Abstract
Plateau potentials are prolonged membrane depolarizations that are observed in hippocampal pyramidal neurons when spiking and Ca2+ entry occur in combination with muscarinic receptor activation. In this study, we used whole-cell voltage clamping to study the current underlying the plateau potential and to determine the cellular signaling pathways contributing to this current. When combined with muscarinic stimulation, depolarizing command potentials that evoked Ca2+ influx elicited a prolonged tail current (Itail) that had an extrapolated reversal potential of −20 mV.Itail was not observed when intracellular Ca2+ levels were chelated with 10 mmintracellular BAPTA, and Itail was reversibly depressed in low external sodium. WhenItail was evoked at intervals >3 min, current amplitudes were stable for up to 1 hr. However, at shorter intervals, Itail was refractory, with a time constant of recovery of 43.5 sec. The inhibitors of soluble guanylate cyclase 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one and 6-anilino-5,8-quinolinequinone depressedItail and zaprinast, which blocks cGMP-specific phosphodiesterase, enhancedItail, suggesting that a component ofItail was activated by cGMP. The inhibitors of cyclic nucleotide-gated (CNG) channelsl-cis-diltiazem and 2′,4′-dichlorobenzamil reversibly depressed Itail. However, protein kinase G inhibition had no effect. Therefore, these results indicate that a component of Itail is attributable to activation of CNG channels. We conclude that Ca2+ influx when combined with muscarinic receptor activation activates soluble guanylate cyclase and increases cGMP levels. The increased cGMP activates CNG channels and leads to prolonged depolarization. The cation conductance of the CNG channel contributes to the prolonged depolarization of the plateau potential.
We first reported that Ca2+ influx in combination with muscarinic (m1/m3) or metabotropic glutamate receptor (mGluR) stimulation generates prolonged depolarizations called plateau potentials (PP) in hippocampal pyramidal neurons (Fraser and MacVicar, 1996). We hypothesized that the PP was generated by a Ca2+-activated cation conductance (Crepel et al., 1994; Congar et al., 1997) because chelating intracellular Ca2+ prevented the Na+-mediated depolarization. Other reports have shown that plateau potentials occur in pyramidal neurons of other cortical regions (Klink and Alonso, 1997; Kawasaki et al., 1999). The PP is an attractive candidate for a major intrinsic conductance generating the prolonged depolarization observed during ictal phase of seizures (Dichter and Ayala, 1987; Fraser and MacVicar, 1996). Indirectly supporting this postulate is the observation that the PP is depressed by anticonvulsants, such as topiramate (Palmieri et al., 2000). However, the identity of the cation conductance generating the PP is still unknown.
We have investigated the contribution of cyclic nucleotide-gated (CNG) channels to the depolarizing current during the plateau potential. CNG channels are nonselective cation channels that in some configurations are also permeated by Ca2+ (Zagotta and Siegelbaum, 1996; Zufall et al., 1997). Several species of CNG channels have been cloned (Finn et al., 1996; Zagotta and Siegelbaum, 1996), and the olfactory CNG is expressed in the hippocampus (el-Husseini et al., 1995; Kingston et al., 1996; Bradley et al., 1997; Wei et al., 1998). The CNG channel is most highly expressed in the soma and proximal dendrites of pyramidal cells (Bradley et al., 1997). Cultured hippocampal neurons exhibit a cation current that is activated by a cGMP analog (Kingston et al., 1996; Bradley et al., 1997). Some (Kingston et al., 1996), but not all laboratories (Bradley et al., 1997) have reported the expression of the rod CNG channel in the hippocampus. The CNG channel is a potential candidate for the depolarizing current underlying the PP because cGMP metabolism is increased by muscarinic receptor or mGluR activation (Trivedi and Kramer, 1998; Wotta et al., 1998). Also, Ca2+ influx could potentially activate guanylate cyclase (GC) by stimulating formation of nitric oxide (NO) (Kingston et al., 1999).
Our strategy to delineate the roles for CNG channels in the PP required the use of whole-cell voltage clamping to quantify the tail current (Itail) underlying the PP. We first ensured that we could voltage clamp the tail current, and then we examined the sensitivity of the current to pharmacological antagonists to the CNG channel. We report that blocking soluble GC (sGC) with 6-anilino-5,8-quinolinequinone (LY83583) (Leinders-Zufall and Zufall, 1995) or 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ) (Garthwaite et al., 1995) depressed the generation ofItail. Antagonists to the CNG channel itself (2′,4′-dichlorobenzamil orl-cis-diltiazem) (Zagotta and Siegelbaum, 1996; Wei et al., 1998) also reversibly inhibitedItail; however, blocking the protein kinase activated by cGMP [protein kinase G (PKG)] and antagonists to NO synthase (NOS) had no effect onItail. Therefore, we conclude that Ca2+ influx in combination with muscarinic stimulation leads to cGMP formation that depolarizes pyramidal neurons by opening CNG channels.
MATERIALS AND METHODS
Hippocampal slice preparation. Hippocampal slices were prepared from Sprague Dawley rats, aged postnatal days 15–23. A block of tissue containing the hippocampus and surrounding structures was attached to a mounting tray with cyanoacrylate glue and immersed in chilled (0–4°C) modified oxygenated (95% O2and 5% CO2) artificial CSF (aCSF) containing (in mm): 120 NaCl, 3.0 KCl, 1.3 MgSO4, 2.0 CaCl2, 1.5 KH2PO4, 26 NaHCO3, and 10 d-glucose, pH 7.35. Horizontal slices (400 μm) were cut through the tissue block using a vibratome (VT100; Leica, Willowdale, Ontario, Canada). The slices were then transferred to a storage chamber with oxygenated aCSF and allowed to recover for at least 1 hr at room temperature.
Electrophysiology. Whole-cell voltage-clamp recordings (Hamill et al., 1981) from CA1 neurons within hippocampal slices were obtained using the “blind-patch” technique (Blanton et al., 1989). Slices were individually transferred to a recording chamber located on an upright microscope (Axioskop; Zeiss, Oberkochen, Germany) and submerged in rapidly flowing (1 ml/min) oxygenated aCSF. Bath temperature was maintained at 32–34°C with a Peltier unit and Cambion bipolar controller. The recording electrodes were pulled from 1.5 mm (outer diameter) borosilicate thin-walled glass capillaries (150F-4; World Precision Instruments, Sarasota, FL) in three stages on a Flaming-Brown micropipette puller (model P-87; Sutter Instruments, Novato, CA). Patch electrodes were filled with a solution containing (in mm): 115 Cs-methanesulphonate, 20 KCl, 10 Na-phosphocreatine, 10 HEPES, and 1.1 EGTA, pH 7.25. In some experiments, 10 mm BAPTA was substituted for the EGTA as described. When filled with intracellular solution, patch electrode resistance ranged from 4 to 8 MΩ. All experiments were conducted in extracellular solution containing tetrodotoxin (TTX) (1.2 μm).
Membrane potentials and/or currents were monitored with either an Axoclamp 2A or Axopatch 200B amplifier (Axon Instruments, Foster City, CA), acquired through a Digidata 1200 series analog-to-digital interface onto a Pentium personal computer using Clampex 7.0 software (Axon Instruments). Data were sampled at a rate of 2–5 kHz and were low-pass filtered (four-pole Bessel) at 1–5 kHz. Series resistance was continuously monitored with hyperpolarizing voltage steps. Recordings with series resistance >20 MΩ were rejected from analysis.
All chemicals were purchased from Sigma (St. Louis, MO), Molecular Probes (Eugene, OR), or Calbiochem (La Jolla, CA). Carbachol (Sigma), tetrodotoxin (Sigma), l-cis-diltiazem (Sigma), NG-nitro-l-arginine (l-NNA) (Sigma), and NG-nitro-l-arginine methyl ester, HCl (l-NAME) (Calbiochem) were dissolved in distilled H2O and added to the aCSF from concentrated stocks. ODQ (Sigma), LY83583 (Sigma), 2′,4′-dichlorobenzamil, HCl (Molecular Probes), and zaprinast (Calbiochem) were first made up as a stock in DMSO before being added to the aCSF. The final concentration of DMSO was always ≤0.1%; in control experiments, DMSO at these concentrations did not alterItail. BAPTA, tetracesium salt (Molecular Probes), KT5823 (Calbiochem),l-NNA, and l-NAME were dissolved directly into the patch-pipette solution.
Data analysis. Data were analyzed using Clampfit 8.0 (Axon Instruments). Itail was quantified in each cell by calculating the area under the tail current. Area was calculated by summing the amplitudes of all individual data samples over time, relative to the baseline of prestimulus holding current. Samples were summed from the end of the depolarizing voltage command pulse over a 13 sec time period or to the point at which the current amplitude returned to prestimulus levels. Statistical comparisons were determined using either Student's t test or one-way ANOVA with Tukey's post hoc test (SPSS version 10.0; SPSS, Chicago, IL). In all cases, p < 0.05 was considered significant. Values are reported as mean ± SEM.
RESULTS
The results in this paper were obtained from 134 CA1 pyramidal neurons in the hippocampal slice preparation, using whole-cell patch-clamp techniques. Whole-cell recordings were made with patch pipettes containing a Cs+-based internal solution to reduce K+ currents and to improve space clamp (Colino and Halliwell, 1993). All experiments were performed in the presence of bath-applied TTX (1.2 μm) to block voltage-gated Na+ channels and to prevent Na+-dependent action potentials.
Cholinergic-dependent plateau potential and slow inward tail current
In the absence of carbachol, a nonhydrolyzable cholinergic agonist, positive current injection (800 msec, ≥0.1 nA) resulted in activation of Ca2+-dependent action potentials (n = 15) (Fig.1A). After cessation of current injection, the membrane potential rapidly returned to resting levels. As shown previously with current-clamp recordings (Fraser and MacVicar, 1996), after a 5 min bath application of 20 μm carbachol, current injection evoked Ca2+ spikes that now elicited prolonged PPs (Fig. 1B). The average duration of the PP was 10.3 ± 1.2 sec (range of 3.7–18.4; n = 15). The PPs produced by carbachol were reversible (Fig.1C).
In the presence of carbachol, anItail was observed under voltage clamp in the same cells that revealed a PP in current-clamp mode.A, Typical responses of a hippocampal CA1 pyramidal neuron under current-clamp conditions to hyperpolarizing and depolarizing current injection in control aCSF. Depolarizing current injection elicited robust Ca2+ spikes, and the membrane potential immediately returned to baseline levels after cessation of the current pulse. B, In the presence of 20 μm carbachol (CCH), Ca2+ spike firing evoked by identical stimuli resulted in a long-lasting PP. C, The PP was reversible after carbachol was washed from the slice. D, In the same cell, under voltage-clamp conditions in the absence of carbachol, an 800 msec depolarizing voltage pulse to 0 mV from a holding potential of −70 mV resulted in unclamped Ca2+ currents. At the offset of the pulse, Itail was not observed. E, In the presence of carbachol, a long-lasting inward Itail was induced at the offset of the depolarizing voltage pulse under voltage-clamp conditions. F, Itail was reversible after wash of carbachol. Recordings were made with a Cs+-based intracellular solution and 1.2 μm TTX in the external solution.
Long-lasting Itails were observed under voltage clamp in the same pyramidal neurons, which had displayed a PP in carbachol under current-clamp conditions (n = 15/15) (Fig. 1). We examined Itails at a holding potential of −70 mV after a voltage step to 0 mV for 800 msec before and after carbachol application. The inwardItail was observed in a total of 77 pyramidal cells only after 20 μm carbachol was applied. Tail currents were quantified by measuring the area of the inward current over a 13 sec time period after the end of the voltage command pulse as described in Materials and Methods. Significant increases in Itail areas were induced by carbachol (control area, −0.04 ± 0.02 nA · sec vs carbachol, −1.2 ± 0.2 nA · sec; p < 0.001;n = 15) (Fig.1D,E); when carbachol was washed,Itail diminished to control levels (−0.1 ± 0.02 nA · sec; p < 0.001;n = 15) (Fig. 1F).
Reversal potential and sodium dependence ofItail
Previously, we reported that activation of PPs depends on increased intracellular [Ca2+] ([Ca2+]i) and that the depolarization is attributable to TTX-insensitive Na+ influx through a nonselective cation channel (Fraser and MacVicar, 1996). We investigated the role for TTX-insensitive Na+ influx inItail by replacing extracellular [Na+] ([Na+]o) with equimolar N-methyl-d-glucamine. Reducing [Na+]ofrom 152 to 26 mm significantly reduced the area of Itail (normal aCSF, −1.8 ± 0.3 nA · sec vs low sodium, −0.3 ± 0.1 nA · sec; p < 0.002; n = 6) (Fig.2A,B);Itail recovered after replacement of the NaCl (−1.5 ± 0.3 nA · sec compared with control; p = 0.63; n = 6) (Fig. 2C). The results of these experiments are summarized in the plot ofItail areas in Figure2D.
Itail was dependent on Na+ influx independent of TTX-sensitive channels, varied linearly with voltage, and was blocked by intracellular BAPTA.A, In 20 μm carbachol,Itail was observed after a depolarizing voltage step (800 msec) to 0 mV from a holding potential of −70 mV.B, Reducing [Na+]o from 152 to 26 mm depressed Itail. External NaCl was substituted with equimolarN-methyl-d-glucamine. C, Restoring external NaCl to 152 mm completely reversed the depression of Itail induced by low [Na+]o. D, Summary of the effects of reducing [Na+]o and after washout on the area of Itail. Mean ± SE areas of Itail are plotted (n = 6). *p < 0.002 compared with control data. E, The reversal potential ofItail was assessed by ramping the voltage from −50 to −100 mV in 100 msec in the presence of carbachol. Ramps obtained during Itail were compared with ramps obtained from identical voltages without an evokedItail. F, Subtraction of the ramp currents without Itail(ramp 1) from the ramp currents duringItail (ramp 2) from −65 to −100 mV revealed that Itail varied linearly with membrane potential in the voltage range tested. The reversal potential in this cell determined by linear extrapolation (dotted line) was −20.1 mV. G, Superimposed recordings in carbachol from different CA1 pyramidal neurons from the same slice loaded with intracellular pipette solution containing either 1.1 mm EGTA or 10 mm BAPTA (indicated by arrows). Chelating [Ca2+]i with BAPTA depressedItail. H, Summary of the mean ± SE Itail areas from pyramidal neurons recorded with either (in the micropipette) 1.1 mmEGTA (n = 10) or 10 mm BAPTA (n = 9). **p < 0.001, EGTA compared with BAPTA.
We next examined the reversal potential forItail to see whether a nonselective cation channel underlies the depolarization. A current attributable to a nonselective cation channel should reverse at approximately −20 to 0 mV. The I–V relationship ofItail was obtained by ramping the holding potential from −50 to 100 mV (in 100 msec) whenItail was evoked after a 800 msec depolarizing prepulse to 0 mV in the presence of carbachol (20 μm) (protocol shown in Fig.2E). Ramp currents obtained during theItail (Fig. 2F2) were compared with ramp currents obtained from identical holding potentials without a prepulse to evokeItail (Fig. 2F1). An example I–V plot is shown in Figure 2F. Subtraction of the control ramp current from the ramp current during the Itail revealed an inward current that showed an extrapolated reversal potential in this cell of approximately −20 mV (Fig. 2F). The inward currents revealed by the ramp subtractions were fitted by a computed linear regression (r2 = 0.98–1), which were then used to extrapolate the reversal potential. The average reversal potential estimated by linear extrapolation was −23.6 ± 8.5 mV (n = 7).
Ca2+ dependence ofItail
In the previous report, PPs were abolished after intracellular perfusion with 10 mm BAPTA, a Ca2+ chelator (Fraser and MacVicar, 1996). In this study, when we included 10 mm BAPTA in the intracellular solution to prevent Ca2+increases, the amplitude of Itail (in the presence of 20 μm carbachol) was greatly depressed (Fig. 2G,H). The mean area ofItail after 15–20 min intracellular perfusion with BAPTA was −0.3 ± 0.08 nA · sec (n = 9). As a control, we recorded Itail in other cells in the same slices using pipettes filled with control intracellular solution that contained 1.1 mmEGTA. In the control cells, Itail was evoked in pyramidal neurons in carbachol, and the area ofItail was significantly increased compared with BAPTA-filled cells (−1.4 ± 0.2 nA · sec;n = 10; p < 0.001). These data suggest that, similar to the PP, activation of Itailis dependent on elevations in [Ca2+]i.
Itail is refractory
Before we could examine the actions of pharmacological agents onItail, we had to ensure thatItail could be reliably evoked at a specific intervals and that it was stable over extended periods of time. We found that, if Itail was evoked with intervals <3 min, the area of the current was significantly reduced (Fig.3A). The normalized area ofItail recorded at various intervals is plotted in Figure 3B, showing the gradual recovery with longer intervals. A single exponential curve was fit to the normalized current areas at intervals of 7.5, 15, 30, 45, 60, 90, and 180 sec after an initial Itail was evoked. The time constant of recovery of Itail to 63% was 43.5 sec (n = 9), calculated from the fit curve. Itail was fully recovered at 180 sec. To examine rundown over an extended time,Itail was activated every 180 sec for up to 1 hr (Fig. 3C). After 51 min, the area of theItail was 99.8 ± 22.9% of the first Itail evoked (n= 6) (Fig. 3C). Plateau potentials in subiculum have also been shown to be refractory when elicited in short intervals (Kawasaki et al., 1999).
Itail was refractory when evoked at short intervals but was stable for up to 1 hr when evoked at >3 min intervals. A,Itail was evoked in the presence of 20 μm carbachol with a depolarizing voltage step to 0 mV from a holding potential of −70 mV (left). Another depolarizing voltage step to 0 mV 15 sec after the initialItail was evoked resulted in anItail with a reduced area (middle). A depolarizing voltage step to 0 mV 180 sec after an initial Itail was evoked resulted in a Itail with a similar area (right). B, Plot of the normalized mean ± SE Itail areas at intervals of 7.5, 15, 30, 45, 60, 90, and 180 sec with respect to the initial evokedItail, showing that activation ofItail had a refractory period (n = 9). A single exponential curve was fit (solid line), and the time constant of recovery was calculated to be 43.5 sec (indicated by dotted lines).C, The normalized mean ± SEItail areas evoked every 3 min were plotted against time for 51 min (n = 6).
Itail requires activation of guanylate cyclase independent of PKG activity
To explore the possibility that carbachol and/or Ca2+ act via a cGMP-dependent pathway (Trivedi and Kramer, 1998; Wotta et al., 1998), we tested inhibitors of steps in cGMP signaling. First, we evaluated the effects of inhibitors of sGC. LY83583 (20 μm), an inhibitor of guanylate cyclase and cGMP channels, reduced the area ofItail (control, −1.7 ± 0.4 nA · sec vs LY83583, −0.3 ± 0.07 nA · sec;p < 0.02; n = 5) (Fig.4A,B). In several neurons, the later portion ofItail was blocked in contrast to the complete lack of residual Itail when BAPTA was included in the pipette or [Na+]o was reduced (Fig. 2). ODQ (20 μm), another sGC inhibitor with no reported action on the cGMP channel (Garthwaite et al., 1995), inhibited the activation of Itail(area in ODQ, −0.5 ± 0.1 nA · sec compared with area of control, −1.5 ± 0.2 nA · sec; p < 0.02;n = 5). The results of the experiments with the guanylate cyclase inhibitors are summarized in Figure 4C. Conversely, we examined the effects of inhibiting cGMP-specific phosphodiesterase by bath application of zaprinast (20 μm) to see whether the area ofItail was enhanced when cGMP levels were potentially increased. We used a submaximal concentration of carbachol (5 μm) to elicitItail. To quantify the actions of zaprinast, we first evoked Itail twice in control solution, and then zaprinast was bath applied for >8 min and Itail was again evoked. All values for Itail, including the washout values, were normalized to the first control. The normalized area of the Itail increased from 1.1 ± 0.1 in control to 2.4 ± 0.6 in zaprinast (n = 10;p < 0.05). The enhancement ofItail was reversible as the area decreased to 1.5 ± 0.3 (not significantly different compared with control) when zaprinast was washed out. In the example shown in Figure4D–F, bath application of zaprinast reversibly enhanced the area of Itail. These results indicate that Itail was dependent on increased cGMP levels that could elicitItail by either activating CNG channels directly in CA1 pyramidal neurons or by activating cGMP-dependent protein kinases (PKG).
Itail required sGC activity but was independent of PKG activity. A, In the presence of 20 μm carbachol (CCH),Itail could be evoked with a depolarizing voltage step to 0 mV from a holding potential of −70 mV.B, After bath application of 20 μmLY83583, Itail was depressed.C, Summary of the mean ± SEItail areas in the presence of the sGC inhibitors LY83583 or ODQ or a PKG inhibitor compared withItail controls. Bath application of either 20 μm LY83583 (n = 5) or 20 μm ODQ (n = 5) significantly depressed Itail area. *p< 0.02, compared with control Itail areas. Intracellular perfusion of the PKG inhibitor KT5823 (10 μm) for >30 min did not alterItail area (n = 5, KT5823; n = 5, control).Itail areas from cells perfused with KT5823 were compared with control areas of Itailobtained from pyramidal neurons in the same slices. D,Itails of reduced areas were evoked with a depolarizing voltage step to 0 mV from a holding potential of −70 mV with perfusion of a submaximal dose of carbachol (5 μm).E, Bath application of zaprinast enhanced the evokedItail. F, The increase inItail area by zaprinast was reversible after wash.
Although we found previously that activation of protein kinases and phosphorylation was unnecessary for PP genesis (Fraser et al., 2001), we determined the involvement of PKG, which can be activated by cGMP. Intracellular perfusion of the selective PKG inhibitor KT5823 (10 μm) (Lei et al., 2000) in the patch pipette for 30 min failed to inhibit Itail (Fig.4C). In the same slices in which control intracellular solution was used, there was not a significant difference in the area of Itail (control, −1.8 ± 0.2 nA · sec vs KT5823, −1.6 ± 0.1 nA · sec;p=0.36; n = 5). Therefore, we conclude that a guanylate cyclase—cGMP pathway, independent of PKG, is required for activation of Itail.
Itail is mediated by cGMP-gated cation channels
To test the hypothesis that Itailis mediated by cGMP-gated cation channels, we used bath applications of two CNG channel blockers, 2′,4′-dichlorobenzamil andl-cis-diltiazem (Koch and Kaupp, 1985;Nicol et al., 1987). The first CNG channel blocker that we used was the amiloride derivative 2′,4′-dichlorobenzamil (100 μm). In the example shown in Figure5A–C, application of 2′,4′-dichlorobenzamil reversibly reduced the area of theItail. The results of these experiments are summarized in the plot of theItail area in Figure 5D. Significant changes in Itail area were induced by 2′,4′-dichlorobenzamil (control, −2.1 ± 0.2 nA · sec vs 2′,4′-dichlorobenzamil, −0.3 ± 0.06 nA · sec; p < 0.001; n = 7); this reduction in area of Itail was reversible (−1.6 ± 0.2 nA · sec; p = 0.2 compared with control; n = 7).
Antagonists of cyclic nucleotide-gated channels depressed Itail. A, In the presence of 20 μm carbachol (CCH),Itail was evoked with depolarizing voltage steps to 0 mV from a holding potential of −70 mV. B, Bath application of 100 μm 2′,4′-dichlorobenzamil (DCB) depressed generation ofItail. C, The inhibition ofItail was reversible after wash of 2′,4′-dichlorobenzamil. D, Summary plot of the mean ± SE areas of Itail before, after, and wash of 2′,4′-dichlorobenzamil. Application of 2′,4′-dichlorobenzamil reversibly depressed Itail(n = 7; *p < 0.001) compared with Itail control. E, Summary of the effects of bath application ofl-cis-diltiazem on mean ± SEItail area. Itailwas depressed by l-cis-diltiazem (n = 7; **p < 0.01 compared with control). In three neurons that were stable in the wash for >30 min, Itail partially recovered from the depression induced by l-cis-diltiazem.
We also examined the effects of l-cis-diltiazem on the carbachol-activated Itail.l-cis-diltiazem is the inactive isomer of a Ca2+ channel blocker that has been reported to inhibit CNG channels. The site of action ofl-cis-diltiazem is suggested to be located on the cytoplasmic side of the channel (Koch and Kaupp, 1985;Stern et al., 1986; Haynes, 1992; McLatchie and Matthews, 1992, 1994). However, at physiological pH, ∼50% of thel-cis-diltiazem is unprotonated and can cross the membrane. Therefore,l-cis-diltiazem was bath applied. Figure 5E shows thatl-cis-diltiazem (100 μm) inhibited activation ofItail. The area ofItail was reduced from −1.8 ± 0.4 nA · sec in control to −0.4 ± 0.05 nA · sec after application of l-cis-diltiazem (p < 0.01; n = 7). A subset of these cells (n = 3) were stable for >30 min in wash in carbachol containing aCSF, and Itailrecovered to −0.9 ± 0.1 nA · sec in these cells. Similar to the effects of sGC inhibition, there was still a residualItail observed in several neurons after inhibition of CNG channels with 2′,4′-dichlorobenzamil (Fig. 5) or l-cis-diltiazem.
Effect of nitric oxide synthase inhibitors onItail
We tested the possibility that NO causes the increased guanylate cyclase activity required for activation ofItail by applying two inhibitors of NOS. Slices were bathed for >1 hr in aCSF containing the NOS inhibitors l-NAME (1 mm) orl-NNA (100 μm).Itails evoked in these slices did not differ from those evoked in control slices [control, −1.6 ± 0.5 nA · sec vs l-NAME, −1.4 ± 0.3 nA · sec (p = 1.00; n = 5) or vs l-NNA, −1.3 ± 0.2 nA · sec (p=0.98; n = 5)]. In addition, neither l-NAME (1 mm) (control, −1.6 ± 0.5 nA · sec vsl-NAME, −1.6 ± 0.4 nA · sec;p=1.00; n = 10) norl-NNA (500 μm) (control, −1.6 ± 0.5 nA · sec vs l-NNA, −1.6 ± 0.3 nA · sec; p=1.00; n = 10) had any significant effect onItail when included in the intracellular recording electrode (Fig. 6).
DISCUSSION
The results in this paper indicate that CNG channels contribute to the prolonged depolarization during the cholinergic plateau potential. We used whole-cell voltage-clamp recordings to quantify theItail after voltage steps to activate Ca2+ currents in CA1 pyramidal neurons. A significant Itail was only observed when carbachol was perfused. The reversal ofItail at −20 mV indicated that the current was likely attributable a nonselective cation channel. The inward current was carried principally by Na+ and was blocked by including high concentrations of the Ca2+ chelator BAPTA in the pipette. The current did not exhibit significant rundown over 1 hr when evoked at intervals of >3 min. At shorter intervals,Itail was refractory. Inhibiting sGC with ODQ or LY83835 depressed the current, and, conversely, inhibiting phosphodiesterase, which degrades cGMP, enhancedItail. Two separate antagonists of the CNG channel, l-cis-diltiazem and 2′,4′-dichlorobenzamil, depressedItail. Inhibition of the kinase activated by cGMP (PKG) using protocols shown to be effective in other studies (Lei et al., 2000) had no effect. These results indicate that Ca2+ influx activates sGC, leading to increased cGMP levels. The increased cGMP activates CNG channels, causing a depolarization attributable to the nonselective cation conductance. Surprisingly, we could find no evidence that the activation of sGC was attributable to increased NO because high concentrations of NOS inhibitors had no effect onItail.
Itail was independent of nitric oxide production. A, The effects of the nitric oxide inhibitors l-NAME and l-NNA are summarized in the histogram. Itail area was not depressed by either bath application for >1 hr (1 mm) (n = 5) or intracellular perfusion (1 mm) (n = 10) of l-NAME. Inhibition of nitric oxide synthase by bath application (n = 5) or intracellular perfusion (n = 10) of l-NNA did not affectItail area.
CNG channels are a family of related proteins that consist, in the native form, of α subunits, which form homomeric pores, and β subunits, which modify channel properties and sensitivity to antagonists when coexpressed with α subunits (Zagotta and Siegelbaum, 1996; Wei et al., 1998; Bonigk et al., 1999). Distinct CNG channels were first described in retinal rod and cone cells and in olfactory cells and are classified as CNCα1 (rod CNG channel), CNCα2 (cone CNG channel), and CNCα3 (olfactory CNG channels). There is substantial evidence, however, that CNG channels are more widely expressed and that they play roles in synaptic function in other brain regions, such as the hippocampus (Wei et al., 1998; Kingston et al., 1999). The olfactory CNG channel (CNCα3) is expressed in hippocampal pyramidal neurons, and this channel is highly permeable to Ca2+ (el-Husseini et al., 1995; Kingston et al., 1996; Bradley et al., 1997; Wei et al., 1998). The rod CNG channel was found in the hippocampus by some (Kingston et al., 1996) but not all investigators (Bradley et al., 1997). Although application of membrane-permeable forms of cGMP [8-bromo-cGMP (8-Br-cGMP)] induced an inward current in cultured hippocampal neurons (Leinders-Zufall et al., 1995; Bradley et al., 1997), the functions of CNG channels in the hippocampus are unknown. Increased cGMP is apparently involved in the induction of LTP because guanylate cyclase inhibition and Rp-8-Br-cGMPs, a cGMP-dependent protein kinase antagonist, blocked induction of LTP in CA1 (Zhuo et al., 1994; Arancio et al., 1995). In addition, transgenic mice lacking the olfactory CNG α subunit exhibited an attenuation of LTP (Parent et al., 1998).
Our results indicate that activation of CNG channels contribute to the depolarization during the plateau potential. Our results do not differentiate between olfactory or rod type of CNG channels. There are differences between the two CNG channels that may be used to determine channel type in future experiments. Olfactory CNG channels are blocked by LY83583 (Leinders-Zufall and Zufall, 1995). We observed depression by LY83583; however, LY83583 also inhibits sGC. We also observed inhibition of Itail by ODQ, another inhibitor of sGC. Therefore, our results do not differentiate between these two possible actions of LY83583, and we cannot attribute the actions of LY83583 to either enzyme inhibition or channel block. This would be best resolved in future experiments by directly activating CNG currents in pyramidal neurons in hippocampal slices by cGMP itself and then testing the effects of LY83583. Inhibition of sGC or antagonists of CNG channels often did not totally blockItail. In contrast, there was no residual Itail in BAPTA or low [Na+]o. It is possible that there are multiple components ofItail that we have not resolved.
There is substantial Ca2+ permeation of some forms of CNG channels, particularly the olfactory form (Frings et al., 1995; Leinders-Zufall et al., 1997; Dzeja et al., 1999). The likeliest CNG channel underlying this current in the hippocampus is the olfactory CNG channel (Kingston et al., 1996; Bradley et al., 1997). This implies that Ca2+ influx will occur in pyramidal cells when Itail and CNG channels are activated, which could have profound effects on cell signaling. We have confirmed here that the plateau potential is prevented by BAPTA, which will chelate and attenuate rises in [Ca2+]i. Therefore, Ca2+ influx through CNG channels would be expected to increase the Ca2+ load in pyramidal neurons, thereby leading to further activation of the plateau potential. The degree of Ca2+ permeability of CNG channels is also determined by the β subunits that are coexpressed with α (Dzeja et al., 1999). If the CNG channel configuration in hippocampal slices (Bradley et al., 1997) is identical to that found in hippocampal cell culture (Kingston et al., 1996), then there should be considerable Ca2+ permeability.
Our results point to several key components in the pathway leading to activation of the CNG channel and the plateau potential, as illustrated in Figure 7. Key steps are the Ca2+ influx leading to increased [Ca2+]i and the concurrent activation of muscarinic receptors. We propose that the concurrent stimulation of muscarinic receptors and increased [Ca2+]i leads to activation of sGC and increased cGMP levels. The increased levels of cGMP induces opening of CNG channels, causing the depolarizing current underlying the plateau potential. Both LY83583 and ODQ are potent inhibitors of sGC, and both depressedItail. The two CNG channel antagonists depressed Itail, consistent with our conclusion that a component of the inward current is attributable to the CNG cation current. The extrapolated reversal potential forItail was close to an estimated reversal potential for a mixed cation conductance. We also propose that zaprinast-sensitive phosphodiesterases contribute to the termination of the CNG current because Itail was enhanced in zaprinast. Protein phosphatase activation also contributes to the generation of the plateau potential (Fraser et al., 2001) in a manner similar to the modulation of Ca2+-activated K+ currents by muscarinic receptors (Pedarzani et al., 1998). The transduction pathway between increased Ca2+ and activation of sGC is still to be determined. When we started these experiments, we hypothesized that Ca2+-dependent activation of NOS could increase NO generation, causing sGC activation (Wotta et al., 1998). Alternatively, muscarinic receptors have been linked to activation of endothelial NOS (Han et al., 1998). However, our results do not support any role for NOS in activation of sGC. Quite high concentrations of two different NOS inhibitors had no effect onItail. Therefore, the pathway leading to sGC activation is still to be determined. There are alternative explanations for our observations, such as modification of [Ca2+]i mechanisms by cGMP. Conclusive evidence for this model will come from more direct studies in hippocampal neurons on the CNG channel itself.
Proposed model for the activation ofItail and generation of plateau potentials. We propose that stimulation of muscarinic receptors (m1/m3) coupled to G-proteins in combination with Ca2+ influx through high-voltage-activated Ca2+ channels (HVA) can activate sGC, leading to an increase in intracellular cGMP and opening of CNG channels. The mechanism by which Ca2+ activates sCG does not apparently require nitric oxide. Influx of Na+ and Ca2+ through CNG channel openings mediates Itail and the prolonged depolarization during the plateau potential. Activation of protein phosphatase (PP) is required for plateau potential generation (Fraser et al., 2001), which may serve to increase the sensitivity of CNG channels to cyclic nucleotides. As part ofItail termination, cGMP-specific phosphodiesterase (PDE) can metabolize cGMP to 5′-GMP.
These findings could be relevant to the generation of seizures and epilepsy. A generalized seizure in the whole animal involves a prolonged depolarization, which is termed the tonic component of the ictal seizure (Dichter and Ayala, 1987). Often this is followed by repetitive depolarizations called the clonic phase. There is no doubt that recurrent collaterals are important in synchronizing neuronal activity in networks leading to seizures (MacVicar and Dudek, 1980;Traub et al., 1989; Jefferys, 1998), and it is likely that multiple currents contribute to seizure generation. However, the conversion of neuronal networks from bursting to prolonged tonic depolarizations could entail the enhancement of a prolonged inward current. The CNG current underlying Itail is an excellent candidate for an intrinsic current that could be a key contributor to the ictal tonic depolarization.
Footnotes
This work was supported by grants from Canadian Institutes of Health Research (CIHR). J.B.K. was supported by a studentship from the Savoy Foundation, and B.A.M. held Senior Scientist awards from CIHR and Alberta Heritage Foundation for Medical Research. We thank Dr. P. Schnetkamp for helpful comments.
Correspondence should be addressed to Dr. Brian A. MacVicar, Neuroscience Research Group, Department of Physiology and Biophysics, 3330 Hospital Drive N.W., Faculty of Medicine, University of Calgary, Calgary, Alberta T2N 4N1, Canada. E-mail: macvicar{at}ucalgary.ca.
