Journal of Neurophysiology

Chloride-Sensitive MEQ Fluorescence in Chick Embryo Motoneurons Following Manipulations of Chloride and During Spontaneous Network Activity

Nikolai Chub, George Z. Mentis, Michael J. O'Donovan


Intracellular Cl ([Cl]in) homeostasis is thought to be an important regulator of spontaneous activity in the spinal cord of the chick embryo. We investigated this idea by visualizing the variations of [Cl]in in motoneurons retrogradely labeled with the Cl-sensitive dye 6-methoxy-N-ethylquinolinium iodide (MEQ) applied to cut muscle nerves in the isolated E10–E12 spinal cord. This labeling procedure obviated the need for synthesizing the reduced, cell-permeable dihydro-MEQ (DiH-MEQ). The specificity of motoneuron labeling was confirmed using retrograde co-labeling with Texas Red Dextran and immunocytochemistry for choline acetyltransferase (ChAT). In MEQ-labeled motoneurons, the GABAA receptor agonist isoguvacine (100 μM) increased somatic and dendritic fluorescence by 7.4 and 16.7%, respectively. The time course of this fluorescence change mirrored that of the depolarization recorded from the axons of the labeled motoneurons. Blockade of the inward Na+/K/2Cl co-transporter (NKCC1) with bumetanide (20 μM) or with a low-Na+ bath solution (12 mM), increased MEQ fluorescence by 5.3 and 11.4%, respectively, consistent with a decrease of [Cl]in. After spontaneous episodes of activity, MEQ fluorescence increased and then declined to the pre-episode level during the interepisode interval. The largest fluorescence changes occurred over motoneuron dendrites (19.7%) with significantly smaller changes (5.2%) over somata. Collectively, these results show that retrogradely loaded MEQ can be used to detect [Cl]in in motoneurons, that the bumetanide-sensitive NKCC1 co-transporter is at least partially responsible for the elevated [Cl]in of developing motoneurons, and that dendritic [Cl]in decreases during spontaneous episodes and recovers during the inter-episode interval, presumably due to the action of NKCC1.


Spontaneous neural activity is a characteristic of developing networks in several different vertebrate animals (for reviews, see Ben-Ari 2001, 2002; O'Donovan 1999) and is believed to play an important role in neural and network development (Casavant et al. 2004; Hanson and Landmesser 2004; Milner and Landmesser 1999). In the spinal cord of E10–E12 chick embryos, spontaneous activity occurs as recurring rhythmic episodes ∼1 min in duration that are followed by longer (≤10–15 min) inter-episode intervals (Landmesser and O'Donovan 1984). During these rhythmic episodes, flexor and extensor motoneurons alternate their discharges in a manner resembling locomotion in the mature animal (Landmesser and O'Donovan 1984). At E10–E12, chloride-mediated GABAergic transmission is an important component of the rhythmic synaptic drive to ventral horn neurons during spontaneous episodes (Sernagor et al. 1995). In the chick embryo (Chub and O'Donovan 2001) as in other developing vertebrates (Gao and Ziskind-Conhaim 1995; Rohrbough and Spitzer 1996; Serafini et al. 1995), GABA depolarizes spinal cord neurons and is functionally excitatory. This action is the result of high intracellular chloride concentration ([Cl]in), which creates an outward driving force for Cl ions during the activation of GABAA channels. In the hippocampus, depolarizing GABAA responses become hyperpolarizing later in development (Ben-Ari et al. 1989; Khazipov et al. 2004), and this process itself appears to require activation of GABAergic transmission (Ganguly et al. 2001). In spontaneously active developing spinal networks, GABAergic depolarizations periodically trigger intracellular Ca2+ elevations (Lev-Tov and O'Donovan 1995; O'Donovan et al. 1994; Wenner and O'Donovan 2001), which are believed to be important for structural development of neurons and the maturation of network connections (Ben-Ari 2002; Demarque et al. 2002; Sernagor et al. 2003). Therefore the short- and long-term regulation of [Cl]in during development is of considerable interest.

We have shown previously that [Cl]in is reduced in ventral spinal neurons after an episode and is restored before the next episode (Chub and O'Donovan 2001). Because these measurements were made using whole cell electrodes, the [Cl]in would have been modified by the intracellular pipette solution. Furthermore, they could only be made infrequently because they required computing the current-voltage relation for locally applied GABAA agonists. Therefore in the present work, we have employed a noninvasive optical method to monitor the [Cl]in changes in spinal motoneurons during spontaneously occurring episodes of activity. For this purpose, we developed a new technique for loading neurons with the Cl-sensitive 6-methoxy-N-ethylquinolinium iodide (MEQ) dye. Traditionally, the use of MEQ requires synthesis of the cell-permeable form of the dye—dihydro-MEQ (DiH-MEQ)—which is then bath-applied to label neurons (Biwersi and Verkman 1991). In the present work, we retrogradely loaded MEQ into motoneurons by applying the dye to the cut ends of muscles nerves. This method has the advantages of labeling a specific cell population and not requiring an initial reduction reaction to render the dye cell-permeant.

We have used the technique to resolve [Cl]in changes in motoneurons in response to bath application of the GABAA receptor agonist isoguvacine, bumetanide a Na+/K/2Cl co-transporter (NKCC1) inhibitor, and to investigate the changes of intracellular chloride during spontaneously occurring episodes of activity.


Fertile White Leghorn chicken eggs from a local supplier were incubated at 37.5–37.7°C in a humidified incubator with circulated air and an egg turner (Sportsman 1502, G.Q.F. Manufacturing). On the day of the experiments, E10–E12 chick embryos were removed from the egg and quickly placed in the dissection chamber with continuously re-circulating and cooled to 12–15°C Tyrode's solution, which contained (in mM) 139 NaCl, 5 KCl, 17 NaHCO3, 1 MgCl2, 3 CaCl2, and 12 glucose; saturated with a mixture of 95% O2-5% CO2 to pH 7.4–7.5. The dissection technique was similar that used in earlier studies (Chub and O'Donovan 2001; Landmesser and O'Donovan 1984) and therefore will be described only briefly. The embryo was pinned to the silicone-elastomer Sylgard-covered bottom of the dissection chamber. After the embryo was eviscerated, and a ventral laminectomy was performed. A portion of the thoracic and the lumbosacral part of the spinal cord (T4–LS8) with the sartorius (SART) and femorotibialis (FEM) motor nerves was dissected from the vertebral column and hind limb muscles. All dorsal roots were cut to eliminate labeling of afferent terminals. The dura and pia mater were removed close to the lateral edge of the spinal cord to improve visualization of the labeled motoneurons.

In the initial experiments, we used bath-loading of the cell-permeable chloride-sensitive MEQ dye (Biwersi and Verkman 1991). MEQ is a largely cell-impermeant organic compound, which exhibits a maximum emission at 420 nm in response to illumination at 360 nm. This fluorescence is collisionally quenched in a concentration-dependent manner by Cl without shifting the emission spectra. It is known that other halide anions, such as I, Br, and SCN (Biwersi and Verkman 1991) can also quench the MEQ fluorescence. For bath-application, the dye must be initially reduced to a lipophilic Cl-insensitive form DiH-MEQ and allowed to enter cells where it can be re-oxidized to the Cl-sensitive MEQ (Biwersi and Verkman 1991). In our initial experiments, we found that the DiH-MEQ did not penetrate deeply into the cord and was only slowly re-oxidized to Cl-sensitive MEQ dye once inside cells. For example, in E10–E11 preparations, 3–4 h after bath application of the DiH-MEQ (50 μM for 60 min) some cells still exhibited green fluorescence indicating incomplete intracellular DiH-MEQ re-oxidation.

For these reasons, we developed another labeling technique in which motoneurons were retrogradely loaded with the Cl-sensitive MEQ by applying the dye to the cut end of motor nerves. This technique is similar to retrogradely labeling neurons with calcium-sensitive dyes (O'Donovan et al. 1993). For this type of labeling, the spinal cord was isolated together with attached muscle nerves. In most experiments, both the SART and FEM nerves were pulled into the tip of a plastic pipette by applying gentle mouth suction to the end of the pipette. Another smaller diameter (insertion) plastic tube, connected to a syringe, was used to deliver the MEQ dye (50–100 mM solution in Dulbecco's phosphate-buffered saline with calcium and magnesium, DPBS, Mediatech) into the pipette holding the motor nerves. After 2–4 h of dye application, the MEQ solution was extracted with the insertion tube and the motor nerves were removed from the suction pipette by applying weak positive pressure. To verify that MEQ specifically labeled motoneurons, in six experiments, we loaded the muscles nerves with MEQ and the retrograde tracer Texas Red Dextran (40–50 mM in DPBS; 10,000 MW; Invitrogen). We used Texas Red Dextran because this dye does not cross gap junctions or synapses and has been widely used to retrogradely label motoneurons (Bonnot et al. 2005; Mentis et al. 2005). During the 8–10 h required for dye loading, the spinal cord was continuously superfused with re-circulating at 21–23°C Tyrode's solution.

The preparation was then transferred to a recording chamber mounted on the stage of an inverted microscope (Nikon Diaphot 300) equipped for epi-fluorescence. For MEQ fluorescence measurements, we used illumination from a 100-W xenon-tungsten lamp, and a fluorescence filter combination UV-2E/C (Nikon) with 20× magnification using a Fluor 0.75 (160/0.17) objective (Nikon). An intensified CCD (XR GEN III+, Stanford Photonics Inc) or a digital (CoolSNAP ES, Photometrics) camera was used for capturing images. The duration of image acquisition was set by an electronic shutter controlled by MetaMorph software (Molecular Devices). This software was also used for capturing images and off-line analysis. We calculated the fractional change in fluorescence (ΔF/F) following various experimental manipulations by first measuring the mean fluorescence within a region of interest (ROI) over the motoneurons. In all cases, unless otherwise mentioned, mean background fluorescence (measured from an unlabeled area) was subtracted from this measurement and the resulting time series was divided by the initial or the predrug fluorescence. The percentage changes of this normalized fluorescence are reported.

We found that MEQ fluorescence decayed over time in the absence of illumination as reported previously and attributed to a gradual leakage of dye from the cells (Biwersi and Verkman 1991). This rate of decay was measured and subtracted from records shown in Figs. 58. Fluorescence decay for long time periods (7 h) was exponentially fitted with SigmaPlot (SPSS), and the resulting time constants are reported.

The electrical activity from motor nerves was recorded together with the optical activity and synchronized by recording the opening and closing of the electronic shutter. Recordings from SART or FEM nerves were made using plastic suction electrodes connected to differential amplifier (AI402 SmartProbe, Axon Instruments). Signals were filtered (DC –2 kHz or 0.1–2 kHz, Cyber Amp 380), digitized (DigiData 1322A) and analyzed with pCLAMP 9 software (Axon Instruments). Rhythmic motor activity occurred spontaneously in Tyrode's solution warmed to 27–29°C.

In some experiments, we used a “low-Na+” bath solution, which comprised (in mM) 144 Tris-Cl, 25 Tris-OH, 5 KHCO3, 12 NaHCO3, 1 MgCl2, 3 CaCl2, and 12 glucose and was saturated with 95% O2-5% CO2; pH 7.3. This solution has the same [Cl] as control Tyrode's solution, but higher osmolarity. Therefore for these experiments, 40 mM sucrose was added to the control Tyrode's solution to minimize the osmolarity difference between control and “low-Na+” solutions.

After the recordings, some of the spinal cords were fixed by immersion in 4% paraformaldehyde (in PBS) for 4 h at 4°C. The cords were subsequently embedded in warm 5% agar (in PBS) and 50-μm sections cut with a Vibratome. The sections were collected in wells and processed for either morphological analysis or immunohistochemistry. To further confirm motoneuron identity, we used a polyclonal antibody against choline acetyltransferase (ChAT) raised in goat (1:100, Chemicon). To maximize the binding sites of the antibodies, the immunohistochemistry was performed in free floating sections. Initially, the sections were washed once for 10 min in PBS, followed by three washes in PBS-T (0.1% Triton in PBS). Nonspecific binding of the secondary antibodies was minimized by block with Normal Donkey Serum (1:10 in PBS-T). The primary antibodies (diluted in PBS-T) were incubated for 8–12 h at room temperature. The primary antibodies were subsequently washed six times (10 min each) in PBS-T. Immunoreactive sites were revealed by the binding of secondary antibodies conjugated to a specific fluorochrome (Jackson Immuno Research Laboratories). ChAT was visualized by anti-Goat-Cy5. The secondary antibodies were incubated for 3 h at room temperature and were subsequently washed six times (10 min each) in PBS. Finally, the sections were mounted on glass slides and coverslipped.

Images were obtained with confocal microscopy using a four-channel two-photon (Chameleon, Coherent) 510META (Carl Zeiss, Germany) confocal microscope equipped with three single photon lasers (488, 543, and 633 nm). Images from sections containing three fluorochromes (MEQ, Texas-Red Dextran, and ChAT) were acquired by multi-tracking. The MEQ dye was excited with the two-photon laser-locked mode at 720 nm and images acquired with a 390–465 nm filter. Texas Red Dextran was excited with a 543 nm laser and acquired with a 565–615 nm filter. ChAT immunoreactivity was excited with a 633 nm laser and acquired with a 650–710 nm filter.

Chemicals were purchased from Sigma, unless otherwise mentioned. Drugs were prepared as a 10 mM stock solution in distilled H2O and were applied to the continuously recirculating Tyrode's solution. All data are expressed as means ± SE. Statistical analyses were performed by t-test, and a significant difference was assumed when P < 0.05.


Six spinal cords were retrogradely co-labeled with MEQ and Texas Red Dextran (Figs. 1 and 2) to confirm that MEQ labels motoneurons only. In all the experiments, the majority of Texas-Red-Dextran-positive neurons were also co-labeled with MEQ dye although there were a few instances where some Texas Red Dextran motoneurons were not MEQ loaded and vice versa. Motoneurons were localized in the ventrolateral region of the ventral horn consistent with the known location of chick lumbar motoneurons. No cells were observed outside the motoneuron nucleus, nor were any dorsal root afferents labeled (Fig. 1D). Confocal images revealed that the MEQ dye labeled motoneuron dendrites and axons as well as the cell bodies (Fig. 1E). As illustrated in Fig. 2, which shows motoneurons labeled for choline acetytransferase (ChAT), a majority of motoneurons could be co-labeled with MEQ and Texas Red Dextran.

FIG. 1.

Retrograde labeling of motoneurons with the chloride-sensitive 6-methoxy-N-ethylquinolinium iodide (MEQ) dye and co-labeling with 2 dyes: MEQ and the retrograde neuronal tracer Texas Red Dextran. A: bright-field image showing the isolated spinal cord of an E11 chick embryo at low magnification (2× objective). The preparation was placed ventral side down on the glass bottom of a superfusion chamber mounted on the stage of an inverted microscope. The T7 to LS3 segments are shown. B: epi-fluorescence image (10× objective) of MEQ labeled motoneurons located in LS2–LS3 segments. C, 13: confocal images from a different preparation of a single optical plane from a spinal cord in which motoneurons were loaded with MEQ (C1; shown in blue) and Texas Red Dextran (C2; shown in red) in the LS1 segment. Overlay of the 2 fluorochromes is shown in C3. The somata of motoneurons and their axons (arrows) are clearly co-labeled with both dyes. D: a transverse section of the spinal cord from a different preparation showing LS2 motoneurons back labeled with MEQ. The area indicated by the dotted rectangle is shown at higher magnification in E. E: projection image from a z-stack of optical planes (10 optical planes, 2-μm optical thickness/plane; z-axis distance: 20 μm) showing a cluster of motoneuron somata labeled with MEQ. Note the presence of MEQ fluorescence in medially and dorsally projecting dendrites and in motoneuron axons exiting the cord (arrows). The motoneurons were retrogradely labeled through a suction pipette filled with the MEQ dye (B, D, and E) or with a mixture of MEQ and Texas Red Dextran (C, 1–3).

FIG. 2.

Motoneurons retrogradely co-loaded with MEQ and Texas Red Dextran exhibit choline acetyltransferase (ChAT) immunoreactivity. A, 1–4: transverse section of the ventral horn of an E10 chick spinal cord showing the dyes MEQ (A1) and Texas Red Dextran (A2) in retrogradely loaded motoneurons. The majority of motoneurons were co-labeled with both dyes. A3: ChAT immunoreactivity reveals that most motoneurons have been retrogradely filled with MEQ and/or Texas Red Dextran. A4: overlay of all 3 fluorochromes. B, 1–3: higher-magnification images of single motoneurons (single optical planes) showing co-localization of 3 fluorochromes. Note that MEQ, Texas Red Dextran, and ChAT immunoreactivity were seen in the dendrites as well as the somata of motoneurons.

We found that the motoneuronal MEQ fluorescence progressively declined over several hours. To quantify this change, we measured the fluorescence every 5 min using brief illumination (600 ms) to minimize dye bleaching. Figure 3 shows an example of the decay of the MEQ fluorescence measured over the somata of motoneurons during the 7-h recording time period. In this experiment, the decay was well fitted with a single exponent (Fig. 3B). Fluorescence decay during shorter time periods (∼1 h) could be linearly approximated (Fig. 3B, inset). The average time constant of fluorescence decay measured over a 7-h period in four experiments was 0.31 + 0.04 h (n = 6, 4 experiments). The decay rates measured from linear fit for short time intervals (∼1 h) were 14.5 ± 1.4% h−1 for the first 3 h and 5.1 ± 0.6% h−1 for the next 4 h (n = 6, 4 experiments; P < 0.05). We attribute this fluorescence decline to a gradual leakage of the MEQ dye from the labeled cells (Biwersi and Verkman 1991).

FIG. 3.

Time-dependent decay of the MEQ fluorescence in labeled motoneurons. A: image (20 video frames averaged) of the MEQ fluorescence of retrogradely loaded LS2 motoneurons in the cord of an E12 chick embryo. The location of the motoneuron pool (viewed from the ventral aspect) is marked by 2 dotted lines. B: decay of the neuronal MEQ fluorescence during a 7-h period. Fluorescence was measured over the solid rectangle shown in A. To reduce possible photo-bleaching, the MEQ-labeled motoneurons were only briefly illuminated (for 600 ms), and images were taken once every 5 min. Note that the 7-h fluorescence decay was exponential. Gray line is a mono-exponential fit with time constant of 0.36 h. Inset: shorter (1 h) time periods (segments a and b) could be fitted by linear regression (gray lines). Calibration is 20 min and 20% ΔF/F.

To establish the sensitivity of MEQ fluorescence to photobleaching, we continuously illuminated the labeled motoneurons for 200–300 s. The intensity of the MEQ fluorescence was measured (every 300th frame) in an ROI located over the labeled motoneurons and plotted against time. Figure 4 shows an example of the decline in fluorescence, measured from the somata of two labeled motoneurons, during 200 s of continuous illumination. The mean rate of photo-bleaching, was relatively small 0.031 ± 0.01% s−1 (n = 6; 4 experiments). This rate was significantly lower than reported in earlier experiments using laser illumination (Inglefield and Schwartz-Bloom 1997). The linear regression fit (Fig. 4B) was used to estimate the decline of the fluorescence intensity under continuous illumination.

FIG. 4.

Limited photobleaching of the MEQ fluorescence during continuous illumination. A: image from MEQ-labeled motoneurons, located in the LS2 spinal segment. The location of the LS2 motoneuron pool and the LS2 ventral root is marked by dotted lines. B: relative intensity of the neuronal MEQ fluorescence during continuous 200-s illumination of the spinal cord. Each data point is the mean fluorescence intensity measured from the somata of 2 labeled motoneurons (cells a and b on A, and marked by circles and triangles on the graph). MEQ fluorescence was measured from every 300th image during the 200-s illumination period. The fluorescence decay curves for the 2 cells were linearly fitted (solid gray lines). Note that, measurements were made on single video frames the signal is noisier (in A and B) than the averaged data shown in Fig. 3.

In the next set of experiments, we established if changes in the fluorescence of the MEQ labeled motoneurons would accompany manipulations designed to alter [Cl]in. At this stage of development (E10-12), GABAA currents are depolarizing with reversal potentials around −30 mV; ∼20 mV higher than the resting membrane potential of ventral horn neurons (Chub and O'Donovan 2001). Therefore activation of GABA-A, Cl-conductances should result in an outwardly directed Cl flux and a reduction of [Cl]in. We found that bath-application of 100 μM GABAA agonist isoguvacine for 10 min, in the presence of 0.5 μM tetrodotoxin (TTX) to block presynaptic effects, led to an increase of MEQ fluorescence and a depolarization of the MEQ labeled motoneurons that was recorded from the muscle nerve (Fig. 5). Because MEQ dye is quenched collisionally by neuronal Cl ions, an increase of the dye fluorescence corresponds to an reduction of [Cl]in (Jayaraman and Verkman 2000). We also found that the isoguvacine-induced changes of [Cl]in differed between the soma and dendrites of the labeled motoneurons. The largest percentage changes of fluorescence were recorded over the dendrites 16.7 ± 1.5%, and smaller changes were detected over the somata 7.4 ± 0.5% (n = 32, 6 experiments; P < 0.05). The time course of the fluorescence transient recorded from labeled motoneurons corresponded closely with the depolarization electrotonically recorded from the SART nerve (Fig. 5B, bottom). The slow potential recorded electrotonically from the muscle nerve has been shown to have the same time course as the intracellularly recorded membrane potential of individual motoneurons (O'Donovan 1989). Because Cl is the dominant ionic current activated by isoguvacine, the time course of the slow potential will represent the time course of the Cl efflux. The mean amplitude of the depolarization simultaneously recorded from the SART nerve was 0.94 ± 0.12 mV (n = 9, 4 experiments).

FIG. 5.

Bath-application of the GABAA receptor agonist isoguvacine depolarized and increased the fluorescence of MEQ-labeled motoneurons. A: digital image (time exposure: 600 ms) of MEQ-labeled motoneurons located in the LS1 spinal segment. B: simultaneous optical and electrical responses to isoguvacine application (100 μM, 10 min) were measured over 3 somatic (depicted in A by dotted lines with letters a, b, c) areas, 3 dendritic (d, e, f) areas and from the SART nerve (bottom). Action potentials were blocked with TTX (0.5 μM in bath). Note the largest fractional changes in fluorescence occur over the motoneuron dendrites. This fluorescence increase indicates outflow of Cl during GABAA receptors activation.

We have hypothesized that the elevated [Cl]in in ventral spinal neurons results from the action of an inwardly directed Cl co-transporter (Chub and O'Donovan 2001; Marchetti et al. 2005). Consistent with this idea, bumetanide, an antagonist of the inward chloride transporter NKCC1 (Haas 1994; Payne et al. 2003; Plotkin et al. 1997; Sun and Murali 1999) can reduce the frequency of spontaneous episodes in the isolated spinal cord of the chick embryo (Marchetti et al. 2005). To test if such a transporter was active in motoneurons, we bath-applied bumetanide and measured the corresponding changes in MEQ fluorescence. Figure 6 shows the effect of 60-min bath application of 20 μM bumetanide. We found that bumetanide increased the neuronal MEQ fluorescence by 5.3 ± 0.8% (n = 11, 4 experiments).

FIG. 6.

Bath application of bumetanide, a blocker of the Na+/K+/2Cl co-transporter (NKCC1), increased the fluorescence of MEQ-labeled motoneurons. A: digital image (time exposure: 600 ms) of MEQ-labeled motoneurons located in the LS1 spinal segment. B: 60-min bath-application of 20 μM bumetanide increased neuronal MEQ fluorescence, consistent with a decrease of [Cl]in. Fluorescence was measured over 3 areas (a, b, c) delineated by the dotted lines in A.

Additional evidence for the participation of the NKCC1 co-transporter in neuronal Cl accumulation was obtained by examining the fluorescence changes to a low-Na+ extracellular solution (see methods). In contrast to other known neuronal Cl transport pathways (Mount et al. 1998; Payne et al. 2003), only NKCC1-dependent Cl accumulation requires an inwardly directed Na+ gradient. Therefore low extracellular [Na+] as well as bumetanide can inhibit NKCC1-mediated Cl in-flux (Russell 2000). Figure 7 shows effect of 40-min bath application of 12 mM Na+ solution on motoneuronal MEQ fluorescence. In these experiments, application of the low Na+ solution increased MEQ fluorescence by 11.4 ± 0.9% (n = 11, 3 experiments), consistent with the participation of NKCC1 in inwardly directed neuronal Cl transport.

FIG. 7.

MEQ fluorescence is increased in “low-Na+” bath solution. A: image (20 video frame average) of MEQ-labeled LS3 motoneurons. The location of the motoneuron pool is marked by 2 dotted lines. B: fluorescence was measured over the soma of the 3 motoneurons (marked by arrowhead with letters a, b, c). The time of the low extra-cellular Na+ (12 mM) solution application is marked by a thick gray bar above the curves.

In the next set of experiments, we investigated the changes of MEQ fluorescence accompanying spontaneously occurring episodes. For this purpose, the electrical activity from FEM or SART motor nerves was simultaneously recorded with the MEQ fluorescence. Figure 8, A and B, shows a typical example of the fluorescence changes (measured every 2 min) over three different areas of the labeled motoneurons: axons in the ventral root, motoneuron somata, and dendrites. Spontaneous rhythmic episodes 40–60 s in duration occurred every 8–12 min (Fig. 8B, top). We found that the MEQ fluorescence increased after a spontaneous episode then slowly declined to the pre-episode level during the interepisodes interval consistent with a decrease of [Cl]in after an episode and its recovery during inter-episode intervals. Another, longer (80 min) example of the fluorescence changes measured repetitively (every 1 min) in the areas of the motoneuron dendrites and somata in the spontaneously active spinal cord is shown in Fig. 8C.

FIG. 8.

Variations of motoneuronal MEQ fluorescence during spontaneously occurring episodes are largest over the region of motoneuron dendrites. A: grayscale image (top) and pseudo-colored image (bottom) of the MEQ-labeled motoneurons located in LS3–LS2 spinal segments. Three colored regions of interest (ROI) are placed over the area of motoneuron dendrites (red crescent), motoneuron somata (blue oval), and motoneuron axons in the ventral root (green rectangle). The ventral roots and the motoneuron pool are delineated by dotted lines. B: electrical activity (AC record: 10 Hz to 3 kHz) from the FEM nerve is displayed simultaneously with changes of neuronal MEQ fluorescence. The 1st spontaneous episode (about 50 s in duration) is marked by an arrow. Measurements were made on images acquired every 2 min (marked by the blue vertical lines superimposed on the FEM nerve record). The changes of MEQ fluorescence over the 3 ROIs (A, bottom) are shown separately (same color as ROI). The time of occurrence of the episodes is identified by the pink bars. C: longer record (80 min) of the spontaneous electrical activity and MEQ fluorescence. Electrical activity from SART nerve is displayed simultaneously with changes of MEQ fluorescence (as in B). Fluorescence was measured from dendrites (red line) and somata (blue line). Bottom: control measurement of the background fluorescence from a nonneuronal area (black line). Note that while the somatic fluorescence changes are smaller than over the dendrites, they are clearly distinguished from the background fluorescence signal.

For these experiments, percentage changes of the MEQ fluorescence (ΔF/F) were calculated as the mean motoneuronal fluorescence after the episode minus the fluorescence before the episode and divided by the pre-episode fluorescence. The largest post-episode fluorescence increase was observed over the motoneuron dendrites 19.7 ± 1.8% and smaller changes were recorded over motoneuron somata 5.2 ± 0.6% (n = 49, 3 experiments; P < 0.05). No fluorescence changes were detected over motoneuron axons.


The results of this paper show that motoneurons in the isolated spinal cord of the chick embryo can be retrogradely labeled with the membrane-impermeant Cl-sensitive MEQ dye. This obviates the need to convert MEQ into the Cl-insensitive, cell-permeant DiH-MEQ and its subsequent intracellular conversion to MEQ (Biwersi and Verkman 1991). In addition, we have demonstrated that the elevated [Cl]in of developing chick motoneurons described previously (Chub and O'Donovan 2001) is controlled at least in part by the action of a bumetanide-sensitive NKCC1 co-transporter. Finally, we have established that motoneuron dendrites, rather than somata, are the major locus for changes of intracellular chloride during spontaneous episodes of network activity.

The rates of neuronal MEQ fluorescence decay and photobleaching in our experiments were similar to those employing bath-loading of DiH-MEQ in fibroblast cultures (Biwersi and Verkman 1991) and in hippocampal (Sah and Schwartz-Bloom 1999) or brain slice preparations (Fukuda et al. 1998; Inglefield et al. 1999; Shibata et al. 2004). In addition, the fluorescence of MEQ-labeled motoneurons was increased by bath-application of the GABAA agonist isoguvacine consistent with Cl outflow. These findings indicate that intracellular MEQ behaves similarly whether loaded retrogradely or by bath-application. Of course, the advantage of the retrograde method is that a specific population of neurons can be loaded allowing the cellular origin of the Cl signals to be defined unambiguously.

We used two methods to demonstrate that MEQ labeling was restricted to motoneurons. First, we identified motoneurons by retrogradely labeling them with Texas Red Dextran. Because of the high molecular weight of this dye (10,000), it does not leak through gap junctions and remains localized within motoneurons. Motoneurons were also identified by their ChAT immunofluorescence. When these procedures were used in conjunction with MEQ labeling, we found that only motoneurons were labeled by the Cl-sensitive dye.

In our previous work, we used whole cell recording to show that the [Cl]in falls after an episode and then is restored during the interepisode interval (Chub and O'Donovan 2001). We hypothesized that the restoration of [Cl]in was due to the action of an inwardly directed NKCC1 co-transporter. The results of the present work provide functional evidence for the presence of NKCC1 on chick embryo motoneurons at E10–E12. Specifically, the fluorescence of MEQ-loaded motoneurons was increased by the NKCC1 blocker bumetanide and by a low [Na+] in the extracellular solution. Both of these procedures effectively decreased the [Cl]in. The presence of this transporter on developing chick motoneurons is in agreement with other studies on immature neurons (Plotkin et al. 1997; Rohrbough and Spitzer 1996; Sun and Murali 1999) where electro-neutral NKCC1 co-transport was demonstrated. Our data do not exclude the possibility that other chloride ion transporters, for example, the Cl/HCO3 exchanger (Kaila 1994; Staley and Proctor 1999), may also be active at this stage of development in chick embryo motoneurons.

Perhaps the most significant finding in the present work was the observation that periodic variations in the [Cl]in accompany the discharge of spontaneously active motoneurons. We found that [Cl]in falls after an episode and slowly recovers during the interepisode interval to the pre-episode level. These activity-dependent variations of MEQ fluorescence were consistent with our recent data indicating that [Cl]in decreased ≤15 mM after an episode (Chub and O'Donovan 2001). These earlier data were obtained with whole cell electrodes and were therefore complicated by dialysis of the [Cl]in by the pipette solution. The fact that we were able to detect significant post-episode changes using whole cell recording raised the possibility that the most significant [Cl]in changes were occurring in the dendrites. Consistent with this hypothesis we found that the largest changes of MEQ fluorescence during spontaneous episodes were over motoneuron dendrites. This observation may indicate that GABAergic synapses, which are active during spontaneous episodes, are not uniformly distributed on the motoneuron surface. Alternatively, it may indicate that dendrites are more sensitive to the ion redistribution because of their smaller cytoplasmic volume in comparison with that of the soma. Finally, it is possible that the density of Cl transporters might be different on the dendrites and on the soma, which could be investigated using immunocytochemistry. Additional progress in understanding the spatial distribution of [Cl]in requires the use of more sensitive chloride dyes and more sophisticated imaging such as that afforded by two-photon confocal microscopy (Marandi et al. 2002). In comparison with patch-clamp recording, Cl-dependent changes of MEQ fluorescence are relatively small. As a result, some important neural network events like local [Cl]in changes after quantal GABA release (e.g., spontaneous miniature postsynaptic currents) remain beyond detection of the MEQ dye.

In conclusion, our data show that the neuronal Cl outflow during an episode exceeds the inwardly directed neuronal Cl transport capacity. Therefore after spontaneous episodes, [Cl]in is reduced and then slowly restored during the interepisode interval. Recently we have shown that bumetanide can block spontaneous episodes in the isolated chick spinal cord following antagonism of glutamatergic transmission (Marchetti et al. 2005). In addition, the rate of spontaneous activity is greatly slowed following blockade of GABAA receptor activity (Chub and O'Donovan 1998; Marchetti et al. 2005). These findings, together with the results of the present work, are consistent with the idea that the bumetanide-sensitive NKCC1 co-transporter is responsible, in part, for restoring the [Cl]in during the inter-episode interval. This activity-dependent variation of [Cl]in leads to corresponding changes in the chloride equilibrium potential and the driving force for GABAergic synaptic potentials. A recent modeling study (Marchetti et al. 2005) has shown that these variations in excitability can themselves lead to the expression of spontaneous episodes of activity. Thus the action of GABAergic Cl conductances and the slower inward pumping of Cl by NKCC1 may be an essential mechanism underlying the genesis of spontaneous activity in the developing spinal cord.


This research was supported by the Intramural Research Program of the National Institute of Neurological Disorders and Stroke.


The authors thank C. VanDunk for help with histology and immunohistochemistry.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract