that they cannot be classified into a few distinctive groups (Parra et al. 1998).

The average values of calcium extrusion rates (), estimated according to eqn (9) in dendrites of hippocampal cultured neurons, are also listed in Table 2. Values in excitatory and inhibitory cells were not statistically different from each other; however, they were higher than the values measured in somata of motoneurons reported elsewhere (60 s^-1, Lips & Keller, 1998; 140 s^-1, Palecek et al. 1999). A higher surface-to-volume ratio might be responsible for these differences (Eilers et al. 1995). The mean calcium decay time constants, calculated from and _S, in the absence of exogenous Ca²⁺ buffers were 176 and 429 ms in excitatory and inhibitory cells, respectively.

The _S(dCa) values showed no significant statistical difference when they were grouped according to firing pattern (Student's t test, P > 0·01). Only _S values between FS and IS cells showed a marginally significant difference (Student's t test, 0·01 < P < 0·05), suggesting that the amount of endogenous buffer expressed in these cultured hippocampal neurons does not correlate with their intrinsic firing patterns at room temperature.

Furthermore, _S showed no dependence on culture age (between 13 and 18 days) or distance of the ROI from the somata (Fig. 5B and C).

No evidence for diffusible buffer in excitatory hippocampal neurons

The value of _S reported above for excitatory neurons is quite similar to that found in bovine chromaffin cells, where it has been shown that most of the underlying buffer is fixed to cellular structures and therefore does not wash out during prolonged whole cell recordings (Zhou & Neher, 1993). In order to verify that excitatory neurons also have an immobile buffer, the _S in a given cell was compared during wash-in and wash-out of indicator dye by patching a cell twice successively with two different concentrations of fura-2 (Fig. 6). First, the cell was loaded with fura-2 (150 µM) for 5 min, during which 12 action potentials were evoked in order to monitor [Ca²⁺] transients (Fig. 6C). Following retraction of the first patch pipette, the cell was patched with a second pipette containing a lower concentration (17·5 µM) of fura-2, and the stimulation protocol was repeated (Fig. 6D). The F_iso at the start of the second patch was slightly lower than that at the end of the first patch, indicating that fura-2 had diffused from the proximal dendrite into the distal dendritic tree between the two recording episodes. The fura-2 concentrations during the loading and unloading curves were calculated assuming that the F_iso at the final plateau of the second loading curve corresponds to the F_iso of 17·5 µM fura-2. Values for d[Ca]_t = 0 (amplitudes at the peak of [Ca²⁺] transients) obtained from loading and unloading of fura-2 were almost identical (see Fig. 6E), indicating that no substantial amount of diffusible buffer was washed out during loading and unloading with fura-2. However, time constants () from the late period of the second whole cell recording were slightly longer than those from the first patch. Together with the slight increase of corresponding A values, these results suggest that some deterioration of the Ca²⁺ extrusion mechanism occurred at late times during the second whole cell recording.

		View larger version [in this window] [in a new window]
Figure 6. Loading and unloading with fura-2 A, fura-2 fluorescence image of an excitatory cell under study taken at the end of the first whole cell patch recording (excitation 380 nm, 20 ms exposure). The ROI on the dendrite was set 15 µm distal to the root of the dendrite. Scale bar = 20 µm. B, fura-2 loading curve during the first () and the second () somatic whole cell episode using pipettes containing 150 and 17·5 µM fura-2, respectively. C and D, [Ca2+] transients evoked during the first (wash-in phase) and the second (wash-out phase) whole cell recording. All transients elicited during the wash-in phase, and 13 transients during the wash-out phase are shown. E, analysis of the [Ca2+] transients. Two sets of data obtained from wash-in () and wash-out phases () were superimposed as in Fig. 4D. Ea, S() values were estimated as 53·26 and 67·7 in wash-in and wash-out phases, repectively. The last 6 points in the plot were excluded from the fit. Eb, S(dCa) values were estimated as 55·75 and 49·3 in wash-in and in wash-out phases, respectively. Ec, plot of A versus B'.

Anomalous cases

Four out of 25 cells investigated did not have constant A values throughout the experiment. Electrophysiologically, all of them except one, which had no autaptic current, had inhibitory autaptic current. The firing patterns in response to long depolarizing currents were fast spiking in two cells and regular spiking in the other two. A increased characteristically during the first few minutes of the experiment and then became stable. As a consequence of the increasing A, the _S() was estimated to be lower than _S(dCa). The mean _S(dCa) was estimated at 166·7 ± 40·9, significantly higher than the mean value for _S(), 62·6 ± 27·3.

In the fast-spiking (f_max = 38, a = 0·17) cell shown in Fig. 7 the earliest [Ca²⁺] transients could be fitted better with double exponentials, although single exponentials were sufficient to fit the transients during the late period of the experiment. When estimating the area of the transient (the equivalent of A) using the double exponential fit, the values of the early period matched quite well with the A values of the late, stable period (Fig. 7E).

		View larger version [in this window] [in a new window]
Figure 7. An anomalous case A, fura-2 fluorescence image of the anomalous cell (excitation 380 nm, exposure 20 ms). A ROI was set at a distance of about 33 µm from the margin of the soma. The cell had been cultured for 12 days and showed inhibitory autaptic currents. The fluorescence originating from the patch pipette is also shown as a triangular shadow on the far right-hand side. Scale bar = 20 µm. B, the time course of Fiso from the ROI during a somatic whole cell recording with 150 µM fura-2. The exponential curve fit (continuous line) has a time constant of 120 s. C, a representative example of a double exponential fit to a [Ca2+] transient. The [Ca2+] transient was evoked at 35 s after break-in. Its decay phase was fitted both with single (dotted curve) and double (continuous curve) exponential fits according to: 10 nM + (108·6 nM × exp(-2·21t)) for the former, 10 nM + (25·4 nM × exp(-0·47t)) + (95·0 nM × exp(-3·82t)) for the latter, where t is given in seconds. The time integrals of single and double exponential curves were 0·049 and 0·079 µM s, respectively. The resting [Ca2+] level at the next transient, which was evoked 17 s later, is marked with a triangle () at the end of the fitted curve. In the inset, the same [Ca2+] trace with exponential fits is depicted on an expanded time scale. D, the first 10 [Ca2+] transients evoked between 18 and 240 s after break-in. The numbers in front of Ca2+ traces represent the whole cell recording times (s) at which records were taken. The second [Ca2+] transient is illustrated on a compressed time scale in C. Single (dotted curve) and double (continuous curve) exponential fits were superimposed on the [Ca2+] transients. The following equations were used for the double exponential fitting to the first, third and fourth [Ca2+] transients: 9 nM + (34·6 nM × exp(-0·61t)) + (82·6 nM × exp(-4·26t)), 10 nM + (28·4 nM × exp(-0·455t)) + (73·1 nM × exp(-3·14t)) and 12 nM + (13·6 nM × exp(-0·316t)) + (60·0 nM × exp(-1·68t)), respectively. E, plot of time integrals of single () and double () exponential fits. The first 4 time integrals derived from double exponential fits, combined with the 6 A values at later times, show less dependence on B' than the time integrals of single exponential fits only (note slopes of lines fitted to the former and the latter; they are 0·022 and 0·17 nM s, respectively). F, plot of 1/dCa as a function of B'. The x-axis intercept of the line fitted to the data points revealed a S(dCa) of 171·3. G, plot of decay time constant as a function of B'. The x-axis intercept of the line (continuous line) fitted to all values revealed a S() of 29·7. H, an action potential evoked by a depolarizing current injection (3 ms, 500 pA) via a somatic patch pipette in current clamp mode.

Besides the possibility that the double exponential fit is coincidental, and therefore inaccurate, we considered two possible reasons for an increase in A: (1) a decline in Ca²⁺-ATPase activity or (2) an increase in Ca²⁺ influx during the whole cell patch experiment. Neither reason explains why such a mismatch preferentially occurs in inhibitory neurons. Moreover, the similarity of time integrals of the double exponential curves to the late stable value of A indicates that the ratio d[Ca]_T/ remained constant throughout the experiment as in other cells, since the single compartment model specifically requires that the time integral of the [Ca²⁺] transient equals d[Ca]_T/, regardless of the properties of buffers inside the compartment (Neher, 1998). This implies that a double exponential fit, which estimates the time integral more accurately than the product A will provide a better description for the first few [Ca²⁺] transients.

The mismatch between _S(dCa) and _S(), despite the uniform time integral of [Ca²⁺] transients, indicates that at least one of the assumptions underlying the linear approximation and the single compartment model is violated in the anomalous cells. Assuming that the diffusion coefficient of Ca²⁺ (D_Ca) is 50 µm² s^-1, the characteristic time for diffusion (r²/D_Ca) across the ROI in Fig. 7A (radius r = 1·5 µm) would be about 50 ms, much less than the time constant of [Ca²⁺] decay (400 ms). Therefore, the geometric aspects of the single compartment model are likely to be valid in this cell. If so, the double exponential decay indicates that there must be a buffer, which does not satisfy the requirements for the linear approximation model. Two possibilities for such buffers are saturable buffers and slow buffers. The former have a high affinity for Ca²⁺ and are present in low concentrations in the dendrites; thus, they become readily saturated within the dynamic range of [Ca²⁺] transients evoked by single action potentials. Therefore, the Ca²⁺ binding ratio of saturable buffers cannot be considered to be constant. The latter have binding kinetics so slow that Ca²⁺ cannot fully equilibrate with the buffers during [Ca²⁺] transients. Parvalbumin (PV) is the most plausible candidate for a slow buffer in hippocampal inhibitory neurons, since (1) it has been found exclusively in inhibitory neurons and not in excitatory cells (Celio, 1986); (2) it is known to have slow binding kinetics for Ca²⁺ in the presence of millimolar Mg²⁺ (Hou et al. 1992); and (3) it is readily saturated due to its high affinity for Ca²⁺ and relatively low concentration in hippocampal neurons (Plogmann & Celio, 1993). Moreover, the double exponential nature of dendritic [Ca²⁺] transients was most prominent in fast-spiking inhibitory neurons, whose action potentials are characterized by large after-hyperpolarizations (Fig. 7H). These large after-hyperpolarizations are consistent with the already-known electrophysiological properties of PV-containing neurons (Kawaguchi et al. 1987). Unfortunately, the double exponential pattern of Ca²⁺ decay was not consistently observed in all the anomalous cases, which could be attributed to the insufficient resolution of our Ca²⁺ measuring system, or to variable concentrations of PV in the neurons. Moreover, slowly decaying phases are easily masked by slow overall fluctuations of [Ca²⁺]_rest during whole cell patch experiments.

However, Pfyffer et al. (1984) reported that almost no PV-positive cells were present in brain cell cultures derived from newborn mice. We, therefore, confirmed the presence of PV in our cultured hippocampal neurons by immunohistochemistry (Fig. 8). The morphology of the PV-containing cells varied but consisted mainly of multipolar cells resembling pyramidal-like basket cells or bipolar or bitufted cells as described previously in vivo (Nitsch et al. 1990). The multipolar cells were characterized by relatively large somata and extensive dendritic arborizations, while the bipolar (or bitufted) ones had small elliptical cell bodies with few, relatively fine processes.

		View larger version [in this window] [in a new window]
Figure 8. Micrograph of PV-immunoreactive neurons from rat hippocampal culture Pyramidal-like basket cells (arrows) with relatively large somata and many processes were observed in the culture dishes. A second population of cells with small cell bodies and few fine processes (arrowheads) were reminiscent of the second group of PV-immunoreactive cells found in the CA1 region in vivo (Nitsch et al. 1990). Scale bar, 100 µm.m

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

In the present study, fluorometric Ca²⁺ measurements were combined with characterization of autaptic currents of rat cultured hippocampal neurons to examine possible correlations between Ca²⁺ buffering properties and the electrophysiological (neurotransmitter) type. The results demonstrate that there is a statistically significant difference in the Ca²⁺ binding ratios of fast endogenous buffers (_S) in neuronal dendrites between GABAergic and glutamatergic neurons. Furthermore, the higher values for _S in inhibitory neurons are consistent with previous reports concerning the distribution of Ca²⁺ binding proteins (CaBPs) in the hippocampus, since most pyramidal cells of the cerebral cortex and hippocampus do not contain any of the known members of EF-hand family CaBPs with the exception of calmodulin (Baimbridge et al. 1992).

What is responsible for the higher Ca²⁺ binding ratio in inhibitory neurons? Comparison with relevant immunohistochemical studies on hippocampus

The estimates of _S in this study cannot be directly compared with other immunohistochemistry (IHC) reports and the properties of cultured neurons are not necessarily similar to those of neurons in vivo because of profound differences in the cellular environment which influences developmental processes. Another significant drawback of experiments in cell culture is the lack of knowledge on the exact locational origin within the hippocampus. Such knowledge is essential for the comparison of data on Ca²⁺ binding ratios with IHC reports, most of which deal with the correlation between CaBP immunoreactivity of hippocampal neurons and their anatomical locations within the hippocampus. Furthermore, the expression of CaBPs is known to be dependent on the developmental stage of the nueron (Jiang & Swann, 1997). Considering that most IHC studies were performed on adult rats, while this study was performed with 2-week-old cells, developmental age becomes another important factor limiting comparison of these results to past findings. However, hippocampal cortical neurons are composed of principal (or pyramidal and granule) and non-pyramidal cells, and thus the histological classification can still provide useful information about their neurotransmitter type, since principal cells are excitatory and non-pyramidal cells are largely GABAergic. For this comparison, only PV, calbindin-D_28K (CB) and calretinin (CR) will be considered since these three CaBPs are known to be the major calcium buffer proteins in hippocampal neurons (Baimbridge et al. 1992).

The major CaBP relevant to excitatory cells (or principal cells) in hippocampus seems to be CB, since principal cells positive for CaBPs other than CB are found only in the hilar region of the dentate gyrus (mossy cells; Fujise et al. 1998). Immunoreactivity of principal cells for CB is not homogeneous: dentate granule cells and some of the CA1 pyramidal cells are immunoreactive (IR) for CB, while CA3 and CA4 pyramidal cells have no immunoreactivity for any known CaBPs (Baimbridge & Miller, 1982). In CA1 pyramidal cells, values for _S between 168 and 187 were found (Helmchen et al. 1996). These values are considerably higher than those of the excitatory neurons (57-60) of this study. Equally low values for _S have been reported in adrenal chromaffin cells (40; Zhou & Neher, 1993), hypoglossal motoneurons (41; Lips & Keller, 1998) and spinal motoneurons (50; Palecek et al. 1999). These cell types form a group which have the lowest _S values among the cells which have been investigated thus far. Together with the fact that no evidence for mobile buffers was found in the excitatory cells of this study, the estimate for _S in excitatory neurons seems to represent a value typical for cells that do not express high levels of any particular CaBP.

Expression patterns of CaBPs in non-principal cells are more complicated. Essentially, PV-, CB-, and CR-IR hippocampal GABAergic neurons represent non-overlapping subpopulations of non-pyramidal cells in rat hippocampus (Miettinen et al. 1992). Among these three major CaBPs, PV and CR are present exclusively in non-principal cells with the exception of mossy cells, which are excitatory CR-IR cells (Miettinen et al. 1992). Recent studies have found that most CR-IR non-principal cells (> 90 %) are GABAergic, regardless of the presence of spiny dendritic structures (Miettinen et al. 1992; Martinez et al. 1999). Although CR is known to have four high affinity Ca²⁺ binding sites with a K_d of 1·5 µM and one low affinity site with a K_d of 0·5 mM (Schwaller et al. 1997), nothing is yet known about its binding rate constants for Ca²⁺. Since Mg²⁺ does not compete with Ca²⁺ for the binding sites of CR, the on-rate is most probably not as slow as that of PV (Schwaller et al. 1997). Thus, CR should be a plausible candidate to explain the difference in _S between inhibitory and excitatory cultured neurons, since CR-IR cells represent a larger population of GABAergic neurons than those containing CB. However, the existence of other unknown CaBPs in GABAergic neurons cannot be ruled out, since the population of cells positive to any one of the three CaBPs represents only half of all hippocampal inhibitory neurons (Miettinen et al. 1992).

Is PV responsible for the double exponential decay of [Ca²⁺] transients induced by single action potentials?

Several previous reports are consistent with the argument that PV is responsible for the anomalous cases of [Ca²⁺] decay that we observed in some inhibitory cells (see Results section). Such time courses are incompatible with the single compartment model, in which all buffers are assumed to be fast (Neher, 1998). The distinguishing property of PV is its high affinity for Ca²⁺ and the slowness in its Ca²⁺ binding kinetics. The accompanying paper (Lee et al. 2000) indicates that this slow kinetic property causes biexponential [Ca²⁺] transients when is comparable to the equilibration rate of Ca²⁺ binding to PV. In other words, the slowness, and not the saturability, of the buffer is responsible for the biexponential decay in [Ca²⁺] transients.

Double exponential decays similar to those observed in our experiments were observed in pyramidal cells after trains of action potentials (Helmchen et al. 1996). These findings seem initially contradictory to the argument presented here (that PV is responsible for this double exponential decay), since pyramidal cells are known to lack PV. However, in pyramidal cells, double exponential decays were observed only following trains of action potentials when the amplitude of [Ca²⁺] transients was greatly increased and when [Ca²⁺] was maintained at a high level for seconds. Single exponentials were sufficient to explain the [Ca²⁺] transients evoked by single action potentials in pyramidal cells (Helmchen et al. 1996). Also, the double exponential decay observed in that study cannot be attributed to the slow kinetics of a Ca²⁺ buffer, since [Ca²⁺] was maintained high for about 2 s during trains of action potentials (24 Hz), long enough for Ca²⁺ buffers to equilibrate with Ca²⁺. Saturation of a rapid buffer during the period of elevated [Ca²⁺] is a more likely explanantion for the complicated time course of [Ca²⁺] decay in such a case. Therefore, the biexponential decay following very short stimuli of small to moderate amplitude, reported here, is most probably indicative of the presence of a slow buffer. In an accompanying paper this prediction is tested by injecting PV into adrenal chromaffin cells, where [Ca²⁺] can be very easily and accurately measured (Lee et al. 2000).

	REFERENCES

Top Abstract Introduction Methods Results Discussion References

Baimbridge, K., Celio, M. & Rogers, J. (1992). Calcium-binding proteins in the nervous system. Trends in Neurosciences 15, 303-308	[Medline]
Baimbridge, K. & Miller, J. (1982). Immunohistochemical localization of calcium-binding protein in the cerebellum, hippocampal formation, and olfactory bulb of the rat. Brain Research 245, 223-229	[Medline]
Bekkers, J. & Stevens, C. (1991). Excitatory and inhibitory autaptic currents in isolated hippocampal neurones maintained in cell culture. Proceedings of the National Academy of Sciences of the USA 88, 7834-7838	[Abstract]
Celio, M. R. (1986). Parvalbumin in most -aminobutyric acid-containing neurons of the rat cerebral cortex. Science 231, 995-997	[Medline]
Connors, B. & Gutnick, M. (1990). Intrinsic firing patterns of diverse neocortical neurons. Trends in Neurosciences 13, 99-104	[Medline]
Eilers, J., Callewaert, G., Armstrong, C. & Konnerth, A. (1995). Calcium signaling in a narrow somatic submembrane shell during synaptic activity in cerebellar Purkinje neurons. Proceedings of the National Academy of Sciences of the USA 92, 10272-10276	[Abstract]
Fierro, L. & Llano, I. (1996). High endogenous calcium buffering in Purkinje cells from rat cerebellar slices. The Journal of Physiology 496, 617-625	[Abstract]
Fujise, N., Liu, Y., Hori, N. & Kosaka, T. (1998). Distribution of calretinin immunoreactivity in the mouse dentate gyrus: II. Mossy cells, with special reference to their dorsoventral difference in calretinin immunoreactivity. Neuroscience 82, 181-200	[Medline]
Helmchen, F., Imoto, K. & Sakmann, B. (1996). Ca2+ buffering and action potential-evoked Ca2+ signalling in dendrites of pyramidal neurons. Biophysical Journal 70, 1069-1081	[Abstract]
Hou, T.-T., Johnson, J. D. & Rall, J. A. (1992). Effect of temperature on relaxation rate and Ca2+, Mg2+ dissociation rates from parvalbumin of frog muscle fibres. The Journal of Physiology 449, 399-410	[Abstract]
Jiang, M. & Swann, J. W. (1997). Expression of calretinin in diverse neuronal population during development of rat hippocampus. Neuroscience 81, 1137-1154	[Medline]
Kawaguchi, Y., Katsumaru, H., Kosaka, Y., Heizmann, C. & Hama, K. (1987). Fast spiking cells in rat hippocampus (CA1 region) contain the calcium-binding protein parvalbumin. Brain Research 416, 369-374	[Medline]
Lee, S.-H., Schwaller, B. & Neher, E. (2000). Kinetics of Ca2+ binding to parvalbumin in bovine chromaffin cells: implications for [Ca2+] transients of neuronal dendrites. The Journal of Physiology 525, 419-432.	[Abstract/Full Text]
Lips, M. B. & Keller, B. U. (1998). Endogenous calcium buffering in motoneurones of the nucleus hypoglossus from mouse. The Journal of Physiology 511, 105-117	[Abstract/Full Text]
Martinez, A., Ruiz, M. & Soriano, E. (1999). Spiny calretinin-immunoreactive neurons in the hilus and CA3 region of the rat hippocampus: Local axon circuits, synaptic connections, and glutaminergic acid decarboxylase 65/67 mRNA expression. Journal of Comparative Neurology 404, 438-448	[Medline]
Miettinen, R., Gulyas, A. I., Baimbridge, K. G., Jacobowitz, D. M. & Freund, T. F. (1992). Calretinin is present in non-pyramidal cells of the rat hippocampus - II. Co-existence with other calcium binding proteins and GABA. Neuroscience 48, 29-43	[Medline]
Neher, E. (1989). Combined fura-2 and patch clamp measurements in rat peritoneal mast cells. In Neuromuscular Junction, ed. Sellin, L. R. & Thesleff, S., pp. 65-76. Elsevier Science Publisher, Amsterdam.
Neher, E. (1998). Usefulness and limitation of linear approximation to the understanding of Ca2+ signals. Cell Calcium 24, 345-357	[Medline]
Neher, E. & Augustine, G. (1992). Calcium gradients and buffers in bovine chromaffin cells. The Journal of Physiology 450, 273-301	[Abstract]
Nitsch, R., Soriano, E. & Frotscher, M. (1990). The parvalbumin-containing nonpyramidal neurons in the rat hippocampus. Anatomy and Embryology 181, 413-425	[Medline]
Palecek, J., Lips, M. B. & Keller, B. U. (1999). Calcium dynamics and buffering in motoneurones of the mouse spinal cord. The Journal of Physiology 520, 485-502	[Abstract/Full Text]
Parra, P., Gulyas, A. & Miles, R. (1998). How many subtypes of inhibitory cells in the hippocampus. Neuron 20, 983-993	[Medline]
Pfyffer, G. E., Bologa, L., Herschkowitz, N. & Heizmann, C. W. (1984). Parvalbumin, a neuronal protein in brain cell cultures. Journal of Neurochemistry 43, 49-57	[Abstract]
Plogmann, D. & Celio, M. (1993). Intracellular concentration of parvalbumin in nerve cells. Brain Research 600, 273-279	[Medline]
Schwaller, B., Durussel, I., Jermann, D., Herrmann, B. & Cox, J. (1997). Comparison of the Ca2+-binding properties of human recombinant calretinin-22k and calretinin. Journal of Biological Chemistry 272, 29663-29671	[Abstract/Full Text]
Zhou, Z. & Neher, E. (1993). Mobile and immobile calcium buffers in bovine adrenal chromaffin cells. The Journal of Physiology 469, 245-273	[Abstract]

Acknowledgements

We are grateful to I. Herfort (Max Planck Institute for Biophysical Chemistry, Göttingen) for culturing hippocampal neurons and to B. Belser (Institute of Histology and General Embryology, Fribourg) for technical help. We thank A. Marty, I. Llano and S. Pyott for their helpful and critical reading of this manuscript. This research was supported by grants from the Behrens-Weise-Stiftung and from the European Union (ERBFMRXCT980236) to E.N., from KOSEF to S.-H.L., and from the Swiss National Science Foundation (grant number: 3100-047291.96) to B.S.

Corresponding author

E. Neher: Max Planck Institute for Biophysical Chemistry, Department of Membrane Biophysics, D-37077 Göttingen, Germany.

Email: eneher{at}gwdg.de

Author's present address

S.-H. Lee: Department of Physiology, Seoul National University College of Medicine, Seoul 110-799, Korea.

This article has been cited by other articles:

				E. Neher Details of Ca2+ dynamics matter J. Physiol., April 15, 2008; 586(8): 2031 - 2031. [Full Text] [PDF]

				Y. Aponte, J. Bischofberger, and P. Jonas Efficient Ca2+ buffering in fast-spiking basket cells of rat hippocampus J. Physiol., April 15, 2008; 586(8): 2061 - 2075. [Abstract] [Full Text] [PDF]

				Y. Kubota, J. A. Putkey, H. Z. Shouval, and M. N. Waxham IQ-Motif Proteins Influence Intracellular Free Ca2+ in Hippocampal Neurons Through Their Interactions With Calmodulin J Neurophysiol, January 1, 2008; 99(1): 264 - 276. [Abstract] [Full Text] [PDF]

				D. Lee, K.-H. Lee, W.-K. Ho, and S.-H. Lee Target Cell-Specific Involvement of Presynaptic Mitochondria in Post-Tetanic Potentiation at Hippocampal Mossy Fiber Synapses J. Neurosci., December 12, 2007; 27(50): 13603 - 13613. [Abstract] [Full Text] [PDF]

				J. Torres-Reveron and M. J. Friedlander Properties of Persistent Postnatal Cortical Subplate Neurons J. Neurosci., September 12, 2007; 27(37): 9962 - 9974. [Abstract] [Full Text] [PDF]

				V. Scheuss, R. Yasuda, A. Sobczyk, and K. Svoboda Nonlinear [Ca2+] Signaling in Dendrites and Spines Caused by Activity-Dependent Depression of Ca2+ Extrusion J. Neurosci., August 2, 2006; 26(31): 8183 - 8194. [Abstract] [Full Text] [PDF]

				L. Kreiner and A. Lee Endogenous and Exogenous Ca2+ Buffers Differentially Modulate Ca2+-dependent Inactivation of CaV2.1 Ca2+ Channels J. Biol. Chem., February 24, 2006; 281(8): 4691 - 4698. [Abstract] [Full Text] [PDF]

				M.-H. Kim, N. Korogod, R. Schneggenburger, W.-K. Ho, and S.-H. Lee Interplay between Na+/Ca2+ Exchangers and Mitochondria in Ca2+ Clearance at the Calyx of Held J. Neurosci., June 29, 2005; 25(26): 6057 - 6065. [Abstract] [Full Text] [PDF]

				D. A. Rusakov, F. Saitow, K. P. Lehre, and S. Konishi Modulation of Presynaptic Ca2+ Entry by AMPA Receptors at Individual GABAergic Synapses in the Cerebellum J. Neurosci., May 18, 2005; 25(20): 4930 - 4940. [Abstract] [Full Text] [PDF]

				I. Ismailov, D. Kalikulov, T. Inoue, and M. J. Friedlander The Kinetic Profile of Intracellular Calcium Predicts Long-Term Potentiation and Long-Term Depression J. Neurosci., November 3, 2004; 24(44): 9847 - 9861. [Abstract] [Full Text] [PDF]

V. Matveev, R. S. Zucker, and A. Sherman