Open Access
11 February 2015 Using simultaneous voltage and calcium imaging to study fast Ca2+ channels
Nadia Jaafari, Elodie Marret, Marco Canepari
Author Affiliations +
Abstract
The combination of fluorescence measurements of membrane potential and intracellular Ca2+ concentration allows correlating the electrical and calcium activity of a cell with spatial precision. The technical advances allowing this type of measurement were achieved only recently and represent an important step in the progress of the voltage imaging approach pioneered over 40 years ago by Lawrence B. Cohen. Here, we show how this approach can be used to investigate the function of Ca2+ channels using the foreseen possibility to extract Ca2+ currents from imaging experiments. The kinetics of the Ca2+ current, mediated by voltage-gated Ca2+ channels, can be accurately derived from the Ca2+ fluorescence measurement using Ca2+ indicators with KD>10  μM that equilibrate in <1  ms. In this respect, the imaging apparatus dedicated to this application is described in detail. Next, we illustrate the mathematical procedure to extract the current from the Ca2+ fluorescence change, including a method to calibrate the signal to charge flux density. Finally, we show an example of simultaneous membrane potential and Ca2+ optical measurement associated with an action potential at a CA1 hippocampal pyramidal neuron from a mouse brain slice. The advantages and limitations of this approach are discussed.

1.

Introduction

Since the introduction of organic voltage-sensitive dyes (VSD),1 the membrane potential (Vm) change associated with neuronal activity could be measured at multiple cellular sites using fluorescence.2,3 This information, spatially and temporally correlated with Ca2+ signals, opens the gate to the understanding of many physiological processes underlying neuronal function.4 More recently, VSD imaging was combined, in single neurons, with Ca2+ fluorescence imaging, initially using sequential recordings5,6 and later using simultaneous recordings with two aligned cameras.7 Under general conditions, this experimental strategy allows correlating the change of intracellular Ca2+ concentration with the underlying Vm. Nevertheless, if the change of intracellular Ca2+ concentration is due exclusively to Ca2+ influx through the plasma membrane, the measurement of Ca2+ concentration can be, in principle, converted into a measurement of Ca2+ current.8 Only in this case, when no Ca2+ release from internal stores occurs, the combined optical measurement of Vm using voltage-sensitive dyes permits the study of the voltage regulation of Ca2+ channels.

The theory of obtaining an optical measurement of a fast Ca2+ current starts with the analysis of the dye-Ca2+ binding reaction occurring in the presence of an endogenous cellular Ca2+ buffer. This scenario has been initially studied by Kao and Tsien,9 who established that the relaxation time of the dye-Ca2+ binding reaction (τR) can be approximated by the equation

τR=1KONDye[Ca2+]+KOFFDye.

For all Ca2+ indicators, the association constant (KONDye) is fast and limited by diffusion, typically to 6×108M1s1, while the dissociation constant KOFFDye may vary two orders of magnitude. Thus, τR at different free Ca2+ concentrations ([Ca2+]) can be estimated as a function of the equilibrium constant KDDye=KOFFDye/KONDye. Whereas for high-affinity indicators with KD<1μM, τR is typically >1ms and depends on [Ca2+], for low-affinity indicators with KD>10μM, τR is faster (typically <200μs) and does not depend on [Ca2+]. This theoretical consideration suggests that a fast Ca2+ current with a duration of a few milliseconds can be reliably tracked by low-affinity indicators.

While the relaxation time of the dye-Ca2+ binding reaction sets the limit of detection of a Ca2+ influx, the measurement of a Ca2+ current is possible only if the total Ca2+ entering the cell through the plasma membrane is proportional to the Ca2+ bound to the dye. In a previous report, we described the conditions under which this requirement is fulfilled.10 In addition, we demonstrated that this is the case of the Ca2+ signal associated with the action potential in the CA1 pyramidal neuron of the hippocampus.10

Here, we describe more in detail how to achieve these measurements. In particular, we describe the necessary apparatus as well as the methodological aspects of this novel approach. We reported tests with five Ca2+ indicators [Oregon Green 488 BAPTA-1 (OGB1), Oregon Green 488 BAPTA-6F (OG6F), Oregon Green 488 BAPTA-5N (OG5N), Bis-Fura2 (BF2), and FuraFF] and, for the low-affinity indicators, confirmed their ability to track a Ca2+ current by performing computer simulations. We also describe an alternative approach of data analysis using the fit of the Ca2+ transient to obtain the Ca2+ current from a reduced number of trials. We finally further discuss how this novel approach overcomes the limitations of Ca2+ current measurements using electrode techniques.

2.

Materials and Methods

2.1.

Slice Preparation, Solutions, and Electrophysiology

Experiments were approved by the Isere prefecture (Authorisation n. 38 12 01) and the specific protocol (n. 197) by the ethics committee of the Grenoble Institute of Neuroscience. Hippocampal slices (250μm thick) were prepared from 21 to 35 postnatal days old C57Bl6 mice as previously described11 using either a VF-200 compresstome (Precisionary Instruments, Greenville, NC) or a Leica VT1200 (Leica, Wetzlar, Germany). Slices were cut in iced extracellular solution and incubated at 37°C for 1 h before use. The extracellular solution used contained (in mM) 125 NaCl, 26NaHCO3, 1MgSO4, 3 KCl, 1NaH2PO4, 2CaCl2, and 20 glucose, bubbled with 95% O2 and 5% CO2. The intracellular solution contained (in mM) 125KMeSO4, 5 KCl, 8MgSO4, 5Na2ATP, 0.3 tris-GTP, 12 tris-phosphocreatine, and 20 HEPES, adjusted to pH 7.35 with KOH. To block Na+ and K+ channels in voltage clamp experiments, the external solution also contained 1μM tetrodotoxin, 5 mM tetraethylammonium (TEA), 4 mM 4-aminopyridine, and 250 nM apamin, and the internal solution also contained 5 mM TEA. In this work, five Ca2+ indicators (all purchased from Invitrogen, Carlsbad, California) were used at the concentration of 1 mM: OGB1 (KD=0.21μM, Ref. 12), OG6F (KD=3μM, as reported by the vendor), OG5N (KD=35μM, Ref. 13), BF2 (KD=0.53μM, as reported by the vendor), and FuraFF (KD=10μM, Ref. 14). Ca2+ indicators were dissolved directly into the internal solution. In voltage-imaging experiments, cells were loaded as previously described4 with the voltage-sensitive dye JPW1114 (Invitrogen). In calibrating experiments using the Ca2+-releasing caged compound nitrophenyl-EGTA (NP-EGTA)15 (Invitrogen), the internal solution also contained 150μM CaCl2. All other chemicals were purchased either from Tocris (Bristol, UK) or Sigma-Aldrich (St. Louis, Missouri). Experiments were performed at 32 to 34°C using an Olympus BX51 microscope equipped with a 60×/1.0NA Nikon objective. Patch-clamp recordings were made using a Multiclamp amplifier 700A (Molecular Devices, Sunnyvale, California), and voltage and current signals were acquired with the analog-to-digital board of the CCD camera. The Vm measured with the patch pipette was corrected for the junction potential (11mV) as calculated with the JPCalc software.16 In voltage clamp recordings, the Ca2+ current was evoked by depolarizing pulses from 70 to 10mV and measured in the soma by subtracting the scaled subthreshold current associated with a voltage step from 70 to 60mV.

2.2.

Imaging System

A schematic of the imaging system is shown in Fig. 1(a). Simultaneous UV/blue light-emitting diode (LED) illumination from the epifluorescence port of the microscope was allowed by a 409 nm dichroic mirror (FF409, Semrock, Rochester, New York). UV light was either at 365 nm for photolysis or at 385 nm for excitation of Fura indicators fluorescence. The 365-nm LED was controlled by an OptoFlash (CAIRN Research Ltd., Faversham, UK) and the other LEDs were controlled by an OptoLED (Cairn). The image could be demagnified either by 0.5 or by 0.25, obtaining the fields of view shown in Fig. 1(b). The two images were separated by a 593 nm dichroic mirror and acquired with a dual-head NeuroCCD-SMQ camera (Redshirt, Decatur, Georgia) at 510±42nm (with the Ca2+ CCD) and at >610nm (with the Vm CCD). The Ca2+ image was, therefore, the mirrored picture of the Vm image. Whereas the camera has 80×80 pixels per head (full resolution), to achieve the speed of 20 kHz, binned stripes of 26×4 pixels were imaged. To achieve simultaneous illumination and detection of OG5N and JPW1114, both indicators were excited at 470 nm. The sensitivity of the VSD at 470 nm was about four times less than that at 532 nm. Thus, while an action potential is associated with a fractional change of fluorescence (ΔF/F0) of 2 to 8% in the proximal apical dendrite at 532 nm, the same signal is associated with ΔF/F0 of 0.5 to 2% at 470 nm. Using this configuration, the emission of JPW1114 detected by the Ca2+ CCD is negligible as shown in Fig. 1(c), and the emission of OG5N detected by the Vm CCD is also negligible as shown in Fig. 1(d). Thus, it is possible to discriminate and quantify Ca2+ fluorescence from Vm fluorescence.

Fig. 1

The imaging system. (a) Schematic of the imaging system including simultaneous UV/blue light-emitting diode illumination and the dual-head CCD camera (illustrated in the picture) for simultaneous Vm and Ca2+ imaging; the image from the 60× objective is demagnified by a coupler before detection. (b) Images of a micrometer slide from the two aligned heads of the camera using 0.5× and 0.25× demagnifing couplers; the distance between two adjacent lines is 10μm. (c) Top: fluorescence subimages of the apical dendrite of a CA1 hippocampal pyramidal neuron filled with 1 mM Oregon Green 488 BAPTA-5N (OG5N); resting fluorescence from the Vm camera is <5% with respect to that from the Ca2+ camera; bottom: the fluorescence transient associated with an action potential is detected with the Ca2+ camera, and it is negligible at the Vm camera. (d) Top: fluorescence subimages of the apical dendrite of another CA1 hippocampal pyramidal neuron filled with JPW1114; resting fluorescence from the Ca2+ camera is <5% with respect to that from the Vm CCD; bottom: the fluorescence transient associated with an action potential is detected with the Vm CCD, and it is negligible at the Ca2+ camera.

NPh_2_2_021010_f001.png

2.3.

Recording and Analysis of Ca2+ and Vm Optical Signals

Ca2+ and Vm recordings were performed at 20 kHz and data were initially expressed as ΔF/F0. The high acquisition rate was necessary to prevent the loss of temporal information while applying the filtering procedure described below. Ca2+ recordings started 20 to 25 minutes after establishing the whole cell configuration, i.e., the time necessary to achieve the equilibrium of the indicator over the dendritic segment of recording. To improve the signal-to-noise ratio (S/N), fluorescence was generally averaged over 9 to 64 trials as specified in each figure legend and corrected for bleaching using trials without a signal. All data analysis was performed using software written in MATLAB® (The Mathworks, Natick, Massachusetts). The JPW1114-ΔF/F0 signal was calibrated in terms of Vm change using a previously demonstrated protocol based on wide-field photorelease of L-glutamate from the caged compound 4-methoxy-7-nitroindolinyl-caged-L-glutamate (MNI-glutamate).17 Briefly, a calibration of the JPW1114-ΔF/F0 signal can be done if an electrical signal of known amplitude is available at all recording sites. In many neuronal types, activation of a large portion of ionotropic glutamate receptors makes them the dominant conductance and the resulting Vm will be 0 mV in all the illuminated area. In CA1 hippocampal pyramidal neurons, a calibration is possible since the resting Vm is uniform over the whole cell.18 Thus, the JPW1114-ΔF/F0 signal associated with a saturating photorelease of L-glutamate will correspond to a change of Vm from the resting Vm to 0 mV. A representative cell where this procedure was applied is shown in Fig. 2(a). The JPW1114-ΔF/F0 signal associated with an action potential was recorded from the initial 50μm apical dendritic segment of a CA1 hippocampal pyramidal neuron [Fig. 2(b)]. After this recording, we added 1μM tetrodotoxin to block voltage-gated Na+ channels and 1 mM MNI-glutamate to the external solution. Starting from the resting Vm (60mV), we uncaged L-glutamate producing a depolarization to 0 mV. The glutamate-induced optical signal [Fig. 2(b), green] provided the calibration for the action potential. Statistical significance was evaluated using either the paired t test or the two-population (2P) t test.

Fig. 2

Calibration of JPW1114-ΔF/F0 in terms of ΔVm. (a) Initial apical dendritic segment of a CA1 hippocampal pyramidal neuron loaded with the VSD JPW1114; the region from where fluorescence was averaged is outlined in yellow. (b) JPW1114-ΔF/F0 associated with an action potential or glutamate photorelease; the calibration protocol is applied from the resting membrane potential, which is assumed to be uniform over the cell; the massive ionotropic glutamate receptors activation drives the illuminated dendritic area to 0 mV. The action potential trace is from an average of 32 trials; the calibration trace was from an average of nine trials.

NPh_2_2_021010_f002.png

2.4.

Computer Simulations

The results of the analysis of the kinetics of the five Ca2+ indicators were compared with those obtained by computer simulations using a previously published theoretical framework.19 Briefly, we simulated the reaction of the dye ([D]) with Ca2+ in the presence of 1 mM of an endogenous buffer ([B]). For comparison with voltage clamp experiments, using the current measured with the patch electrode as Ca2+ current (ICa), we numerically solved the set of differential equations:

d[Ca2+]dt=ICaKOND·[Ca2+]·[D]+KOFFD·[DCa2+]KONB·[Ca2+]·[B]+KOFFB·[BCa2+]d[DCa2+]dt=KOND·[Ca2+]·[D]KOFFD·[DCa2+]d[BCa2+]dt=KONB·[Ca2+]·[B]KOFFB·[BCa2+]d[D]dt=d[DCa2+]dtd[B]dt=d[BCa2+]dt.

The values of the association and equilibrium constants of the endogenous buffer were KONB=2×108M1s1 and KDB=105M, while KOFFB=KONB·KDB. The value of the association constant of all Ca2+ indicators was KOND=6×108M1s1, as reported by Kao and Tsien,9 while KOFFD=KOND·KDD was derived for each different indicator from the KD values reported above. Numerical solutions were obtained using the MATLAB® function “ode45.”

3.

Results

3.1.

Comparative Analysis of the Kinetics of Different Ca2+ Indicators

The theoretical considerations described in the Introduction suggest that the kinetics of a fast Ca2+ current with a duration of a few milliseconds can be reconstructed from the fluorescence change of a low-affinity indicator. The ultimate strategy to test this hypothesis is to verify that the ΔF/F0 Ca2+ signal is linear with the integral of the Ca2+ current measured, under voltage clamp, in the presence of Na+ and K+ blockers (see Materials and Methods). The protocol used in these tests, illustrated in Fig. 3(a), consisted of two depolarizing square pulses. The current associated with the first pulse, from 70 to 60mV not evoking a Ca2+ current, was scaled and subtracted from the current associated with a pulse from 70 to +10mV. This procedure permitted the extraction of the Ca2+ current from the total current recording containing the leak current and the capacity transient. The integral of the Ca2+ current obtained in this way was then compared to the ΔF/F0 signal. Figure 3(b) shows results from five representative cells, each recorded with a different indicator: the first 7 ms of ICa normalized to 1 (black traces), the corresponding normalized ICa integrals (red traces), and the associated normalized ΔF/F0 signals (blue traces). The signals obtained in these experiments were faithfully mimicked by the results of computer simulations, reported on the right, performed as described in the Materials and Methods.

Fig. 3

Kinetics of different Ca2+ indicators investigated in voltage clamp. (a) Voltage clamp protocol consisting of two 16 ms voltage pulses from 70mV, the first one below the threshold for activation of Ca2+ channels and the second one to 20mV; experiments were performed in the presence of Na+ and K+ channel blockers; the Ca2+ current (Ica) was extracted by subtraction of the scaled subthreshold current; the kinetics of the current were analyzed in correlation with that of the Ca2+ fluorescence change in the initial 60μm segment of the apical dendrite; all data were averages of 16 trials. (b) Ica and Ca2+ ΔF/F0 from five representative cells filled with 1 mM of one of the following indicators: Oregon Green 488 BAPTA-1 (OGB1), Oregon Green 488 BAPTA-6F (OG6F), OG5N, Bis-Fura2 (BF2), or FuraFF; the integrals of ICa (ICa) are also shown; signals are normalized to their maxima over the first 7 ms after the pulse beginning; simulations of the dye-Ca2+ binding reaction are reported on the right. (c) From two of the representative cells filled either with OGB1 or with OG5N, the ICa and Ca2+ ΔF/F0 normalized to their maxima over the first 4 ms after the pulse beginning; the difference between the two curves is also reported. (d) Mean±SD of the surface of ICaΔF/F0 (S) obtained for each of the five indicators tested from the number of cells reported; OGB1: 0.185±0.034, N=5; OG6F: 0.049±0.023, N=5; OGB1: 0.007±0.018, N=12; BF2: 0.375±0.150, N=4; FuraFF: 0.002±0.023, N=6; the values of S for OGB1 with respect to those for OG5N, and the values of S for BF2 with respect to those for FuraFF were significantly different (two-population t test, *p<0.001 and **p<0.002, respectively). Panels (a) and (c) were adapted from Ref. 10.

NPh_2_2_021010_f003.png

To quantify the difference of kinetics between the Ca2+ ΔF/F0 signal and the ICa integral we computed the area (S) of the difference of the two curves normalized to 1 over the first 4 ms as shown in Fig. 3(c). The statistics of S for the five groups of cells (N=4 to 12 for each indicator) is illustrated in Fig. 3(d). The values of S with the two low-affinity indicators, OG5N and FuraFF, were statistically different from the values of S with the high-affinity indicators, OGB1 and BF2 (2P t test, p<0.02). In particular, the low-affinity indicators were able to track fast Ca2+ currents. OG5N-ΔF/F0 has an S/N more than times three larger than that of FuraFF-ΔF/F0 when associated with the same Ca2+ signal. For this reason, OG5N should be used for measuring relative small Ca2+ currents as those reported here. However, FuraFF can be better combined with JPW1114 for simultaneous Vm and Ca2+ imaging7 and should be used for larger Ca2+ currents where fluorescence averaging over many trials is less critical. The tests presented here suggest that all organic Ca2+ indicators with KD for Ca2+ of >10μM can be potentially used to faithfully extract the time course of a fast Ca2+ current.

3.2.

Calibration of the OG5N-ΔF/F0 Signal and Extraction of the Ca2+ Current

An important aspect of the Ca2+ imaging technique is its ability to provide an estimate of the intracellular Ca2+ concentration. In general terms, this estimate is not straightforward since the important information may involve knowing the free Ca2+ concentration ([Ca2+]FREE) as well as the Ca2+ that binds to the dye and to the endogenous buffer. However, for a quantitative estimate of ICa, the important information is exclusively the total Ca2+ concentration ([Ca2+]TOT) entering the cell. The estimate of [Ca2+]TOT corresponding to a particular ΔF/F0 signal will allow determining the charge concentration and deriving the corresponding ICa volume density (ICa,/V).

A way to produce a [Ca2+]TOT of known amplitude is to release Ca2+ from a caged compound. The cell in Fig. 4(a) was filled with an intracellular solution, also containing 300μM of NP-EGTA and 150μM CaCl2. NP-EGTA before photolysis is a high-affinity Ca2+ chelator (KD=80nM). The KD decreases by a factor of 12,500 after photolysis.15 While in the pipette 50% of the NP-EGTA is bound to Ca2+, this percentage is unknown in the cell where it is regulated by the global Ca2+ homeostasis. However, when the cell of Fig. 4(a) was depolarized from 70 to 10mV, a persistent OG5N-ΔF/F0 signal >20% was observed. Since this signal typically corresponds to >10 times the signal associated with one action potential and [Ca2+]FREE associated with an action potential in this dendrite type is >100nM,20 [Ca2+]FREE at 10mV is >1μM. Since KD of NP-EGTA is 80 nM, at Vm=10mV, >92% of NP-EGTA is always bound to Ca2+. Under this condition, as shown in Fig. 4(a), the Ca2+ indicator is still not saturated since a further depolarization step of +50 mV produced an OG5N-ΔF/F0 signal >20%. Having established that at Vm=10mV, 300μM is bound to NP-EGTA and available for release, the calibration protocol illustrated in Fig. 4(b) consisted of applying a sequence of 16 UV pulses (of 20 ms duration every 5 s) and measuring the OG5N-ΔF/F0 signal associated with each pulse. The [Ca2+]TOT released at the pulse k follows the geometric progression:21

[Ca2+]TOT(k)=α·[300μM-j=0k1[Ca2+]TOT(j)],
where [Ca2+]TOT (k=0)=0. The OG5N-ΔF/F0 corresponding to a given [Ca2+]TOT is obtained by fitting this geometric progression to the sequential decrease of the OG5N-ΔF/F0 amplitude. In this particular cell, we obtained that an OG5N-ΔF/F0 of 1% corresponded to [Ca2+]TOT=18μM. From this calibration the ICa/V associated with the action potential was derived in the following way, as shown in Fig. 4(c). First, the ΔF/F0 signal was converted into a charge-to-volume ratio (Q/V) using the equation
Q/V=[Ca2+]TOT·2e·NAdm3,
where e is the fundamental charge, dm is the decimeter unit, and NA is the Avogadro number. Second, the converted signal was filtered using a Savitzky-Golay filter and preserving the kinetics of the original signal.22 Finally, ICa/V was obtained by applying the time derivative to the filtered signal as shown in Fig. 4(c). Previously reported data analysis using this calibration procedure indicated that the [Ca2+]TOT corresponding to an OG5N-ΔF/F0 signal of 1% varies from 16 to 24μM.10 Thus, a standard estimate of OG5N-ΔF/F0=1% corresponding to·[Ca2+]TOT=20μM can be applied in cells where the calibration protocol is not directly performed.

Fig. 4

Calibration of Ca2+ currents using Ca2+ photorelease. (a) OG5N-ΔF/F0 from the initial 80μm segment of the cell shown on the top (region of fluorescence average outlined) associated with a depolarization from 70 to 10mV (top) and from 10 to +40mV (bottom). (b) From the same cell filled with a solution containing 300μM nitrophenyl-EGTA and 150μM CaCl2, sequential UV-flashes (pulses) at 10mV produced an OG5N-ΔF/F0 signal of decreasing amplitude (left); the OG5N-ΔF/F0 amplitude versus the pulse was fitted with a geometric sequence to obtain the efficiency of each pulse (α=0.2 in this example); considering the Ca2+ photoreleased by the first pulse, a signal of 1% corresponded to [Ca2+]TOT=18μM in this specific cell. (c) Calibrated OG5N-ΔF/F0 (middle trace) associated with an action potential (top trace) and corresponding ICa/V (bottom trace). Data are from an average of 64 trials. Panel (a) was adapted from Ref. 10.

NPh_2_2_021010_f004.png

3.3.

Estimate of ICa/V Signals by ΔF/F0 Filtering or Curve Fitting

The OG5N-ΔF/F0 signal associated with an action potential can be measured from a 50μm dendritic segment with good S/N, which can be further improved by averaging several trials. The extraction of ICa/V, however, is based on the calculation of the time derivative, which requires the signal noise to be smaller than the signal change between two consecutive samples. A way to obtain this condition is to temporally filter the signal up to the limit of distortion of its kinetics. In our previous report, we have shown that the Savitzky-Golay algorithm is an optimal filtering tool for improving the S/N of the OG5N-ΔF/F0 signal without significant temporal distortion using time windows of up to 20 to 30 samples.10 The application of this procedure allows extracting ICa/V from relatively large dendritic compartments and averaging several trials. The challenge for the improvement of data analysis was to go beyond the limitations of the filtering approach. This can be done by fitting the raw or the filtered OG5N-ΔF/F0 signal with a model function obtaining a noiseless curve that mimics the time course of the OG5N-ΔF/F0 signal. While the choice of the model might be arbitrary, a simple function that resembles the time course of the OG5N-ΔF/F0 transient is the sigmoid. In particular, the OG5N-ΔF/F0 signal normalized to its asymptotic value can be mimicked by the products of N sigmoid functions Y:

Y(t)=j=1Nsigm11+eφj·(tθj),
where t is time, ϕj and θj are the parameters to be determined by the fit, and NSigm is the arbitrary number of sigmoid functions. We tried several integer values for NSigm, and we found that 3 was the minimum number of sigmoid functions always providing a fit with the same kinetics of the OG5N-ΔF/F0 signal. The product of three sigmoid functions was, therefore, set as the general model to fit the signal, and its ability to extract ICa/V either from small dendritic regions or from relatively large regions in single trials was assessed in the cell reported in Fig. 5(a). In particular, we analyzed signals associated with an action potential from an average of 32 trials or from a single trial [Fig. 5(b)]. Figure 5(c) shows the ΔF/F0 signals from the average of 32 trials in a large dendritic region of ~50μm length (region A) and in a small dendritic site of 2.4μm (region B). The same panel (right traces) also shows the ΔF/F0 signal in region A from the single trial. Each raw signal was filtered with the Savitzky-Golay algorithm at a time window of 10, 20, or 40 samples. Alternatively, it was fitted with the three-sigmoid model. The filtering algorithm improves the S/N but starts producing a temporal distortion at 40 samples. The consequent distortion of the extracted ICa/V is evident in the case of the average of 32 trials in region A shown in the left column of Fig. 5(d). While in this case the ICa/V obtained by filtering with a time window of 20 samples has an acceptable S/N, the same filtering approach applied to extract ICa/V either from region B or from region A in a single trial generates a noisy signal (middle and right traces). In all cases, however, the fit approach generates noiseless curves that faithfully reproduce the time course of the OG5N-ΔF/F0 signals from which ICa/V signals can be extracted. This technical improvement in data analysis is critical for many applications. In particular, when the spikes occurrence is stochastic, ICa/V must be extracted from single trials. Alternatively, the gain in analysis performance can be spent to extract ICa/V from small or relatively dim regions.

Fig. 5

Extraction of the Ca2+ current using filtering or curve fitting. (a) Fluorescence images of a cell filled with OG5N; region A comprises 50μm of dendrite from where fluorescence was averaged; region B is a single pixel. (b) Action potential signal from an average of 32 trials (left) or from a single trial (right). (c) From the same trials above, OG5N-ΔF/F0 signals from the average of 32 trials in region A (left), from the average of 32 trials in region B (middle), or from the single trial in region B (right); the signals were either unfiltered (top, blue traces), filtered with the Savitzky-Golay algorithm at time windows of 10 (SG10), 20 (SG20), and 40 (SG40) samples, or fitted by the product of three sigmoid functions (FIT); the filter algorithm and the fit were implemented by the “smooth” and “lsqcurvefit” MATLAB® functions, respectively. (d) From the filtered or the fitted traces above, ICa/V were calculated for the three cases: average of 32 trials in region A (left), average of 32 trials in region B (middle), or single trial in region A (right); the approach of data fitting permits ICa/V extraction from sites of 2.4μm using averaged trials or from larger regions in single trials.

NPh_2_2_021010_f005.png

3.4.

Simultaneous Vm and ICa/V Optical Measurements

To illustrate the ability of simultaneous Vm and ICa/V optical measurements to investigate Ca2+ currents and the behavior of underlying Ca2+ channels, we report here an analysis of the ICa/V associated with an action potential in the initial segment of the apical dendrite of the CA1 hippocampal pyramidal neuron. In the cell of Fig. 6(a), action potentials were evoked by current injection and recorded optically in the first 50μm dendritic segment as reported in the top of Fig. 6(b). One action potential was evoked starting from Vm=60mV and another action potential was evoked starting from Vm=80mV. The simultaneously recorded ICa/V signals are shown in the bottom of Fig. 6(b). The ICa/V associated with the action potential starting from Vm=80mV was larger than that of the ICa/V associated with the action potential starting from Vm=60mV. The time course of ICa/V can be precisely correlated with the time course of Vm as shown in Fig. 6(c). In 14 cells in which this protocol was applied, the mean ±S.E.M. of the ICa/V peak in the first 50μm dendritic segment was 7.2±1.7pA/μm3 for the action potential starting at Vm=60mV and 10.1±1.4pA/μm3 for the action potential starting at Vm=80mV as shown in Fig. 6(d). The ICa/V was consistently larger for the action potential starting at Vm=80mV (p<0.001, paired t test). As demonstrated in a previous report,10 the different ICa/V amplitude observed under the two Vm conditions was mainly due to activation of T-type voltage-activated Ca2+ channels, which recovered from inactivation when the cell was hyperpolarized to Vm=80mV.23

Fig. 6

Dendritic ICa/V associated with an action potential at different starting Vm. (a) Fluorescence images of a cell filled with OG5N and JPW1114; the region from where fluorescence was average is outlined. (b) Optical measurements of Vm (top traces) and ICa/V (bottom traces) with an action potential starting either at Vm=60 or 80mV. (c) ICa/V superimposed to the optically recorded action potential in the two cases. (d) Mean±SEM of the peak ICa/V associated with an action potential starting either at Vm=60 or 80mV from 14 cells; the peak ICa/V was significantly higher for the action potential starting at 80mV (p<0.001, paired t test).

NPh_2_2_021010_f006.png

4.

Discussion

Since the introduction of organic VSD and thanks to the work of Lawrence B. Cohen and colleagues, it was expected that the improved S/N of Vm measurements24 could open the gate to applications where an imaging approach could eventually overcome the intrinsic limitations of electrode techniques. The more recent introduction of fluorescent Vm probes that could be directly injected into cells25 allowed recordings of Vm of <1mV.26

The patch clamp technique, however, is routinely used to measure not only Vm, but also the membrane current in the voltage clamp mode. Our current understanding of ion channel function is mainly derived from the use of the patch clamp techniques. This measurement, in a single-electrode or two-electrode voltage clamp, consists of maintaining the cell at a given Vm by compensating the cell current with an injected current.27 The current measured with the electrode is, therefore, the summation of the filtered currents from all different cellular regions, including remote regions where Vm is unclamped.28 To study the function of native ion channels at different sites of neurons and under physiological electrical activity, this approach is practically not applicable for a series of reasons. First, the voltage clamp condition is not applicable in a cell where Vm is not uniform. Second, although local current measurements from a limited number of cellular sites can be obtained using dendritic patch clamp recordings,29 not only the number of recording sites is limited with respect to an imaging approach but each current is also the summation of currents from different sites, filtered with respect to the site of recording. In other words, a voltage clamp current measurement carries no information about the site of origin of the current. Third, the physiological behavior of a channel occurs when Vm changes physiologically, which is not the condition of the voltage clamp. Injecting a current to dynamically clamp the cell and reproducing a physiological Vm change, for instance, an action potential, does not solve the problem since the potential drop due to the series resistance and the space-clamp of the cell will generate a nonuniform distortion of the Vm form. Fourth, separation of a current mediated by a specific ion requires the pharmacological block of all other currents. Obviously, these channels are necessary for a physiological Vm change. For instance, an action potential requires Na+, K+, and Ca2+ channels, and the study of a specific current during this event would require the block of the other two currents.

For Ca2+ currents, the imaging approach described in this report potentially overcomes all these limitations. In contrast to patch clamp recordings, these calcium optical currents can be measured in conditions of a physiological change of Vm. The measured currents are confined to the sites where they are recorded. The additional information on local Vm change, necessary to correlate the behavior of the conductance with its biophysical properties, is obtained by performing combined Vm and Ca2+ imaging experiments. The possibility to extract ICa from Ca2+ imaging experiments was foreseen in the past. Thus, as expected by the pioneering study of Kao and Tsien9 and by the computer simulations reported here, the kinetics of ICa can be reconstructed without distortion and correlated with Vm change. The crucial information on the Vm at the same site of the ICa is given by the concomitant Vm optical measurement. In this way, it is possible to follow the activation and deactivation of the channel as a function of the Vm change during physiological activity. The optical measurement of ICa is, however, affected by some limitations. First, obviously, it cannot be performed when a concomitant Ca2+ release from internal stores occurs. Second, it cannot be applied to measure a slow ICa that occurs at the same time scale of Ca2+ sequestration and extrusion. This, as shown in a previous report,10 will introduce a negative current artifact in the ICa measurement. Third, in the optical measurement, the calibration can be provided in terms of ICa/V, while the evaluation of ICa requires precise morphological reconstruction of the site of recording.

The optical measurements of ICa became possible only recently, thanks to the remarkable advance in light sources and detection devices for Vm and ion imaging.30 The simultaneous measurement of Vm and ICa/V can be also achieved using the low-affinity indicator Fura-FF and the optical settings described in a previous article.7 The S/N of Fura-FF ΔF/F0 signal, however, is substantially lower compared to the equivalent OG5N-ΔF/F0 signal. Another possibility for simultaneous measurements of Vm and ICa/V using OG5N is its combination with a red-excitable VSD, such as those described by Yan et al.31,32 Such indicators would allow combining simultaneous blue illumination for OG5N with red optimal illumination for the VSD. Among these indicators, we tested Di-2-ANBDQ(F)PTEA, and we found that its fluorescence excited at 470 nm decreases in intensity when Vm is depolarized (data not shown). Thus, when exciting Di-2-ANBDQ(F)PTEA at 470 and at 630 nm simultaneously, the ΔF/F0 signal associated with a Vm change is smaller than the equivalent ΔF/F0 signal obtained by red illumination only. Finally, for signals that are consistent from trial to trial, measurements of Vm and ICa/V can be obtained sequentially with the combination of JPW1114 and OG5N by alternating optimal excitation of Ca2+ fluorescence at 470 nm with optimal excitation of Vm fluorescence at 532 nm.

The present approach should drastically improve our understanding of the physiological function of Ca2+ channels by providing the possibility to explore the biophysics of native channels during physiological activity in subcellular loci of the complex neuronal architecture. The method described here is, therefore, another important brick in the long history of membrane potential imaging applications.

Acknowledgments

We thank Philippe Moreau for technical help and Jean-Claude Vial for useful discussions. This work was supported by the Agence Nationale de la Recherche (Grant Voltimagmicro, program Emergence-10, Labex Ion Channels Science and Therapeutics: program number ANR-11-LABX-0015 and National Infrastructure France Life Imaging Noeud Grenoblois).

References

1. 

L. B. Cohen et al., “Changes in axon fluorescence during activity: molecular probes of membrane potential,” J. Membr. Biol., 19 (1), 1 –36 (1974). http://dx.doi.org/10.1007/BF01869968 JMBBBO 0022-2631 Google Scholar

2. 

L. B. Cohen, B. M. Salzberg and A. Grinvald, “Optical methods for monitoring neuron activity,” Annu. Rev. Neurosci., 1 171 –182 (1978). http://dx.doi.org/10.1146/annurev.ne.01.030178.001131 ARNSD5 0147-006X Google Scholar

3. 

M. Zochowski et al., “Imaging membrane potential with voltage-sensitive dyes,” Biol. Bull., 198 (1), 1 –21 (2000). http://dx.doi.org/10.2307/1542798 BIBUBX 0006-3185 Google Scholar

4. 

M. Canepari, K. Vogt and D. Zecevic, “Combining voltage and calcium imaging from neuronal dendrites,” Cell. Mol. Neurobiol., 28 (8), 1079 –1093 (2008). http://dx.doi.org/10.1007/s10571-008-9285-y CMNEDI 0272-4340 Google Scholar

5. 

M. Canepari, M. Djurisic and D. Zecevic, “Dendritic signals from rat hippocampal CA1 pyramidal neurons during coincident pre- and post-synaptic activity: a combined voltage- and calcium-imaging study,” J. Physiol., 580 (2), 463 –484 (2007). http://dx.doi.org/10.1113/jphysiol.2006.125005 JPHYA7 0022-3751 Google Scholar

6. 

M. Canepari and K. E. Vogt, “Dendritic spike saturation of endogenous calcium buffer and induction of postsynaptic cerebellar LTP,” PLoS ONE, 3 (12), e4011 (2008). http://dx.doi.org/10.1371/journal.pone.0004011 1932-6203 Google Scholar

7. 

K. E. Vogt et al., “High-resolution simultaneous voltage and Ca2+ imaging,” J. Physiol., 589 (3), 489 –494 (2011). http://dx.doi.org/10.1113/tjp.2011.589.issue-3 JPHYA7 0022-3751 Google Scholar

8. 

B. L. Sabatini and W. G. Regehr, “Optical measurement of presynaptic calcium currents,” Biophys. J., 74 (3), 1549 –1563 (1998). http://dx.doi.org/10.1016/S0006-3495(98)77867-1 BIOJAU 0006-3495 Google Scholar

9. 

J. P. Kao and R. Y. Tsien, “Ca2+ binding kinetics of fura-2 and azo-1 from temperature-jump relaxation measurements,” Biophys. J., 53 (4), 635 –639 (1988). http://dx.doi.org/10.1016/S0006-3495(88)83142-4 BIOJAU 0006-3495 Google Scholar

10. 

N. Jaafari, M. De Waard and M. Canepari, “Imaging fast calcium currents beyond the limitations of electrode techniques,” Biophys. J., 107 (6), 1280 –1288 (2014). http://dx.doi.org/10.1016/j.bpj.2014.07.059 BIOJAU 0006-3495 Google Scholar

11. 

N. Jaafari et al., “Economic and simple system to combine single-spot photolysis and whole-field fluorescence imaging,” J. Biomed. Opt., 18 (6), 060505 (2013). http://dx.doi.org/10.1117/1.JBO.18.6.060505 JBOPFO 1083-3668 Google Scholar

12. 

B. L. Sabatini, T. G. Oertner and K. Svoboda, “The life cycle of Ca(2+) ions in dendritic spines,” Neuron, 33 (3), 439 –452 (2002). http://dx.doi.org/10.1016/S0896-6273(02)00573-1 NERNET 0896-6273 Google Scholar

13. 

M. Canepari and D. Ogden, “Kinetic, pharmacological and activity-dependent separation of two Ca2+ signalling pathways mediated by type 1 metabotropic glutamate receptors in rat Purkinje neurones,” J. Physiol., 573 (1), 65 –82 (2006). http://dx.doi.org/10.1113/jphysiol.2005.103770 JPHYA7 0022-3751 Google Scholar

14. 

R. Schneggenburger and E. Neher, “Intracellular calcium dependence of transmitter release rates at a fast central synapse,” Nature, 406 (6798), 889 –893 (2000). http://dx.doi.org/10.1038/35022702 NATUAS 0028-0836 Google Scholar

15. 

G. C. Ellis-Davies and J. H. Kaplan, “Nitrophenyl-EGTA, a photolabile chelator that selectively binds Ca2+ with high affinity and releases it rapidly upon photolysis,” Proc. Natl. Acad. Sci. USA, 91 (1), 187 –191 (1994). http://dx.doi.org/10.1073/pnas.91.1.187 1091-6490 Google Scholar

16. 

P. H. Barry, “JPCalc, a software package for calculating liquid junction potential corrections in patch-clamp, intracellular, epithelial and bilayer measurements and for correcting junction potential measurements,” J. Neurosci. Methods, 51 (1), 107 –116 (1994). http://dx.doi.org/10.1016/0165-0270(94)90031-0 JNMEDT 0165-0270 Google Scholar

17. 

K. E. Vogt et al., “Combining membrane potential imaging with L-glutamate or GABA photorelease,” PLoS ONE, 6 (10), e24911 (2011). http://dx.doi.org/10.1371/journal.pone.0024911 1932-6203 Google Scholar

18. 

S. Gasparini and J. C. Magee, “Phosphorylation-dependent differences in the activation properties of distal and proximal dendritic Na+ channels in rat CA1 hippocampal neurons,” J. Physiol., 541 (3), 665 –672 (2002). http://dx.doi.org/10.1113/jphysiol.2002.020503 JPHYA7 0022-3751 Google Scholar

19. 

M. Canepari and F. Mammano, “Imaging neuronal calcium fluorescence at high spatio-temporal resolution,” J. Neurosci. Methods, 87 (1), 1 –11 (1999). http://dx.doi.org/10.1016/S0165-0270(98)00127-7 JNMEDT 0165-0270 Google Scholar

20. 

M. Maravall et al., “Estimating intracellular calcium concentrations and buffering without wavelength ratioing,” Biophys. J., 78 (5), 2655 –2667 (2000). http://dx.doi.org/10.1016/S0006-3495(00)76809-3 BIOJAU 0006-3495 Google Scholar

21. 

M. Canepari et al., “Photochemical and pharmacological evaluation of 7-nitroindolinyl-and 4-methoxy-7-nitroindolinyl-amino acids as novel, fast caged neurotransmitters,” J. Neurosci. Methods, 112 (1), 29 –42 (2001). http://dx.doi.org/10.1016/S0165-0270(01)00451-4 JNMEDT 0165-0270 Google Scholar

22. 

A. Savitzky and M. J. E. Golay, “Smoothing and differentiation of data by simplified least squares procedures,” Anal. Chem., 36 1627 –1639 (1964). http://dx.doi.org/10.1021/ac60214a047 ANCHAM 0003-2700 Google Scholar

23. 

C. C. Kuo and S. Yang, “Recovery from inactivation of T-type Ca2+ channels in rat thalamic neurons,” J. Neurosci., 21 1884 –1892 (2001). JNRSDS 0270-6474 Google Scholar

24. 

W. N. Ross et al., “A large change in dye absorption during the action potential,” Biophys. J., 14 (12), 983 –986 (1974). http://dx.doi.org/10.1016/S0006-3495(74)85963-1 BIOJAU 0006-3495 Google Scholar

25. 

S. Antic and D. Zecevic, “Optical signals from neurons with internally applied voltage-sensitive dyes,” J. Neurosci., 15 (2), 1392 –1405 (1995). JNRSDS 0270-6474 Google Scholar

26. 

M. Canepari et al., “Imaging inhibitory synaptic potentials using voltage sensitive dyes,” Biophys. J., 98 (9), 2032 –2040 (2010). http://dx.doi.org/10.1016/j.bpj.2010.01.024 BIOJAU 0006-3495 Google Scholar

27. 

B. Sakmann and E. Neher, “Patch clamp techniques for studying ionic channels in excitable membranes,” Annu. Rev. Physiol., 46 455 –472 (1984). http://dx.doi.org/10.1146/annurev.ph.46.030184.002323 ARPHAD 0066-4278 Google Scholar

28. 

S. R. Williams and S. J. Mitchell, “Direct measurement of somatic voltage clamp errors in central neurons,” Nat. Neurosci., 11 (7), 790 –798 (2008). http://dx.doi.org/10.1038/nn.2137 NANEFN 1097-6256 Google Scholar

29. 

G. J. Stuart and N. Spruston, “Probing dendritic function with patch pipettes,” Curr. Opin. Neurobiol., 5 (3), 389 –394 (1995). http://dx.doi.org/10.1016/0959-4388(95)80053-0 COPUEN 0959-4388 Google Scholar

30. 

R. Davies, J. Graham and M. Canepari, “Light sources and cameras for standard in vitro membrane potential and high-speed ion imaging,” J. Microsc., 251 (1), 5 –13 (2013). http://dx.doi.org/10.1111/jmi.2013.251.issue-1 JMICAR 0022-2720 Google Scholar

31. 

W. L. Zhou et al., “Intracellular long-wavelength voltage-sensitive dyes for studying the dynamics of action potentials in axons and thin dendrites,” J. Neurosci. Methods, 164 (2), 225 –239 (2007). http://dx.doi.org/10.1016/j.jneumeth.2007.05.002 JNMEDT 0165-0270 Google Scholar

32. 

P. Yan et al., “Palette of fluorinated voltage-sensitive hemicyanine dyes,” Proc. Natl. Acad. Sci. USA, 109 (50), 20443 –20448 (2012). http://dx.doi.org/10.1073/pnas.1214850109 1091-6490 Google Scholar

Biography

Nadia Jaafari obtained her PhD from the University of Marseille and worked for several years in the laboratory of J.M. Henley at Medical Research Council Centre for Synaptic Plasticity in Bristol. She is currently an INSERM senior research associate in the laboratory of Marco Canepari.

Elodie Marret graduated from the University Joseph Fourier in Grenoble, and she is currently an INSERM technician in the laboratory of Marco Canepari.

Marco Canepari is an INSERM first-class researcher (CR1) working in the Laboratoire Interdisciplinare de Physique in Grenoble. He obtained his PhD at SISSA/ISAS in Trieste, and he worked in several laboratories as a postdoctoral scientist, including the laboratory of Dejan Zecevic at Yale University School of Medicine.

© 2015 Society of Photo-Optical Instrumentation Engineers (SPIE) 2329-423X/2015/$25.00 © 2015 SPIE
Nadia Jaafari, Elodie Marret, and Marco Canepari "Using simultaneous voltage and calcium imaging to study fast Ca2+ channels," Neurophotonics 2(2), 021010 (11 February 2015). https://doi.org/10.1117/1.NPh.2.2.021010
Published: 11 February 2015
Lens.org Logo
CITATIONS
Cited by 13 scholarly publications.
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Calcium

Action potentials

Luminescence

Calibration

Electronic filtering

Neurons

Optical testing

Back to Top